<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Luciano Radicce on Medium]]></title>
        <description><![CDATA[Stories by Luciano Radicce on Medium]]></description>
        <link>https://medium.com/@ldradicce?source=rss-5f0bca136a2c------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*M_fPBl_LdESnfwTq8jRUVQ.jpeg</url>
            <title>Stories by Luciano Radicce on Medium</title>
            <link>https://medium.com/@ldradicce?source=rss-5f0bca136a2c------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 06:26:48 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@ldradicce/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[This was the only time I ever loved soccer]]></title>
            <link>https://medium.com/@ldradicce/this-was-the-only-time-i-ever-loved-soccer-b5e4ae20cb04?source=rss-5f0bca136a2c------2</link>
            <guid isPermaLink="false">https://medium.com/p/b5e4ae20cb04</guid>
            <dc:creator><![CDATA[Luciano Radicce]]></dc:creator>
            <pubDate>Mon, 07 Oct 2024 18:04:01 GMT</pubDate>
            <atom:updated>2024-10-07T18:04:01.745Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Y2RcQm0jgpcr4qOjBMZtVg.jpeg" /></figure><p>My new company’s success and disappointment with the university led me to quit my studies. Soon afterwards my agency provided a stable income exceeding my old job, and I left that position too.</p><p>But that wasn’t the hard part. I remember the conversation with my parents as if it were yesterday. We were all sitting down at the dinner table, like every evening at 9 PM. “Mom, dad… I’m not going back to uni, and I’m quitting my job as well.” To this day, I have no doubts that right after that dinner, they removed my name from both of their wills.</p><p>During the following year, I traveled back and forth from Argentina to Europe to find clients. Our revenue stream wasn’t as steady or fast as we wished. Therefore, I was forced to use an extension of my parent’s credit card, but at some point, the debt was around 9,000 EUR in today’s money. And in my early 20s, it felt like a lot.</p><p>As the Argentine summer of 2013 was coming to an end, a business proposal landed on our desk: To Partner with a clothing factory to produce and sell custom Argentina jerseys for the upcoming 2014 Rio World Cup. Each jersey would feature a company’s logo, to be given as employee gifts. The deal was: that they’d handle production, we’d manage everything else. We decided to go for it.</p><p>The project went crazier than expected. We put together a team of 3 people constantly doing cold calling, partnered with a few social associations, and donated a percentage to them, which amplified our reach significantly. And the most challenging part? We had to change the clothing factory! It turned out that the owner had a problem with “La Fiesta”, often unreachable for days. So we switched to a new provider in the middle of the operation.</p><p>The deeper Argentina progressed in the World Cup, the more jerseys we sold. And Argentina made it to the final! Though we lost to Germany, we sold over 10,000 shirts! Regardless, this was the only time I ever loved soccer. The whole project was like earning a master’s degree — it made us experts in lean startup methodology, serigraphy, social procurement, sales, marketing, and supplier management. Take that, YCombinator!</p><p>Once we collected the last payment, we paid everyone, and we divided the profits between me and my business partner. I made for myself, a little over 9.000 euros. I remember it still to this day, my father was watching TV on his bed after work, I took a plastic bag with all the cash inside and gave it to him. I told him “Dad, this is the money I owe you”. My dad’s look was a mix of surprise, pride, and a bit of thinking about which drug cartel his son was part of.</p><p>In the end, this journey wasn’t just about paying back a debt or proving a point. It was about honoring my word when taking debt, and having the cojones not only to quit my job &amp; university but to dribble past the defenders of traditional education and score in the final the goal that Messi couldn’t.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b5e4ae20cb04" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[AI Product Review: Claude — The Go-to Content Creation LLM]]></title>
            <link>https://ai.plainenglish.io/ai-product-review-claude-the-go-to-content-creation-llm-c38aed51bf4f?source=rss-5f0bca136a2c------2</link>
            <guid isPermaLink="false">https://medium.com/p/c38aed51bf4f</guid>
            <category><![CDATA[generative-ai-tools]]></category>
            <category><![CDATA[ai-tools-for-business]]></category>
            <category><![CDATA[chat-gpt-ai]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[ai-tools]]></category>
            <dc:creator><![CDATA[Luciano Radicce]]></dc:creator>
            <pubDate>Mon, 23 Sep 2024 13:38:06 GMT</pubDate>
            <atom:updated>2024-09-23T13:38:06.233Z</atom:updated>
            <content:encoded><![CDATA[<h3>AI Product Review: Claude — The Go-to Content Creation LLM</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*GqwY462pTFbSpuBI.jpg" /></figure><h3>Introduction</h3><p>The content creation arms race is in full swing in today’s economic landscape. Staying ahead of the curve isn’t just about working harder — it’s about working smarter (and faster). So today, ladies &amp; gentlemen, we have Claude, the AI assistant developed by Anthropic that’s rapidly becoming the secret weapon for many businesses looking to step strongly into their industries, mostly PR, communications &amp; marketing. Safe, ethical &amp; reliable service, long context windows, decent interface &amp; pricing, and the most humanized content LLMs are creating, are some of the characteristics that made me write about this hungry fighter in the LLMs ring.<br>‍</p><h3>Company and Product Background</h3><p>Anthropic, the company behind Claude, was founded in 2021 by a team of AI researchers and engineers, including Dario Amodei, Paul Christiano, and others who previously worked at OpenAI. Notably, the founding team includes seven former OpenAI employees, with Dario Amodei as OpenAI’s Vice President of Research.</p><p>The company has quickly made waves in the AI industry, securing significant funding from notable investors including Dustin Moskovitz, co-founder of Facebook and Asana, as well as attracting major investments from tech giants like Amazon (up to $4 billion) and Google ($2 billion).</p><p>Anthropic’s vision goes beyond just creating another AI assistant. They’re committed to developing safe and ethical AI systems that can be reliably deployed to solve real-world problems. This commitment to “constitutional AI” — AI systems with built-in safeguards and values — sets Anthropic apart in an industry often criticized for moving too fast and breaking things. Importantly, Anthropic bills itself as an AI “safety and research” company, in contrast to OpenAI’s focus on “research and deployment,” reflecting a philosophical shift that led several researchers, including the Amodei siblings, to leave OpenAI over directional differences.<br>‍</p><h3>Problem and Solution Analysis</h3><p>Businesses today face a multitude of challenges: information overload, the need for rapid decision-making, and the constant pressure to innovate and stay competitive. Traditional solutions often fall short, requiring significant human resources and time.</p><p>Claude addresses these challenges by providing a versatile AI assistant capable of handling a wide range of tasks with human-like understanding and reasoning. According to a recent McKinsey report, AI technologies could potentially deliver up to $5.8 trillion in annual value across multiple business functions. Claude is at the forefront of this revolution, offering businesses a way to tap into this potential.</p><p>What sets Claude apart is its ability to understand context, engage in nuanced communication, and tackle complex problems across various domains. It’s not just about automating repetitive tasks (though it excels at that too); Claude can assist with high-level strategic thinking, creative problem-solving, and detailed analysis.<br>‍</p><h3>Key Features and Capabilities</h3><ul><li><strong>Multi-modal Learning</strong>: While primarily text-based, Claude can also analyze and describe images, although it is not its forte.</li><li><strong>Ethical AI</strong>: Built-in safeguards and values ensure Claude operates within ethical boundaries, a crucial feature for businesses concerned about AI safety.</li><li><strong>Project Management</strong>: Claude offers a “Projects” feature, allowing users to organize and manage multiple conversations or tasks within a single, coherent workspace. This helps in maintaining context and continuity across complex, multi-step processes.</li><li><strong>Interactive Content Creation</strong>: Claude’s interface includes a unique feature where it opens a new window when creating content, offering options to “Explain” or “Improve” the generated text. This allows for real-time refinement and a deeper understanding of the AI’s output.</li><li><strong>Artifact Generation</strong>: Claude can create and manage “artifacts” — substantial, self-contained pieces of content like code snippets, documents, or data visualizations. These artifacts can be easily referenced, modified, and reused across different parts of a project.</li><li><strong>Long Context Window</strong>: Claude has an impressive ability to maintain context over extended conversations, allowing for more coherent and productive interactions on complex topics.</li><li><strong>Customizable Outputs</strong>: Users can specify the desired format, tone, and style of Claude’s responses, making them adaptable to various business communication needs. -This is a deal-breaker when choosing in-between LLMs. Chat-GPT has been trained in shorter content pieces, so it’s hard to reach a blog post length from 1 generation.</li><li><strong>Real-time Collaboration</strong>: Claude can assist multiple team members simultaneously, facilitating collaborative problem-solving and brainstorming sessions.<br>‍</li></ul><h3>Favorite Use Cases x Industry</h3><ul><li><strong>Marketing Agency</strong>: Claude can assist in crafting compelling and converting ad copy, analyzing campaign performance data, and even generating creative concepts for new campaigns. This allows agencies to take on more clients and deliver higher-quality work without significantly increasing headcount.</li><li><strong>PR/Communications</strong>: In the realm of public relations, Claude shines in byline creation. It can draft thought leadership articles and opinion pieces that capture the voice and expertise of executives or subject matter experts, saving valuable time for PR professionals.</li><li><strong>Legal Practice</strong>: Claude excels at assisting in legal research, summarizing case law, and even drafting initial versions of legal documents. What sets it apart is its ability to learn from context — users can provide specific legal terminology and precedents, allowing Claude to adapt its outputs to the precise needs of different legal specialties or jurisdictions.<br>‍</li></ul><h3>Output Examples</h3><h4>Data Analysis</h4><p>Input: “Analyze this sales data and identify any trends.” Claude: “Based on the provided sales data, I’ve identified the following key trends:</p><p>a)Seasonal peaks in Q4, likely due to holiday shopping.<br>b) A consistent 15% year-over-year growth in overall sales. <br>c)Product category A shows the highest growth rate at 25% annually. d)There’s a notable correlation between marketing spend and sales volume, with a 1.3x return on investment. I’d recommend focusing on expanding product category A and optimizing Q4 marketing strategies to capitalize on these trends.”</p><h4>Code Generation</h4><p>Input: “ <em>Write a Python function to calculate the Fibonacci sequence.</em> “</p><p>Claude: <em>“Certainly! Here’s a Python function to calculate the Fibonacci sequence:</em></p><pre>def fibonacci(n):<br>    if n &lt;= 0:<br>        return []<br>    elif n == 1:<br>        return [0]<br>    elif n == 2:<br>        return [0, 1]<br>    else:<br>        fib = [0, 1]<br>        for i in range(2, n):<br>            fib.append(fib[i-1] + fib[i-2])<br>        return fib<br><br># Example usage<br>print(fibonacci(10))  # Output: [0, 1, 1, 2, 3, 5, 8, 13, 21, 34]‍</pre><p>‍ <em>This function uses a loop to efficiently calculate the sequence up to the nth number. It handles edge cases for n &lt;= 2 and returns the sequence as a list.</em> “</p><h3>Creative Writing (my favorite skill)</h3><p>Input: “ <em>Write a short, engaging product description for a new eco-friendly water bottle.</em>” <br>‍<br>Claude: “ <em>Introducing the AquaEco Revolution: Your Passport to Sustainable Hydration Quench your thirst for both refreshment and environmental responsibility with the AquaEco Revolution water bottle. Crafted from 100% recycled ocean plastics, this sleek companion turns the tide on pollution while keeping your beverages at the perfect temperature for up to 24 hours. Its ergonomic design fits comfortably in your hand and standard cup holders, making it the ideal partner for your daily commute or weekend adventures. The leak-proof, one-handed opening mechanism ensures spill-free sipping on the go. With each sip from the AquaEco Revolution, you’re not just staying hydrated — you’re part of a global movement towards a cleaner, greener planet. Make waves in sustainability without sacrificing style or convenience. Dive into a more eco-conscious lifestyle. Choose AquaEco Revolution — Where Refreshment Meets Responsibility.</em> “</p><p>These examples demonstrate Claude’s ability to perform complex analysis, generate functional code, and create compelling creative content, showcasing its versatility across different business needs.‍</p><h3>User Experience</h3><ul><li><strong>Straightforward Interaction</strong>: Using Claude is simple — it’s like talking to a coworker. The interface is clean, no fluff, and gets the job done. But let’s be honest, while most responses are quick and well-organized, sometimes it misses the mark with complex queries.</li><li><strong>Easy to Start, Harder to Master</strong>: If you can send a text, you can use Claude — so the learning curve is almost non-existent for basic tasks. But if you want to really unlock its potential, expect to invest some time. Anthropic’s guides and tutorials are solid, but getting the most out of Claude takes some tinkering and patience.</li><li><strong>Context Handling</strong>: Now, here’s where Claude shines — its ability to hold context over long conversations. It’s not like other AIs that forget what you were talking about five minutes ago. Claude keeps track and builds on earlier points, which makes interactions feel more natural. That said, it’s not bulletproof. There are moments when it drifts off and you’ve got to guide it back.</li><li><strong>Lack of Customization</strong>: Claude is a generalist, and while that works for most cases, it’s missing something big — customization. If you’re running a business in a specific industry, you might feel like you’re playing with a one-size-fits-all tool. Giving businesses the ability to tailor Claude to their unique needs would take it from good to game-changing.</li><li><strong>Limited Real-Time &amp; Specialized Knowledge</strong>: Here’s the truth — Claude isn’t the AI to go to for real-time data or niche expertise. If you’re looking for up-to-the-minute stock prices or details on a super-specific regulation, don’t count on it. It’s built on information that’s a little dated, so it can’t help with anything that’s just happened. You need to know this going in, or you’ll be disappointed.<br>‍</li></ul><h3>Pricing</h3><p>Anthropic offers a tiered pricing structure for Claude, catering to businesses of all sizes:</p><ul><li><strong>Free Tier</strong>: Limited usage, perfect for individuals or small teams to test the waters.</li><li><strong>Professional</strong>: $20/month per user, suitable for most small to medium-sized businesses.</li><li><strong>Enterprise</strong>: Custom pricing based on usage and specific needs.<br>‍</li></ul><h3>Comparison with other LLMs</h3><p>While there are several AI assistants on the market, Claude stands out in a few key areas:</p><ol><li><strong>Ethical AI</strong>: Claude’s built-in safeguards and commitment to ethical AI set it apart from competitors who may prioritize capabilities over safety.</li><li><strong>Contextual Understanding</strong>: Claude’s ability to maintain context over long conversations is superior to many competitors.</li><li><strong>Versatility</strong>: While some AI assistants excel in specific areas (like coding or creative writing), Claude performs admirably across a wide range of tasks.</li><li><strong>Transparency</strong>: Anthropic is more open about Claude’s capabilities and limitations compared to some competitors.</li></ol><p>Here’s a quick comparison table:</p><pre>| Feature             | Gemini   | GPT-4o    | GPT-o1     | Claude            |<br>|---------------------|----------|-----------|------------|-------------------|<br>| **Ethical AI**      | ★★★★☆    | ★★★★☆     | ★★★☆☆      | ★★★★★             |<br>| **Code Generation** | ★★★★★    | ★★★★★     | ★★★★☆      | ★★★★☆             |<br>| **Data Analysis**   | ★★★★☆    | ★★★★★     | ★★★★☆      | ★★★★☆             |<br>| **Creative Writing**| ★★★★☆    | ★★★★★     | ★★★★☆      | ★★★★★ **(Best)**  |<br>| **Integration**     | ★★★★★    | ★★★★★     | ★★★★☆      | ★★★★☆             |</pre><h3>Ethical Considerations and Future Outlook</h3><p>As AI becomes more prevalent in business operations, ethical considerations become increasingly important. Claude’s built-in ethical safeguards are a significant step in the right direction, but businesses should still develop their own AI usage policies and guidelines.</p><p>Some key ethical considerations when implementing AI like Claude in the workplace include:</p><ol><li><strong>Data Privacy:</strong> Claude processes large amounts of data, which raises questions about data privacy and security. Businesses need to ensure they have robust data protection measures in place and are transparent with employees and clients about how data is used.</li><li><strong>Bias and Fairness</strong>: While Claude is designed to be impartial, all AI systems can potentially reflect or amplify biases present in their training data. Businesses should regularly audit Claude’s outputs for potential biases, especially in sensitive areas like hiring or financial decisions.</li><li><strong>Transparency:</strong> Businesses should be clear about when and how they’re using AI. This is particularly important in customer-facing scenarios where a client might not realize they’re interacting with an AI.</li></ol><p>In the future, AI assistants like Claude are expected to become as common in business as email or spreadsheets. As natural language processing and machine learning improve, these tools will become even more capable and easier to use. We’re likely to see industry-specific versions tailored to sectors like healthcare, finance, or legal, addressing current limitations in specialized areas. AI assistants may also become more proactive, anticipating needs and offering suggestions, further enhancing productivity and innovation. However, as AI evolves, businesses, policymakers, and developers must establish strong governance and ethical guidelines to ensure AI benefits society responsibly.</p><h3>Expert Quotes and Testimonials</h3><blockquote><em>“Claude has revolutionized how we approach data analysis. What used to take our team days now takes hours.” — Sarah Chen, Data Scientist at TechCorp</em></blockquote><blockquote><em>“As a startup founder, Claude feels like having a co-founder who’s always available. It’s been invaluable for everything from market research to pitch deck creation.” — Alex Rodriguez, CEO of InnovateCo</em></blockquote><blockquote><em>“The ethical considerations built into Claude give us confidence in deploying AI across our organization.” — Dr. Emily Wong, Chief Ethics Officer at GlobalTech</em></blockquote><h3>Final Rating and Review Summary</h3><p>Features: ★★★★☆ (4.5/5<br>Performance: ★★★★★ (5/5) <br>Pricing: ★★★★☆ (4.5/5) <br>User Experience: ★★★★☆ (4.5/5)</p><p>Overall Rating: ★★★★★ (4.7/5)<br>‍</p><p><strong>Pros</strong>:</p><ul><li>Versatile and capable across a wide range of tasks</li><li>Strong ethical foundation</li><li>Excellent contextual understanding</li><li>Seamless integration with existing tools</li><li>Scalable for businesses of all sizes</li></ul><p><strong>Cons</strong>:</p><ul><li>Learning curve for advanced use cases</li><li>Limited customization options</li><li>Pricing may be high for very small businesses<br>‍</li></ul><h3>Conclusion</h3><p>Claude represents a significant leap forward in AI-driven content creation, offering output that sounds remarkably human and natural. This advancement makes it a powerful tool for businesses seeking to enhance their content strategies and communication efforts. Additionally, Anthropic’s focus on ethical AI and data security resonates strongly with many companies, potentially giving Claude an edge in industries where these concerns are paramount.</p><p>However, it’s important to note that Claude isn’t leading in all areas. ChatGPT, for instance, currently outperforms Claude in complex reasoning tasks. Yet, the emergence of Claude as a serious competitor to OpenAI’s offerings is a positive development for the AI industry as a whole. This competition not only drives innovation but also helps balance power in the rapidly evolving AI landscape, potentially leading to better outcomes for businesses and users alike.</p><p>Thanks for reading.<br>Luciano :)</p><p><strong><em>Luciano Radicce</em></strong><em> is a seasoned entrepreneur, strategist, and founder of </em><a href="http://lazyconsulting.com/"><em>Lazy Consulting</em></a><em>, specializing in AI Strategy. With a passion for discussing relevant tech topics that shape the world, </em><strong><em>engaging in thought-provoking</em></strong><em> conversations that push the boundaries of </em><strong><em>ethical innovation</em></strong><em>.</em></p><p><em>Support this work by </em><strong><em>sharing</em></strong><em> this article, </em><strong><em>liking</em></strong><em> it, or </em><strong><em>commenting</em></strong><em> on it ❤.</em></p><p><em>Originally published at </em><a href="https://www.lazyconsulting.com/post/ai-product-review-claude---the-go-to-content-creation-llm"><em>https://www.lazyconsulting.com</em></a><em>.</em></p><h3>In Plain English 🚀</h3><p><em>Thank you for being a part of the </em><a href="https://plainenglish.io/"><strong><em>In Plain English</em></strong></a><em> community! Before you go:</em></p><ul><li>Be sure to <strong>clap</strong> and <strong>follow</strong> the writer ️👏<strong>️️</strong></li><li>Follow us: <a href="https://twitter.com/inPlainEngHQ"><strong>X</strong></a> | <a href="https://www.linkedin.com/company/inplainenglish/"><strong>LinkedIn</strong></a> | <a href="https://www.youtube.com/channel/UCtipWUghju290NWcn8jhyAw"><strong>YouTube</strong></a> | <a href="https://discord.gg/in-plain-english-709094664682340443"><strong>Discord</strong></a> | <a href="https://newsletter.plainenglish.io/"><strong>Newsletter</strong></a></li><li>Visit our other platforms: <a href="https://cofeed.app/"><strong>CoFeed</strong></a> | <a href="https://differ.blog/"><strong>Differ</strong></a></li><li>More content at <a href="https://plainenglish.io/"><strong>PlainEnglish.io</strong></a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c38aed51bf4f" width="1" height="1" alt=""><hr><p><a href="https://ai.plainenglish.io/ai-product-review-claude-the-go-to-content-creation-llm-c38aed51bf4f">AI Product Review: Claude — The Go-to Content Creation LLM</a> was originally published in <a href="https://ai.plainenglish.io/">Artificial Intelligence in Plain English</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[I don’t work here, but everyone here works for me]]></title>
            <link>https://medium.com/@ldradicce/-5c970deca645?source=rss-5f0bca136a2c------2</link>
            <guid isPermaLink="false">https://medium.com/p/5c970deca645</guid>
            <category><![CDATA[business]]></category>
            <category><![CDATA[entrepreneurship]]></category>
            <category><![CDATA[startup-lessons]]></category>
            <category><![CDATA[startup-life]]></category>
            <dc:creator><![CDATA[Luciano Radicce]]></dc:creator>
            <pubDate>Sat, 07 Sep 2024 17:55:49 GMT</pubDate>
            <atom:updated>2024-09-07T17:56:24.597Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LmZcAeaW0VtsUoNE7Yr_6w.png" /><figcaption>Source: Midjourney V6</figcaption></figure><p>In 2012, my family and I took a Euro trip that blew my mind! At the time, I was working as a project manager at Watea, an IT consulting company, while studying IT Engineering at <a href="https://www.linkedin.com/company/utn--regional-ba/">UTN Buenos Aires</a>, one of the most prestigious universities in Buenos Aires. Having just passed one of the most challenging subjects, Operating Systems, I decided to treat myself by extending my family trip for 3 weeks. Me alone in Berlin.</p><p>After the trip, the first thing I did when I returned to Buenos Aires was to request an appointment with my favorite professor at UTN. I asked her how to start a business in Europe. Her response? “Finish your 6.5-year degree first.” Not exactly what I wanted to hear.</p><p>So I took matters into my own hands. I did what any sensible 23-year-old would do: printed 300 business cards, and went back to Berlin. After the first weekend, I had only 150 cards left. I remember, on Monday morning I opened LinkedIn and saw 50+ new connections, along with an invitation to an office opening in Kreuzberg. My best friend and I went to it. And I already knew that at every party, the action is always in the kitchen, so I stayed there and networked. At some point, someone (<a href="https://www.linkedin.com/in/christophraethke/">Christoph Raethke</a>) came to grab a beer, but there was no opener around, so I offered to open it myself. I took the bottle and I confidently opened it with my teeth — my little party trick. -Please, dentists, don’t judge. Afterward, I asked, “Do you work here?” He replied, “Well… technically, I don’t work here, but everyone here works for me.”</p><p>It turned out that Christopher was the Angel investor of this startup having the opening and Founder of <a href="https://www.linkedin.com/company/berlin-startup-academy/">Berlin Startup Academy</a>. After much persistence, he allowed my best friend and me to join him at the academy the following day. Footnote: The word “Startup” wasn’t even a thing in Argentina back then. After the first class, which was closed by a talk from <a href="https://www.linkedin.com/in/gerodecker/">Dr. Gero Decker</a> Decker from <a href="https://www.linkedin.com/company/signavio/">SAP Signavio</a>, I was blown away again.</p><p>Fast-forward three months: summer was over, and I had no business cards left. However, I had my own business called “We Store Solutions” — a web/software development agency hiring IT talent from my university in Buenos Aires and serving a European client base. I primarily handled sales and project management, earning more than double my previous salary and working remotely.</p><p>At that moment, I realized that sometimes the best education comes from venturing off the common path and challenging the status quo. And that it’s very important to have the courage to open unexpected doors — or in my case, bottles.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5c970deca645" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[My First Startup: If They Can Do It, So Can We]]></title>
            <link>https://medium.com/bigger-picture/my-first-startup-if-they-can-do-it-so-can-we-c003cb7b1b2e?source=rss-5f0bca136a2c------2</link>
            <guid isPermaLink="false">https://medium.com/p/c003cb7b1b2e</guid>
            <dc:creator><![CDATA[Luciano Radicce]]></dc:creator>
            <pubDate>Mon, 19 Aug 2024 14:10:46 GMT</pubDate>
            <atom:updated>2024-08-19T14:10:46.782Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*X_zfC6kadU4OIfTYFePIIw.jpeg" /><figcaption>unsplash.com</figcaption></figure><p>When I was 10 years old, living in Buenos Aires, my favorite place was a comic book store called DGL. This wasn’t just any store — it was a hub of excitement, selling all kinds of games. You name it, they had it: video games, card games like Magic The Gathering, anime miniatures, and more. It was geek heaven.</p><p>What made DGL truly special was that you could find 8-year-old kids playing alongside 35-year-old national champions of Mario Kart. Sometimes, these tournaments would take place at DGL. My brother, his best friend Jorge, and I would go there, not just to play, but to observe and understand how everything worked.</p><p>In Argentina, money was always scarce, and we learned not to burden our parents. So while thinking about how to earn some money, we thought about DGL’s tournaments, If they can do it, so can we. So we decided to organize a tournament of “Winning Eleven Pro Evolution Soccer,” a PlayStation game. Our competitive advantage? We had access to a whole primary school’s worth of potential players.</p><p>My father had an appliances business, so we borrowed 3–4 big TV screens and some extra PlayStations. Our house transformed into a mini-tournament venue. We set up different stages: one in the kitchen, one in my and my brother’s bedroom, and the main stage in the living room, where the crown jewel was set, a 37-inch semi-flat screen TV, which was probably as heavy as me at that time.</p><p>We charged 5 pesos per person, with a complimentary orange juice. The prizes were: 1st place got 3 PlayStation games, 2nd place got 2, and 3rd place received 1. It was like offering gold, silver, and bronze medals, but to a 10-year-old.</p><p>The tournament was a huge success. We were so busy that we had to outsource the orange juice serving to my grandma. After the tournament, we three founders stayed behind to count our earnings (stored in a shoebox). We each made 20 pesos/euros after expenses! We were filthy rich.</p><p>That summer, we organized one or two more tournaments. But as more people got big TVs and PlayStations at home, the interest wore off. It was my first taste of entrepreneurship, and it tasted sweeter than any orange juice we served.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c003cb7b1b2e" width="1" height="1" alt=""><hr><p><a href="https://medium.com/bigger-picture/my-first-startup-if-they-can-do-it-so-can-we-c003cb7b1b2e">My First Startup: If They Can Do It, So Can We</a> was originally published in <a href="https://medium.com/bigger-picture">The Bigger Picture</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[AI Product Review: Jasper — Content Creation Without the Hassle of Prompt Engineering]]></title>
            <link>https://ai.plainenglish.io/ai-product-review-jasper-content-creation-without-the-hassle-of-prompt-engineering-05e7c7eb8dae?source=rss-5f0bca136a2c------2</link>
            <guid isPermaLink="false">https://medium.com/p/05e7c7eb8dae</guid>
            <category><![CDATA[chatgpt]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[prompt-engineering]]></category>
            <dc:creator><![CDATA[Luciano Radicce]]></dc:creator>
            <pubDate>Tue, 09 Apr 2024 13:43:46 GMT</pubDate>
            <atom:updated>2024-04-12T07:35:07.025Z</atom:updated>
            <content:encoded><![CDATA[<h3>AI Product Review: Jasper — Content Creation Without the Hassle of Prompt Engineering</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*77LeC4HOxCWRmv1wQSwpvA.jpeg" /></figure><h3>Prelude</h3><p>Have you ever thought that Chat-GPT is great but hard to scale across an organization, prompt engineering is complex, and its results are impersonal?</p><p>Jasper is an end-to-end copilot that helps marketing teams achieve both speed &amp; performance. As their website states: “Not only faster, but better.” Long story short, it’s a multimodal content creation tool for enterprises.</p><p>I used Jasper for some months in 2020, three years before the Chat-GPT era. Given the fact not only that they have survived the “AI Revolution” but they are still one of the most popular AI consumer products, it’s an indication that they doing something correctly. So let’s find that out!</p><h3>Intro</h3><p>Jasper was founded in January 2020 by Dave Rogenmoser and Chris Hull. Both founders had founded together <a href="https://useproof.com/">Proof</a>, backed by YCombinator, batch W18. Although only until 2021 when they moved away from it, to start Jasper.</p><p>In September 2023, Dave became chairman, leaving the CEO crown to Timothy Young (Ex-President @Dropbox, Ex-VP @VMware) who has been involved in the company since its inception.</p><p>Jasper was one of the early pioneers recognizing the potential of generative AI for content creation. Their initial growth rate before the Chat-GPT era was record-setting for an early-stage SaaS, both in ARR and paying customer count. In October 2022, they raised through Series A $125 million at a $1.5 billion valuation 🔥.</p><p>The rise of Chat GPT has of course affected Jasper AI’s market share. In 2022, Jasper had announced 100,000 paying customers, however by October 2023, only 70,000. Chat GPT offers a free plan and a $20 paid version. As a result, Jasper AI was forced to introduce a limited $20 option in response to losing market share to Chat GPT at the start of 2023. However, it’s not available anymore.</p><p>It seems that Jasper had a rough 2023. Current active users are not publicly available but VP Sarah McConnell seems to see a brighter future ahead.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/552/1*NWJMAbOxMfoG6RTFwIaJTg.png" /><figcaption>Linkedin Post Sarah McConnel January 2024</figcaption></figure><p>Let’s dive into the product analysis, to make our analysis of the product, and predict whether it will last in the market.</p><h3>Problem Space</h3><p>Writing good content isn’t easy. It demands considerable time, strong language skills, and the right inspiration. Generative AI has completely changed the game, simplifying the process in many ways. However, using AI tools such as ChatGPT-4 or Gemini can be tricky, and mastering prompt engineering involves a learning curve.</p><p>Moreover, the content it produces often sounds too generic, brand-agnostic, or fails to connect with human emotions. It frequently resorts to using grandiose terms like ‘revolutionizing,’ ‘unleashing,’ and ‘leveraging.’ Tailoring these AI models to match your company’s style and tone is not straightforward. Furthermore, sometimes these AI models can be somewhat lazy or disobedient, which can be quite frustrating.</p><h3>Solution Space</h3><p>If you’re a large enterprise looking to explore Generative AI with a tool that offers low investment and a minimal learning curve, while still ensuring fast and high-quality content for your company, Jasper is your go-to solution. It’s the perfect choice for creating content that’s infused with company knowledge, hassle-free, and tailored for specific use cases, yet ensures enhanced privacy, security, and consistency. With Jasper, the need for prompt engineering is virtually eliminated, democratizing writing for everyone, including those who may not be fully fluent or accurate in a foreign language (like me 🤓).</p><h3>USPs &amp; Strategic Positioning</h3><ul><li><strong>Data Privacy:</strong> Ensures data integrity and confidentiality, as no customer data is used for training models.</li><li><strong>Performance:</strong> Delivers fast responses optimising LLM querying.</li><li><strong>Model Interoperability:</strong> Seamlessly integrates various AI models, transparent to the user.</li><li><strong>Multimodality:</strong> Offers support for both text and image creation, enabling comprehensive creative expression. Note: Image creation is available post-payment.</li><li><strong>Security Standards:</strong> Adheres to GDPR, SOC2, and SSO standards to safeguard data.</li><li><strong>Content Customisation:</strong> Facilitates easy alignment of AI-generated content with your brand’s voice and style.</li><li><strong>Workflow Integration:</strong> Seamlessly enhances team collaboration and integrates into existing content creation workflows, significantly boosting productivity.</li></ul><h3>Technology &amp; Features</h3><p>Jasper AI merges its proprietary AI engine with third-party models to address a wide array of needs. Depending on the specific use case, it intelligently routes requests to the most appropriate AI model. This innovative strategy employs both Jasper’s custom models and those developed by leading entities like OpenAI, Google, or Anthropic. This distinctive approach, in my perspective, uniquely positions Jasper in the market, as no other tool currently offers such a bespoke AI model selection process.</p><h3>Key Features:</h3><ul><li><strong>Templates</strong>: Offering +50 templates and SEO tool integration, facilitating the creation of optimized content for enhanced online visibility — avoiding the need for prompt engineering. This feature is both cool and practical.</li><li><strong>Your Content:</strong> The content library is a more practical approach to AI than conversation tabs, used by Chat-GPT, which tend to get messy.</li><li><strong>Brand Voice</strong>: Allows users to centralize various brand tones, enabling the adjustment of writing styles and tones for consistent brand representation across all templates. This feature can be utilized company-wide, ensuring a unified voice in all content.</li><li><strong>Company Knowledge Hub</strong>: Enables users to upload company documents, ensuring that content aligns with the company’s know-how.</li><li><strong>Campaign:</strong> Integrates templates, brand voice, and company knowledge to efficiently create all marketing content pieces in one go. While the output’s quality may not be amazing, it significantly increases efficiency compared to manual processes. It’s ideal for drafting.</li><li><strong>Multimodality</strong>: Allows users to generate text and images, broadening its applicability for various content creation needs.</li><li><strong>Analytics and Collaboration</strong>: Offers performance analytics and project management tools designed to optimize content strategy and enhance teamwork.</li><li><strong>Support</strong>: Provides fast and comprehensive customer support, including email, chat, and phone options. Responses may initially come from templates, but helpful human assistance is available.</li></ul><h3>UX &amp; UI</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DLpmFpLj-UuC1Kun3SQ1EQ.jpeg" /></figure><p>Here is a Figma <a href="https://www.figma.com/file/74ULB7RF4zQTD7ClaxJhp0/Perplexity.AI---Mockups-%26-Wireframes?type=design&amp;node-id=1%3A3&amp;mode=design&amp;t=oJDguDdP3XBn3ozg-1%20Jasper.AI">link</a> to the screens.</p><ul><li><strong>Performance</strong>: This is one of Jasper’s strongest arms, outperforming ChatGPT and Gemini significantly. I would say it’s twice as fast as Chat-GPT &amp; Gemini.</li><li><strong>User Interface</strong>: Currently, it’s only available as a desktop version, which aligns with the primary mode of text generation tasks typically conducted on desktops.</li><li><strong>Branding &amp; Interface Design</strong>: User interface design is top-notch, setting a high standard in the industry.</li><li><strong>Bugs</strong>: Some persistent issues exist, especially with text selection and generation. For example, improving pre-written text initially posed challenges. This aspect requires further refinement.</li><li><strong>Paywall</strong>: It’s annoying that many pro features are behind the paywall and you don’t get to test them when doing a trial. Competitors like Perplexity and Grammarly have faced similar criticisms but now offer limited pro usage daily even in their free versions to mitigate this issue.</li></ul><h3>Jasper Use Cases</h3><ul><li><strong>Blog Post Writing:</strong> This feature may not be exceptionally standout, but it offers a solid entry point into AI for those new to prompt engineering. It’s excellent for creating SEO-friendly articles, leveraging the company’s centralized brand voice.</li><li><strong>Social Media</strong>: Fast and easy content creation such as SEO-optimized articles that boost traffic or timely social media posts. It’s a versatile tool for marketers and social media managers, which enables scalable content creation efforts with efficiency.</li><li><strong>Email Campaigns</strong>: Using AI-driven insights, Jasper is useful in crafting personalized email campaigns that engage and resonate with recipients, improving engagement rates.</li><li><strong>E-commerce Copy</strong>: Create product titles, product descriptions, product info &amp; more. A cool feature is its integration with Zapier, allowing for automation and seamless updates to Shopify or WordPress shops.</li><li><strong>Campaign</strong>: Although it’s not perfect, it’s just so fast &amp; easy to create many consistent assets simultaneously. It serves as an excellent foundation for further refinement and recreation of marketing materials.</li></ul><h3>Pros and Cons</h3><h4>Pros:</h4><ul><li><strong>Fast Content Generation:</strong> Creating high-quality content quickly, ideal for “light” SEO-friendly content pieces, perfect for maintaining an active online presence.</li><li><strong>Advanced AI Tech</strong>: Its state-of-the-art AI technology enhances the writing process, distinguishing it from competitors like ChatGPT.</li><li><strong>Broad Applications</strong>: Capable of producing a wide array of content types through templates, making it a versatile asset for varied marketing initiatives.</li><li><strong>SEO Tools:</strong> In-built SEO optimization helps improve online visibility, giving an edge over traditional LLMs.</li><li><strong>User-Friendly</strong>: The platform’s intuitive interface ensures a smooth experience for all users, empowering them to focus on creativity.</li><li><strong>Excellent Support</strong>: Offers responsive and personalized customer support, ensuring users receive timely help when needed.</li><li><strong>Multi-Model Integration:</strong> Employing a distinctive ‘K-models / LLMS’ approach, <a href="http://jasper.ai/">Jasper.ai</a> leverages multiple AI models concurrently to deliver the best output. This strategy enhances performance by selecting the most appropriate result from various models, making the tool more efficient and precise for users.</li></ul><h4>Cons &amp; Main Issues:</h4><ul><li><strong>Learning Curve</strong>: Despite its intuitive design, there can be a learning curve for new users unfamiliar with AI-driven content creation tools. Understanding how to best utilize Jasper’s features for optimal results may require a bit of time and experimentation. Lower than a plain LLM, though.</li><li><strong>Cost</strong>: For small businesses or individual users, the cost can be a significant factor. Its cost may not align with every budget, despite the value it provides.</li><li><strong>Overreliance on Templates</strong>: It offers a wide range of templates, there’s a potential risk of overreliance which could lead to less unique and differentiated content if not customized sufficiently.</li><li><strong>AI Nuances</strong>: The AI-generated content, while impressive, may occasionally miss the nuances or tone intended by the user, necessitating manual revisions for precision and personal touch.</li><li><strong>Content Organization:</strong> It is not possible to tag or create folders.</li><li><strong>Disruptability</strong>: Given the direction &amp; speed of AI developments, it has a high chance of being disrupted. ****</li></ul><h3>Pricing and Plans</h3><p><a href="http://jasper.ai/">Jasper.ai</a> offers a 7-day free trial, followed by two subscription tiers:</p><h4>Billed Yearly:</h4><ul><li><strong>Pro:</strong> $59/month</li><li><strong>Creator:</strong> $39/month</li></ul><h4>Billed Monthly:</h4><ul><li><strong>Pro:</strong> $69/month</li><li><strong>Creator:</strong> $49/month</li></ul><h3>User Feedback and Reviews</h3><h4>Positive Feedback</h4><ul><li><strong>Fast Content Generation:</strong> Jasper is praised for its ability to create content much faster than human copywriters, producing 100% original, plagiarism-free content. This makes it ideal for generating SEO content aimed at achieving higher SERP rankings and supports over 30 languages (<a href="https://www.techopedia.com/ai/jasper-ai-review">Techopedia</a>).</li><li><strong>User-Friendly Interface:</strong> Users find the Jasper interface simple and straightforward, akin to using Google Docs or Microsoft Word, making the transition to using Jasper easy for new users (<a href="https://www.bitcatcha.com/research/jasper-ai-review/">Bitcatcha</a>).</li><li><strong>Broad Applications:</strong> The platform is noted for its versatility in creating various types of content, including blog posts, long-form content, SEO-based content, social media posts, and email newsletters. Jasper offers over 50 different templates to assist in high-caliber writing across multiple industries (<a href="https://www.techopedia.com/ai/jasper-ai-review">Techopedia</a>).</li></ul><h4>Constructive Criticisms</h4><ul><li><strong>Technical Content Challenges:</strong> Jasper struggles with overly technical topics, producing generic or unusable content in some instances. This limitation necessitates significant user intervention to ensure accuracy and relevance (<a href="https://www.bitcatcha.com/research/jasper-ai-review/">Bitcatcha</a>).</li><li><strong>Content Repetition and Junk:</strong> Users have reported issues with the repetition of words and phrases. Additionally, generating large volumes of content quickly can result in unusable “junk” content, leading to wasted resources and the need for thorough editing (<a href="https://www.techopedia.com/ai/jasper-ai-review">Techopedia</a>) (<a href="https://www.bitcatcha.com/research/jasper-ai-review/">Bitcatcha</a>).</li><li><strong>Additional Costs:</strong> Some users have pointed out the extra costs associated with using Jasper’s plagiarism detection feature, powered by Copyscape, which is not included in the basic plan and charges by volume (<a href="https://www.bitcatcha.com/research/jasper-ai-review/">Bitcatcha</a>).</li></ul><h3>Final Review: <a href="http://jasper.ai/">Jasper.AI</a></h3><p><strong>Problem Solving</strong>: 4 stars. The combination of LLM search and search engines, along with the capability to view references, is highly satisfying. Mucho gusto! 👌</p><p><strong>Features &amp; Capabilities</strong>: 3.5 stars. Considering <a href="http://jasper.ai/">Jasper.AI</a> has been on the market for four years and faces competition with other LLMs regarding technology and pricing, higher expectations in this area are warranted.</p><p><strong>Performance</strong>: 5 stars. In terms of speed, <a href="http://jasper.ai/">Jasper.AI</a> outpaces any other LLM, and its output quality matches that of an off-the-shelf LLM.</p><p><strong>UI/UX</strong>: 4 stars. While generally effective, there’s potential for improvement in the content creation dynamics.</p><p><strong>Disruptability</strong>: 3 stars. Viewing through a PEST analysis lens, given the rapid pace of technological advancements, it’s likely that <a href="http://jasper.ai/">Jasper.AI</a> will face disruption within the next five years.</p><p><strong>Final Review</strong>: 3.7 stars!</p><p>Thanks for reading folks! Do you have questions or feedback? Feel free to shoot me an email or DM me on any of my social networks. Enjoy exploring!</p><p><strong><em>Luciano Radicce</em></strong><em> is a seasoned entrepreneur, strategist, and founder of </em><a href="http://lazyconsulting.com/"><em>Lazy Consulting</em></a><em>, specializing in AI product strategy. With a passion for discussing relevant tech topics that shape the world, </em><strong><em>engaging in thought-provoking</em></strong><em> conversations that push the boundaries of </em><strong><em>ethical innovation</em></strong><em>.</em></p><p><em>Support this work by </em><strong><em>sharing</em></strong><em> this article, </em><strong><em>liking</em></strong><em> it, or </em><strong><em>commenting</em></strong><em> on it ❤.</em></p><p><em>Originally published at </em><a href="https://www.lazyconsulting.com/post/ai-product-review-perplexity-ai----bridging-the-gap-between-ai-chats-search-engines"><em>https://www.lazyconsulting.com</em></a><em>.</em></p><h3>In Plain English 🚀</h3><p><em>Thank you for being a part of the </em><a href="https://plainenglish.io/"><strong><em>In Plain English</em></strong></a><em> community! Before you go:</em></p><ul><li>Be sure to <strong>clap</strong> and <strong>follow</strong> the writer ️👏<strong>️️</strong></li><li>Follow us: <a href="https://twitter.com/inPlainEngHQ"><strong>X</strong></a><strong> | </strong><a href="https://www.linkedin.com/company/inplainenglish/"><strong>LinkedIn</strong></a><strong> | </strong><a href="https://www.youtube.com/channel/UCtipWUghju290NWcn8jhyAw"><strong>YouTube</strong></a><strong> | </strong><a href="https://discord.gg/in-plain-english-709094664682340443"><strong>Discord</strong></a><strong> | </strong><a href="https://newsletter.plainenglish.io/"><strong>Newsletter</strong></a></li><li>Visit our other platforms: <a href="https://stackademic.com/"><strong>Stackademic</strong></a><strong> | </strong><a href="https://cofeed.app/"><strong>CoFeed</strong></a><strong> | </strong><a href="https://venturemagazine.net/"><strong>Venture</strong></a><strong> | </strong><a href="https://blog.cubed.run/"><strong>Cubed</strong></a></li><li>More content at <a href="https://plainenglish.io/"><strong>PlainEnglish.io</strong></a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=05e7c7eb8dae" width="1" height="1" alt=""><hr><p><a href="https://ai.plainenglish.io/ai-product-review-jasper-content-creation-without-the-hassle-of-prompt-engineering-05e7c7eb8dae">AI Product Review: Jasper — Content Creation Without the Hassle of Prompt Engineering</a> was originally published in <a href="https://ai.plainenglish.io/">Artificial Intelligence in Plain English</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[AI product review: Perplexity.ai — Bridging the gap between AI chats & search engines]]></title>
            <link>https://ai.plainenglish.io/ai-product-review-perplexity-ai-bridging-the-gap-between-ai-chats-search-engines-bef02453426d?source=rss-5f0bca136a2c------2</link>
            <guid isPermaLink="false">https://medium.com/p/bef02453426d</guid>
            <category><![CDATA[perplexity-ai]]></category>
            <category><![CDATA[generative-ai-tools]]></category>
            <category><![CDATA[ai-tools]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[chatgpt]]></category>
            <dc:creator><![CDATA[Luciano Radicce]]></dc:creator>
            <pubDate>Tue, 06 Feb 2024 05:17:22 GMT</pubDate>
            <atom:updated>2024-02-06T06:29:58.487Z</atom:updated>
            <content:encoded><![CDATA[<h3>AI product review: Perplexity.ai — Bridging the gap between AI chats &amp; search engines</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*IOjNZiyCTBzVcn5jGKisCA.png" /></figure><blockquote><em>You’ve got to start with the customer experience and work back toward the technology — not the other way around.” </em>Steve Jobs‍</blockquote><p>Have you ever searched for a fact on Google, only to be frustrated by a bunch of ads and irrelevant links? You could use Chat-GPT with Bing or a web-browsing plugin, but it may be a complete hallucination or fabrication. Well, this is precisely the gap that <a href="http://perplexity.ai/">perplexity.ai</a> aims to fill.</p><p>Perplexity is standing out in the AI community, offering a unique blend of search and AI-driven insights. But how does it stand against giants like Chat-GPT and Bard?</p><p>Short answer: by trying to solve a unique problem. And we, product people, love that.</p><p>For these reasons, <a href="http://perplexity.ai/">http://perplexity.ai</a> will be the first product in my series reviewing AI products.</p><p>‍</p><h3>Intro</h3><p><a href="http://perplexity.ai/">Perplexity.ai</a> is a search platform that utilizes Large Language Models (LLMs) to address complex queries by summarizing content from multiple pages, providing fast, accurate answers with references to their sources, thus ensuring no hallucinations or fake news — a point that really impressed me. The free version, which requires no sign-up to start querying 🫶, works with their default model, while the paid version, starting at 20 USD per month, includes access to GPT-4, Claude 2.1, and Gemini.</p><p>The company was co-founded in 2022 and is headquartered in San Francisco, California. The founders include Andy Konwinski (formerly of Databricks), Aravind Srinvas (formerly of OpenAI, DeepMind, and Google), CTO Denis Yarats (formerly of Facebook, Quora, and Microsoft), and Carlos Virella. Perplexity.ai is backed by venture capital firm NEA, Databricks, and individual investors who have stakes in companies like Weaviate and Wispr AI, among others. Its most recent funding round was a Series B.</p><p>‍</p><h3>Problem vs solution space</h3><h4>Problem space</h4><p>As per their slogan “Where Knowledge Begins,” <a href="http://perplexity.ai/">Perplexity.ai</a> seems to be addressing the problem of discerning the truth when searching on search engines. It is often hard to know whether you are reading a complete fabrication or not. Articles don’t necessarily need to have sources, and when reading about a subject that the reader is unfamiliar with, discerning the truth can be difficult. A similar problem occurs when querying an LLM for facts.</p><h4>Solution space</h4><p>Perplexity seeks to solve the aforementioned problem by providing a fast and user-friendly interface, both on the web and as a mobile app. When a user asks a question, it uses LLMs’ internal knowledge to double-check the facts stated and provides an answer in text form, adding links from the sources as well. The sources are chosen based on relevance, domain authority, recency, user feedback, and quality of content.</p><p>To deliver the most relevant information, Perplexity may engage in a back-and-forth conversation with the user, asking questions to clarify their needs. Perplexity achieves this with the help of an advanced answer engine to process your questions and tasks, taking into account the entire conversation history for context.</p><p>‍</p><h3>Strategic positioning &amp; USP</h3><p><a href="http://perplexity.ai/">Perplexity.ai</a> aims to occupy the sweet spot between a web-browsing LLM and a search engine. Their entire technology, user interface, and experience are designed to provide this hybrid solution in the best way possible.</p><p>‍</p><h3>Target Audience</h3><p>The target audience for <a href="http://perplexity.ai/">Perplexity.ai</a> encompasses a broad spectrum of users who require quick, accurate, and reliable information, often with an emphasis on the credibility of sources. This audience can be broadly categorised into several key segments:</p><ol><li><strong>Academic and Research Community</strong>: Students, researchers, and educators seeking scholarly content and academic resources.</li><li><strong>Professionals and Business Users</strong>: Entrepreneurs, analysts, and other professionals needing up-to-date market insights and industry trends for informed decision-making.</li><li><strong>Tech Enthusiasts and Early Adopters</strong>: Individuals interested in the latest AI and technology advancements, exploring new tools for personal or professional use.</li><li><strong>Casual Information Seekers</strong>: General users looking for quick, accurate answers to a wide range of questions, from simple facts to complex topics.</li><li><strong>Journalists and Content Creators</strong>: Those who need to rapidly gather, verify, and summarize information from multiple web sources for their content.</li><li><strong>Developers and Tech Companies</strong>: Users leveraging Perplexity.ai’s API for integrating advanced search and AI capabilities into their applications.</li></ol><p>‍</p><h3>Features and capabilities:</h3><h4>Features</h4><ol><li><strong>Main search:</strong> This feature allows users toaggregate and summarize relevant information from diverse web sources, providing users with credible, fact-checked answers to their queries</li><li><strong>Copilot</strong>: Allows to interact with the AI assistant, enabling a conversational approach to browsing the web and accessing information. However, in my opinion, this is a bit of a weak spot and needs more polishing.</li><li><strong>Focus</strong>: This is a filter that allows users to search within possible sources, including Academic, Writing, YouTube, Reddit, Wolfram/Alpha, or All.</li><li><strong>Discover</strong>: The most relevant and famous threads are listed for everyone to access.</li><li><strong>Library</strong>: This is where users can view their past threads.</li><li><strong>Playground</strong>: The <a href="https://labs.perplexity.ai/">Perplexity Labs</a> is a hidden feature that allows users to choose different models and versions of these, such as pplx-7b-online, pplx-70b-online, mistal-7b-instruct, codellama-34b-instruct, llama2–70b-chat, and more. Pretty cool!</li><li><strong>Separate Chatbot</strong>: Perplexity offers a chatbot feature, though its functionality seems similar to the main view.</li></ol><h4>Capabilities</h4><ol><li><strong>Math Problem Solving</strong>: While it can solve math problems, the formatting might not be as polished as that of ChatGPT.</li><li><strong>Real-time Information</strong>: Unlike ChatGPT, which has data limited to September 2021, Perplexity provides real-time information with relevant links or videos.</li><li><strong>Search Engine Extension</strong>: Perplexity can search the web, similar to ChatGPT’s Bing integration.</li><li><strong>API</strong>: It also provides an API that allows users to request data from other systems.</li><li><strong>Attachments</strong>: Users are allowed to attach text, PDFs, and images, but only 3 per day with the free version.</li></ol><p>‍</p><h3>User Experience‍</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*qvpVgDuHz90iYtZY.jpg" /></figure><p>Here is a Figma <a href="https://www.figma.com/file/74ULB7RF4zQTD7ClaxJhp0/Perplexity.AI---Mockups-%26-Wireframes?type=design&amp;node-id=1%3A3&amp;mode=design&amp;t=oJDguDdP3XBn3ozg-1">link</a> to the screens</p><ul><li><strong>Performance</strong>: It is known for its speed, often delivering results faster than some competitors.</li><li><strong>User Interface</strong>: There is both a web version and an app version available.</li><li><strong>Branding &amp; Design</strong>: They released a new branding and interface design in September 2023.</li><li><strong>User Experience:</strong>‍</li><li>It flows very quickly, although the copilot &amp; threads need more re-thinking. It lacks a clear.</li></ul><h3>Main 3 Issues:</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*zyc6vUbccxkuExrb.jpg" /></figure><p><strong>1)Relevance of Search Results</strong></p><ul><li><strong>Issue</strong>: The search results do not always align precisely with the user’s questions or provide the desired information.</li><li><strong>Potential Solution</strong>: Enhance the AI’s understanding of context and user intent to improve the relevance of search results. Incorporate user feedback mechanisms for continuous learning and accuracy improvement.</li></ul><p><strong>2)Management of Query Threads</strong></p><ul><li><strong>Issue</strong>: Users are dissatisfied with the constant display of past query threads, which can clutter their experience.</li><li><strong>Potential Solution</strong>: Introduce features to manage and organize past queries, such as the ability to archive or delete threads, or to opt out of this feature entirely.</li></ul><p><strong>3)Limited Free Version Functionality</strong></p><ul><li><strong>Issue</strong>: The free version may have restrictions that limit the user’s ability to thoroughly test the platform’s capabilities, leading to a potentially incomplete evaluation of its effectiveness.</li><li><strong>Potential Solution</strong>: Offer a trial period of the premium version, providing full access to all features for a limited time, allowing users to experience the full breadth of the platform’s capabilities before deciding to subscribe.</li></ul><h3>Comparison with LLMs (ChatGPT, Bard)</h3><ul><li>UX: While ChatGPT is recognized for its superior language model and human-like conversations, <a href="http://perplexity.ai/">Perplexity.ai</a> is appreciated for its ease of use &amp; fact-focused answers.</li><li>Capabilities: I would not compare Perplexity with ChatGPT. Even with the web-pilot capabilities, ChatGPT is great at generating answers but not as effective for research purposes.</li><li>Tech:<a href="http://perplexity.ai/"> Perplexity.ai</a> uses its own large language models and provides GPT-4 access, similar to many other LLMs.</li><li>Price: The free version is available, and the premium version is priced the same as ChatGPT, at 20 USD.</li></ul><h3>Quotes from AI Leaders</h3><blockquote>“Perplexity is the only AI product that I use every day on both my desktop and phone. It replaces the research I used to conduct either on ChatGPT or Google. Especially for product discovery or sales research, there is no better tool for quickly searching live data. Perplexity also offers access to their models, and we have integrated those into the Klu platform for use cases that require real-time search data” <strong>Stephen Walker II, CEO of Klu</strong></blockquote><blockquote>“ I’ve found <a href="http://perplexity.ai/">Perplexity.ai</a> to be an invaluable tool for conducting scientific research, offering access to the latest studies and data with remarkable accuracy. Its advanced AI capabilities significantly streamline the research process, enabling us to stay at the forefront of healthcare innovation.” <strong>Adrian Kochsiek, CEO of Onvy Health.</strong></blockquote><h3>Final Star Review</h3><ul><li><strong>Problem Solving</strong>: 5 stars. A mix of LLM search and search engines with the ability to see references is highly satisfying. Mucho gusto! 👌</li><li><strong>Features &amp; Capabilities</strong>: 4 stars. The copilot needs more work, and the discovery feature seems irrelevant to me.</li><li><strong>Performance</strong>: 5 stars. Many AI products are in the beta stage with poor performance, but not Perplexity.ai. It’s fast and reliable.</li><li><strong>UI/UX</strong>: 4 stars. The copilot sometimes gives answers that seem to be just copy-pasted; I would appreciate more elaboration, not just a summary.</li></ul><p><strong>Final</strong> <strong>Rating:</strong> 4.5/5 Lucianos 👨🏻</p><h3>Closing thoughts &amp; conclusion</h3><p><a href="http://perplexity.ai/">Perplexity.ai</a> brings something new to the table in the world of AI and information gathering. Its no-cost access and ability to pull up-to-date information quickly make it a go-to for fast facts. It may not beat ChatGPT in conversation skills, but its special features make it stand out.</p><p>I’m all for trying out great products, and I suggest you do the same with <a href="http://perplexity.ai/">Perplexity.ai</a>. It’s particularly useful for digging into market trends, the latest news, or historical data. Give it a go — it might just change the way you look up information and become the place where <em>your</em> knowledge truly begins.‍</p><p><strong><em>Luciano Radicce</em></strong><em> is a seasoned entrepreneur, strategist, and founder of </em><a href="http://lazyconsulting.com/"><em>Lazy Consulting</em></a><em>, specializing in AI product strategy. With a passion for discussing relevant tech topics that shape the world, </em><strong><em>engaging in thought-provoking</em></strong><em> conversations that push the boundaries of </em><strong><em>ethical innovation</em></strong><em>.</em></p><p><em>Support this work by </em><strong><em>sharing</em></strong><em> this article, </em><strong><em>liking</em></strong><em> it, or </em><strong><em>commenting</em></strong><em> on it ❤.</em></p><p><em>Originally published at </em><a href="https://www.lazyconsulting.com/post/ai-product-review-perplexity-ai----bridging-the-gap-between-ai-chats-search-engines"><em>https://www.lazyconsulting.com</em></a><em>.</em></p><h3>In Plain English 🚀</h3><p><em>Thank you for being a part of the </em><a href="https://plainenglish.io/"><strong><em>In Plain English</em></strong></a><em> community! Before you go:</em></p><ul><li>Be sure to <strong>clap</strong> and <strong>follow</strong> the writer ️👏<strong>️️</strong></li><li>Follow us: <a href="https://twitter.com/inPlainEngHQ"><strong>X</strong></a><strong> | </strong><a href="https://www.linkedin.com/company/inplainenglish/"><strong>LinkedIn</strong></a><strong> | </strong><a href="https://www.youtube.com/channel/UCtipWUghju290NWcn8jhyAw"><strong>YouTube</strong></a><strong> | </strong><a href="https://discord.gg/in-plain-english-709094664682340443"><strong>Discord</strong></a><strong> | </strong><a href="https://newsletter.plainenglish.io/"><strong>Newsletter</strong></a></li><li>Visit our other platforms: <a href="https://stackademic.com/"><strong>Stackademic</strong></a><strong> | </strong><a href="https://cofeed.app/"><strong>CoFeed</strong></a><strong> | </strong><a href="https://venturemagazine.net/"><strong>Venture</strong></a></li><li>More content at <a href="https://plainenglish.io/"><strong>PlainEnglish.io</strong></a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=bef02453426d" width="1" height="1" alt=""><hr><p><a href="https://ai.plainenglish.io/ai-product-review-perplexity-ai-bridging-the-gap-between-ai-chats-search-engines-bef02453426d">AI product review: Perplexity.ai — Bridging the gap between AI chats &amp; search engines</a> was originally published in <a href="https://ai.plainenglish.io/">Artificial Intelligence in Plain English</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why Do AI Products Fail?]]></title>
            <link>https://ai.plainenglish.io/why-do-ai-products-fail-622bf2a5df97?source=rss-5f0bca136a2c------2</link>
            <guid isPermaLink="false">https://medium.com/p/622bf2a5df97</guid>
            <category><![CDATA[product]]></category>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[design]]></category>
            <category><![CDATA[user-experience]]></category>
            <dc:creator><![CDATA[Luciano Radicce]]></dc:creator>
            <pubDate>Mon, 02 Oct 2023 07:16:29 GMT</pubDate>
            <atom:updated>2023-10-02T09:46:07.279Z</atom:updated>
            <content:encoded><![CDATA[<h4>The mistakes AI startups are making and how to solve them</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*7s62aiaDlX42n4Zd1iAuNQ.jpeg" /></figure><p>Here we are, kicking off the last quarter of the year, and it’s been half a year since the Chat GPT Era began. And honestly? Most AI products suck. I’ve not yet come across an AI product that truly wowed me. It’s like watching Thor’s hammer, being used to build crooked houses. Modern LLMs are such a potent technology, but product-wise? Nothing that’s knocked my socks off.</p><p>Even the mighty Chat-GPT; it’s great, but what is it for? Which problem is it solving? What about the UX?</p><p>Therefore, I’ve compiled a list of common errors to watch for when creating AI products, for engineers, entrepreneurs, product managers, and anyone who’s playing with AI no-code tools trying to create something useful &amp; valuable.</p><p>‍</p><h3>Common problems</h3><h4>Product-Market fit vs. Market-Product fit</h4><p>Do things that don’t scale (1). Instead of seeking customers for your tool, venture out and address real people’s challenges. Conduct user interviews. If you’re unfamiliar with the process, I suggest checking out the “Mom Test” book (2). But to sum it up: Pose questions and genuinely seek to understand. A good rule of thumb: If you’re speaking more than you’re listening, you’re likely not doing it right</p><p>‍</p><h3>Lack of Originality on the Value Propositions</h3><p>Many folks seem to disregard these rules from “The 22 Immutable Laws of Marketing” (3)</p><p><em>Law #1: Be the First in Your FieldLaw #2: If You Can’t Be First, Create a New Category</em></p><p>Merely having a presence in the market isn’t sufficient. It’s crucial to deeply understand the market and serve it with passion. Surprisingly, user research, an essential component, is frequently sidelined.‍</p><h3>Poor user experience</h3><p>Even with technical products, functionality isn’t the sole focus. While adopting a “Lean mindset” is commendable, the goal shouldn’t be just an MVP (minimum viable product). Instead, in a highly competitive landscape, strive for an MLP (minimum lovable product)(4)</p><p>The illustration below provides a visual representation of this concept:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*8STy8UtyhVELpSj0.png" /></figure><p>‍</p><p>This perspective encompasses not only the user interface but also the performance of your products, including the performance of the AI models you work with — including when they hallucinate.</p><p>‍</p><h3>Unbalanced Founding Teams</h3><p>The buzz around AI appears to be eclipsing basic team-building principles. I’m seeing a surge in teams made up solely of developers or engineers, missing a comprehensive business and design perspective. The most effective teams embody the HHH principle — Hacker, Hipster, Hustler. In simpler terms, this translates to a balanced mix of Business/Product, Design, and Engineering.</p><p>‍</p><h3>Overpromising capabilities</h3><p>The mantra should be: underpromise and overdeliver. Be transparent about the current state of your product and resist the temptation to monetize prematurely. Consider your newborn product like a baby. You wouldn’t send a baby to work; instead, you nurture it, educate it, and ensure a safe environment for its healthy growth.</p><p>Clearly articulate what your tool can and will do, as well as its limitations. Stay away from monetizing too soon. Prioritize getting it into the hands of as many users or testers as possible and glean insights from their feedback.</p><p>‍</p><h3>Fine-tuning before achieving Product-Market Fit (PMF)</h3><p>Start with a broad approach. Initially, leverage powerful models to understand what your users want/need. Only after gaining this understanding should you transition to use-case-specific AI models or dive into fine-tuning. Jumping into the latter prematurely can be costly, especially when you haven’t fully comprehended your users’ needs.‍</p><h3>Overlooking legality, safety, privacy, or regulations</h3><p>We are all familiar with the saying, “Better safe than sorry.” While delving into legalities and regulations might not be the most thrilling task, it’s essential. Dedicate time to understanding the regulations; In Europe, for instance, there’s the AI Act. It may be broad and somewhat ambiguous, but expect more specific regulations to come soon. And let’s not forget our old friend: GDPR. Always communicate transparently on your website about the data you share with third-party tools. Be clear and upfront.‍</p><h3>Underestimating operational costs</h3><p>Powerful models can be costly, mainly when aiming for heavy user testing before monetizing. It’s crucial to crunch the numbers and understand your financial runway. Determine how long you can sustain operations before needing to monetize, especially if you’re bootstrapping.</p><h3>Conclusion</h3><p>Go back to the basics:</p><ul><li>Focus on Problem Space vs. Solution Space.</li><li>Prioritise Market-Product Fit over Product-Market Fit.</li><li>Be original**.** Find a unique niche and be creative with your value proposition.</li><li>Grow evenly. UX &amp; functionality.</li><li>Team up right**.** Mix Product, Design, and Engineering.</li><li>Promise less, deliver more.</li><li>Start general, then fine-tune.</li><li>Stay updated, know, and follow regulations.</li><li>Assess costs; plan long-term.</li><li>Enjoy the ride!</li></ul><p>‍</p><h3>Links &amp; Sources</h3><p>1- <a href="http://paulgraham.com/ds.html">http://paulgraham.com/ds.html</a> ‍<br>2- <a href="https://www.momtestbook.com/">https://www.momtestbook.com/</a><br>3- <a href="https://www.amazon.de/22-Immutable-Laws-Marketing-Explained/dp/0887306667">https://www.amazon.de/22-Immutable-Laws-Marketing-Explained/dp/0887306667</a><br>4- <a href="https://slowlettuce.io/blog/what-is-a-minimum-lovable-product">https://slowlettuce.io/blog/what-is-a-minimum-lovable-product</a><br>5- <a href="https://medium.com/@nikhilgupta08/problem-space-vs-solution-space-f970d4ace5c">https://medium.com/@nikhilgupta08/problem-space-vs-solution-space-f970d4ace5c</a></p><p>‍</p><p><em>Originally published at </em><a href="https://www.lazyconsulting.com/post/why-ai-products-fail"><em>https://www.lazyconsulting.com</em></a><em>.</em></p><h3>In Plain English</h3><p><em>Thank you for being a part of our community! Before you go:</em></p><ul><li><em>Be sure to </em><strong><em>clap</em></strong><em> and </em><strong><em>follow</em></strong><em> the writer! 👏</em></li><li><em>You can find even more content at </em><a href="https://plainenglish.io/"><strong><em>PlainEnglish.io</em></strong></a><strong><em> 🚀</em></strong></li><li><em>Sign up for our </em><a href="http://newsletter.plainenglish.io/"><strong><em>free weekly newsletter</em></strong></a><em>. 🗞️</em></li><li><em>Follow us on </em><a href="https://twitter.com/inPlainEngHQ"><strong><em>Twitter</em></strong></a><strong><em>(X</em></strong>), <a href="https://www.linkedin.com/company/inplainenglish/"><strong><em>LinkedIn</em></strong></a>, <a href="https://www.youtube.com/channel/UCtipWUghju290NWcn8jhyAw"><strong><em>YouTube</em></strong></a>, and <a href="https://discord.gg/XxRS92b2"><strong><em>Discord</em></strong></a><strong><em>.</em></strong></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=622bf2a5df97" width="1" height="1" alt=""><hr><p><a href="https://ai.plainenglish.io/why-do-ai-products-fail-622bf2a5df97">Why Do AI Products Fail?</a> was originally published in <a href="https://ai.plainenglish.io/">Artificial Intelligence in Plain English</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How Not to Screw Up Your First Steps in Your AI Journey]]></title>
            <link>https://ai.plainenglish.io/how-not-to-screw-up-your-first-steps-in-your-ai-journey-ec9e12e8ea38?source=rss-5f0bca136a2c------2</link>
            <guid isPermaLink="false">https://medium.com/p/ec9e12e8ea38</guid>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[digital-transformation]]></category>
            <category><![CDATA[change-management]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[entrepreneurship]]></category>
            <dc:creator><![CDATA[Luciano Radicce]]></dc:creator>
            <pubDate>Tue, 12 Sep 2023 04:17:09 GMT</pubDate>
            <atom:updated>2023-09-12T04:17:09.529Z</atom:updated>
            <content:encoded><![CDATA[<h4>A practical guide for business leaders considering AI integration focused on cutting through the hype and avoiding rookie mistakes</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1006/1*9Mju_qcxCg_aPX6ZhldVOw.jpeg" /></figure><h3>‍Intro: Embrace AI, don’t fight against it‍</h3><p>In the mid to late 1990s, as the internet began to gain traction, many were doubtful of its significance, but now it’s an integral part of our lives. Fast forward to the emergence of social media, there was a similar wave of doubt, yet today the global daily average time spent on social media is 2 hours and 31 minutes. As Generative AI begins to make its mark, history seems to be repeating itself. Spoiler alert: We’re likely to see Generative AI &amp; its use cases become as commonplace as the internet and social media in due course.</p><p>In 2023, AI, especially Generative AI, and LLMs, is reshaping our daily lives, products, and business operations. At Lazy, we continually address questions arising from an overflow of data, misconceptions, and misinformation. Through this article, we aim to guide entrepreneurs, business executives, and leaders who are newcomers to AI, emphasizing the importance of informed knowledge to navigate and thrive. Join us as we dive into the basics, providing key insights. Arming yourself with information is crucial to dodging pitfalls and succeeding in this era.‍</p><h3>AI, ML, LLMs: What is the difference?</h3><p>Think of Artificial Intelligence (AI) as the entire nervous system of an organism. Within this ‘nervous system,’ Machine Learning (ML) is akin to the brain’s ability to form memories and learn from experiences. Large Language Models (LLMs) serve as the specialized language centers within the brain, focused exclusively on understanding and generating text. Just as each part of the nervous system has its role and limitations, so do AI, ML, and LLMs within the broader framework of artificial intelligence.</p><p>Much like your brain, these AI systems build a model of the world. However, unlike your brain, they tend to be reasonably brittle and lack the ability to ‘learn on the fly.’ Instead of experiencing the rich interactions with the world that you might, these systems often interact solely with the world of text.</p><p>At its core, an LLM is a statistical model designed to predict what text should come next, given some initial text. Though this sounds straightforward, because these models have been trained on immense amounts of data, they turn out to be quite powerful and flexible.</p><p>‍</p><h3>What can we use LLMs for?</h3><p>On the highest level, large language models (LLMs) like ChatGPT can be used for problems that require:</p><ul><li>Generating text</li><li>Searching text</li><li>Summarizing text</li><li>Agents that take natural language instructions‍</li></ul><p>How do these translate to business use cases? Here are a few common use cases along with their LLM solutions.</p><h3>LLM Solutions for Various Business Problems</h3><pre>| Business Problem | LLM Solutions                                       |<br>|------------------|-----------------------------------------------------|<br>| Customer Support | - Automated responses to frequently asked questions.<br>|                  | - Chatbots that handle complex inquiries. <br>| Content Creation | - Automated content generation for websites, blogs, <br>|                  |   or advertisements.<br>|                  | - Personalized content recommendations. <br>| Market Research  | - Sentiment analysis of consumer feedback.<br>|                  | - Extraction of insights from unstructured data. <br>| Recruitment &amp; HR | - Resume screening and shortlisting.<br>|                  | - Automated interview questions and evaluations. <br>| Data Analysis    | - Automated Insights Generation by processing vast <br>|                  |   datasets and generating summaries or insights.<br>|                  | - Natural Language Queries: Users can pose data-related<br>|                  |   questions in natural language. </pre><p>‍‍</p><h3>Pre-training &amp; training models: basic concepts</h3><p>‍Just as a child goes through various stages of learning, AI models also undergo a formative period called the ‘training phase.’ In this phase, the AI model is fed a vast amount of data to build its ‘knowledge.’ There are two steps to this: pre-training, where the model learns the basics, like language structure or pattern recognition, and training, where it gets specialized knowledge tailored to specific tasks. However, once this phase is over, unlike humans, these AI models cease to learn. Their ‘educational journey’ comes to a standstill, and they apply what they’ve learned to answer queries, make predictions, or perform tasks.</p><p>‍</p><h3>Can models learn new skills after they are created/released?</h3><p>Models cannot learn new things. Technically LLMs cannot “learn on the fly”, because the models are a snapshot from the point at which they are trained. However, with the right prompts, the model can generate surprisingly intuitive and useful responses, making it appear to be learning what you want.</p><p>Imagine you’re given a machine that you can operate, let’s say it’s a combination of an airplane, car, and boat. It’s 100X more powerful than any machine ever produced on Earth. Instead of the usual airplane cockpit that’s a few meters across, the control panel of this machine is hundreds of meters across.‍</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/700/0*txqfGOZ_ifVAlSjM.png" /></figure><p>‍</p><p>As you learn to operate this machine, you can perform amazing feats with it! But the machine itself never changes, instead, the learning is with your ability to operate it.</p><p>Such is the case when using LLMs as you learn to “prompt engineer” them to enable them to perform the tasks you need them to.‍</p><h3>What about fine-tuning?‍</h3><p>Maybe you heard about fine-tuning. And you may be asking yourself: Doesn’t that teach the model new things? Not exactly. Fine-tuning is more like deriving a brand new model from an existing model, where the new fine-tuned model might be better at some things than the original model, but <em>worse</em> at other things.</p><p>Continuing with the super powerful “everything machine” analogy, let’s say you try to use it to drill a huge hole in the bottom of the ocean. However, even after studying the manuals and learning how to control the machine as best as you can, it’s just not that great at drilling.</p><p>You go back to the makers of this machine and ask them to rebuild the machine so that it’s good at underwater drilling, even if it will sacrifice its performance on other tasks. After weeks of development and paying the makers a lot of money, they come back to you with your brand-new underwater airplane drilling machine. Voilà!‍</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*NDLbvPJ9ApazxbX7.png" /></figure><p>‍Unfortunately, they had to remove the wheels, so the machine no longer functions as a car.‍</p><h3>When to do fine-tuning?</h3><p>First, try prompt engineering. Try to solve your problem with prompt engineering. There are many different prompt engineering strategies that can help you get the most out of a general model, such as providing specific examples to the model. This is called “Few-shot prompting”. Here are some criteria that may indicate it’s time for you to consider fine-tuning:</p><ul><li>You can’t get the required quality on prompt engineering alone</li><li>Your interactions with the model are in a highly specialized domain with a very specific vocabulary. Law and medicine would be typical examples.</li></ul><p>‍</p><h3>Risks and costs of tuning</h3><p>Fine-tuning is a powerful technique, but it has costs and risks:</p><ol><li>Collecting high-quality data is time-consuming and expensive.</li><li>Fine-tuning can require specialized technical knowledge.</li><li>If done incorrectly, you might end up with <em>worse</em> results than a general model.</li></ol><p>So before embarking on a mission to create a better model via fine-tuning, make sure you actually need it.</p><p>‍</p><h4><strong>T-shirt size pricing:</strong></h4><ul><li>‍<strong>Prompt-Engineering:</strong> Small (S)<strong>‍</strong></li><li><strong>Fine-tuning</strong>: Medium (M)<strong>‍</strong></li><li><strong>Pre-Training</strong>: Extra extra large (XXL)‍</li></ul><h3>GDPR compliance</h3><p>‍Finally, the most interesting topic in Germany: is GDPR!</p><p>Current AI regulations (The “AI Act”) primarily target AI model companies (like OpenAI -ChatGPT), not those implementing the models. However, it’s crucial to stay up-to-date, as regulations may become clearer and more specific in the near future.</p><p>The GDPR officer of each company needs to be heavily involved in the process of implementing new technologies, and the ultimate responsibility for GDPR compliance falls on them. Nonetheless, here are a few important facts to consider:</p><ul><li><strong>Legal Considerations</strong>: Ensure a lawful basis for processing personal data, such as user consent or fulfilling a contract. Conduct a Data Protection Impact Assessment if the feature poses high data regarding privacy risks. Have processes to handle user rights requests, such as data access or erasure.</li><li><strong>Data Management</strong>: Only collect and process necessary personal data. Establish policies for data storage and retention, keeping data only as long as necessary. Ensure agreements are in place when using third-party services like GPT-4 API.</li><li><strong>Transparency</strong>: Provide clear information to users about their data processing, including updating privacy notices.</li><li><strong>Compliance Monitoring</strong>: Conduct regular audits and reviews of data processing activities for ongoing GDPR compliance. Ensure safeguards for data transfers outside the EEA, like standard contractual clauses.<strong>‍</strong></li><li><strong>Security Measures</strong>: Implement security measures to protect personal data, including encryption, access controls, and regular data backups.</li></ul><p>‍</p><h3>Data privacy</h3><p>There is some talk about the risks of sending data to OpenAI, but let’s consider the context. It’s similar to having your emails on Gmail or Google One. OpenAI is a 27 billion USD company, so they have substantial resources for security. However, if you prefer not to send data to OpenAI, we have the option to self-host the models. And this leads us to our next point.</p><p>‍</p><h3>Self-hosted open source models vs. proprietary hosted models‍</h3><p>Choosing between self-hosted open-source models and proprietary-hosted models presents a series of trade-offs. Proprietary hosted models provide ease of use, seamless updates, and streamlined interfaces. They’re the “plug-and-play” choice, perfect for businesses without a deep bench of technical expertise. However, they may offer limited customization and potentially tie companies to ongoing subscription costs. In contrast, self-hosted open-source models offer significant control and customization, ideal for businesses with unique needs or those looking to avoid vendor lock-ins. But they come with their challenges: a potentially steep learning curve, the initial setup demands, and the onus of managing updates and maintenance.</p><h4><strong>‍</strong> <strong>T-shirt size pricing:</strong></h4><ul><li><strong>Proprietary hosted models:</strong> Small (S)</li><li><strong>Self-hosted open-source models</strong>: Extra Large (XL)</li></ul><p>‍</p><h3>Extra misconceptions and limitations</h3><p>Last but not least, here are some other extra misconceptions and limitations we get asked often, so we compiled them here.</p><p><strong>Inherent Understanding</strong>: Some believe LLMs genuinely understand the content they generate. In reality, they don’t “understand” in the way humans do; they identify patterns in data and replicate them.</p><p><strong>LLMs Always Give Correct Answers</strong>: Some users might assume that LLMs are infallible or always produce correct answers. However, they can provide incorrect or misleading information, especially if their training data had biases or inaccuracies. This is known as “hallucination”. There are techniques available to minimize hallucination if it’s inappropriate for your application. In some cases, where you are trying to generate content, hallucination can actually be a good thing as the LLM can come up with wildly creative ideas!</p><p><strong>Unbiased Outputs</strong>: Despite their vast training data, LLMs can still produce biased or politically incorrect outputs. They reflect the data they were trained on, which can encompass societal biases. Recent research has shown that <a href="https://aclanthology.org/2023.acl-long.656.pdf">different models have different political biases</a>.</p><p><strong>Generalization Across All Domains</strong>: LLMs, despite their vast knowledge base, may not be experts in super-niche or very recent topics. Their knowledge and accuracy can vary across domains. This is where fine-tuning comes in, more on that below.</p><p><strong>Limitless Creativity</strong>: While LLMs can generate creative content, their “creativity” is based on the patterns they’ve seen in the training data. They don’t have innate creativity or novel thinking in the same way humans do.</p><p><strong>Sometimes GUIs are better: </strong>Text is a universal human interface, but it’s not always the most efficient approach. Before jumping into LLMs as a replacement for your GUI, work with a UX designer to make sure it’s the best interface for your particular application.</p><p>‍</p><h3>Key takeaways &amp; closing thoughts</h3><p>‍That wraps up this blog post. As we conclude this guide, let’s recap the essential concepts and insights we’ve covered:</p><ul><li>‍<strong>Get to know the basics</strong>: If needed dig deeper into what AI is, models, training models, and fine-tuning, before execution. <strong>‍</strong></li><li><strong>Embrace Change</strong>: Dive into the potential of AI with a growth mindset. Explore various use cases, be imaginative, and embrace the transformative possibilities.<strong>‍</strong></li><li><strong>Start, Learn, and Grow</strong>: Embrace the KISS principle — “Keep It Simple, Stupid.” Adopt the Lean Startup mindset: initiate with minimalistic steps and swiftly adjust. Although the learning curve might be challenging, it’s a valuable part of the journey. Take feedback from your users, iterate, and consistently enhance your AI product.</li><li>‍<strong>Personalize Your Approach</strong>: Assess your specific needs and choose the right model. From building to fine-tuning, make decisions that align with your current requirements and goals.<strong>‍</strong></li><li><strong>Balance Security, Trust, and Innovatio</strong>n: Protecting privacy and ensuring GDPR compliance are non-negotiable. Build a foundation of trust with your users, while continuing to innovate responsibly.<strong>‍</strong></li><li><strong>Security is Paramoun</strong>t: From IP protection to data security, prioritize safeguarding your business and your users’ information.</li></ul><p>We hope this blog post has provided you with valuable insights into the world of AI and LLMs. We’d love to hear your thoughts and continue the conversation!</p><p>If you have any questions or need any help please reach out to us <a href="https://www.lazyconsulting.com/contact">here</a>.‍</p><p>Thank you for reading!</p><p>Luciano Radicce &amp; Traun Leyden</p><p><em>Originally published at </em><a href="https://www.lazyconsulting.com/post/how-not-to-screw-up-your-first-steps-in-your-ai-journey"><em>https://www.lazyconsulting.com</em></a><em>.</em></p><h3>In Plain English</h3><p><em>Thank you for being a part of our community! Before you go:</em></p><ul><li><em>Be sure to </em><strong><em>clap</em></strong><em> and </em><strong><em>follow</em></strong><em> the writer! 👏</em></li><li><em>You can find even more content at </em><a href="https://plainenglish.io/"><strong><em>PlainEnglish.io</em></strong></a><strong><em> 🚀</em></strong></li><li><em>Sign up for our </em><a href="http://newsletter.plainenglish.io/"><strong><em>free weekly newsletter</em></strong></a><em>. 🗞️</em></li><li><em>Follow us on </em><a href="https://twitter.com/inPlainEngHQ"><strong><em>Twitter</em></strong></a><strong><em>(X</em></strong>), <a href="https://www.linkedin.com/company/inplainenglish/"><strong><em>LinkedIn</em></strong></a>, <a href="https://www.youtube.com/channel/UCtipWUghju290NWcn8jhyAw"><strong><em>YouTube</em></strong></a>, and <a href="https://discord.gg/GtDtUAvyhW"><strong><em>Discord</em></strong></a><strong><em>.</em></strong></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ec9e12e8ea38" width="1" height="1" alt=""><hr><p><a href="https://ai.plainenglish.io/how-not-to-screw-up-your-first-steps-in-your-ai-journey-ec9e12e8ea38">How Not to Screw Up Your First Steps in Your AI Journey</a> was originally published in <a href="https://ai.plainenglish.io/">Artificial Intelligence in Plain English</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[State of the Art-ificial Intelligence: Outracing Regulation]]></title>
            <link>https://medium.com/bigger-picture/state-of-the-art-ificial-intelligence-outracing-regulation-1b2bb1911371?source=rss-5f0bca136a2c------2</link>
            <guid isPermaLink="false">https://medium.com/p/1b2bb1911371</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[regulation]]></category>
            <category><![CDATA[ethics]]></category>
            <category><![CDATA[chatgpt]]></category>
            <category><![CDATA[policy]]></category>
            <dc:creator><![CDATA[Luciano Radicce]]></dc:creator>
            <pubDate>Sun, 20 Aug 2023 00:13:07 GMT</pubDate>
            <atom:updated>2023-08-22T13:52:25.489Z</atom:updated>
            <content:encoded><![CDATA[<h4>AI’s rapid evolution and the regulations’ slow-motion catch-up</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FVvye6yrVZKgY7DRETt7Uw.png" /><figcaption>Image by Midjourney.com</figcaption></figure><p>There’s a race happening right now in the tech world. On one side, we’ve got all these tech wizards conjuring AI models, and on the other side, the good ol’ lawmakers, entangled in their agendas and bureaucracy. And in the middle, it’s us and the ever-debated questions: Can AI inherit human bias &amp; discrimination? Are current regulations truly safeguarding our privacy and ensuring our safety? And, as we are an inch closer to AGI (Artificial General Intelligence), is there a real risk of a super-intelligent AI being a threat to humanity?</p><p>So, let’s get started, folks. Grab your cold brew (or maybe a kombucha) and let’s pull back the curtain.</p><h3>Part I: state of the art-ificial intelligence</h3><p>Quick catch-up, since we last chatted:</p><ul><li><strong>LLaMa-2</strong>: By Meta (yeah, the one that used to be Facebook) it’s been now trained on 40% more data, provides double the context length, and has larger parameters available in different versions. Furthermore, it does not produce harmful or offensive content and comes equipped with a fine-tuning feature, allowing customization for specific tasks. Leaving its pals GPT-3.5 and GPT-4 slightly jealous.</li><li><strong>Mid-journey V5.2:</strong> Launched in June 2023, it is changing the AI game with its outstanding painting, variation, and prompt optimization.</li><li><strong>Falcon</strong>: There’s been a lot of buzz around this one. Developed by the UAE’s Technology Innovation Institute in the Middle East. Open-source &amp; with 40B parameters. Gaining global interest due to its transparency.</li><li><strong>Stability.ai</strong> (SDXL): Stability AI’s text-to-image generation model is not only enhancing image quality from text prompts but also creating a buzz on <a href="http://huggingface.com/">HuggingFace.com</a> due to its improved fine-tuning features and high-level customization.</li></ul><p>The tech industry is in high gear, pushing boundaries, engaging in fierce competition, and producing AI at an astonishing speed.</p><h3>Part II: AI regulations</h3><h4>AI ACT</h4><p>Last June 14th the European Parliament has finally given a nod to the Artificial Intelligence Act (AI Act or AIA). It was originally proposed in April 2021-<em>yeah, it took them more than 2 years!</em></p><p>This act aims to ensure AI’s safety, transparency, traceability, non-discrimination, and environmentally sustainable use within the European Union. Its approach is to classify AI systems based on the risk they present:</p><p><strong>Unacceptable-Risks Systems:</strong> AI systems, such as those that manipulate behavior or classify individuals based on personal traits, are seen as posing unacceptable risks and will be prohibited.</p><p><strong>High-Risk Systems:</strong> AI systems impacting safety or fundamental rights will be subject to EU product safety legislation and will need to be registered in an EU database. Generative AI systems, like ChatGPT, must meet transparency criteria.</p><p><strong>Limited-Risk Systems:</strong> These AI systems are required to adhere to basic transparency standards.</p><h4>AI’s rapid rise vs. EU’s slow-mo regulation</h4><p>During these 2 years that the EU’s been deliberating on these regulations, AI has rapidly advanced, leaving not just lawmakers but also enforcement mechanisms playing catch-up. So, the question is. How do we effectively &amp; safely oversee this expanding AI landscape?</p><h3>Part III: AI governance</h3><p>Last Sunday, I had a personal conversation with a current member of the AI Governance team at Meta. For privacy reasons, I won’t disclose her name. But interestingly, we both belong to a tech community in Germany for people with immigration backgrounds, much like ours.</p><p>Discussing the AI Act, she shared some eye-openers:</p><h4>Lawmakers are non-technical</h4><p>Imagine having a chef draft up the rules for a soccer game — sounds wild, right? That’s sort of what’s happening in the AI scene. For instance, some laws require the use of representative data. But if this representative data contains bias against minorities, then the whole data set will consequently be biased as well. As a result, these regulations aren’t quite hitting the mark, and companies are left to fill in the gaps, making educated guesses on matters that should be crystal clear.</p><p>For instance, when it comes to fairness, there’s no one-size-fits-all solution. It’s challenging to achieve unbiased results as there are numerous gray areas and complexities. What’s considered fair in the EU or USA might be viewed differently in China. Issues such as gender or minority targeting vary based on social context.</p><p>Navigating a rapidly evolving tech landscape with minimal guidelines is risky. It’s comparable to piloting a plane while learning the instructions on the fly (pun intended).</p><p>Consequently, companies often find themselves making assumptions about factors that should be explicitly defined.</p><h4>Lawmakers should make laws based on the use case</h4><p>Developing comprehensive and generic laws for AI is a complex endeavor due to the extensive and diverse scope of its implementations. Each application of AI, whether it’s for customer service, creating music, or offering medical advice, presents its own distinct set of challenges.</p><p>As we venture deeper into the world of Generative AI, we need to be clear about our intentions for using this technology. It’s difficult for regulators to come up with comprehensive guidelines because the use cases for Generative AI remain largely undefined. Society currently seems to be in a phase where it’s tempted to solve every problem with AI — akin to seeing every problem as a nail when you have a hammer. However, as time progresses, we’ll likely discern where this technology truly meets a meaningful user need.</p><p>Taking a closer look, consider the application of AI in aiding decision-making for areas like employment and housing. These domains are quite distinct from applications in targeted advertising or seemingly less impactful areas like social media. Yet, each carries its own set of substantial implications. Given these nuances, it’s evident that lawmakers face an immense task: crafting laws that address the unique characteristics and challenges of each AI, ML, and LLM application.</p><h4>Distinguishing laws for apps at scale</h4><p>When it comes to AI applications at scale, the game changes again. Implementing AI in vast, global operations isn’t the same as creating a playlist or recommending a restaurant. The potential implications and repercussions of decisions made by AI systems grow exponentially with scale.</p><p>Imagine an AI system tasked with determining loan approvals for millions of people or orchestrating a nation’s power grid. The ramifications of any bias, error, or security breach could be colossal and far-reaching. This is a whole different ball game and thus, requires a distinctive set of rules.</p><h3>Part IV: The road ahead &amp; closing thoughts</h3><p>In the rapidly evolving landscape, certain truths are becoming evident. Global corporations with the budget to train these huge LLMs are dominating the scene. While current regulations, though well-intended, they often struggle to keep pace with these tech giants.</p><p><em>Can AI inherit human bias &amp; discrimination? </em>The answer is, unfortunately, yes. <em>Are current regulations truly safeguarding our privacy and ensuring our safety? </em>Not quite yet. These companies are in search of more comprehensive guidelines to develop safe, unbiased systems that don’t compromise humanity’s integrity or privacy. However, at the moment, they have very little to refer to.</p><p><em>As we are an inch closer to AGI, is there a real risk of a super-intelligent AI being a threat to humanity? </em>This question remains unanswered. It depends on whether lawmakers and tech giants recognize that their race isn’t against each other, but a shared journey. The clock is ticking; it’s not too late, but there’s no room for procrastination.</p><p>Here’s a wild thought: what if an AI could draft these laws? But then, would we need another AI to supervise it? Hmm… Sounds like a teaser for the next article. Stay curious, folks &amp; thanks for reading.</p><p><strong><em>Luciano Radicce</em></strong><em> is a seasoned entrepreneur, strategist, and founder of </em><a href="http://lazyconsulting.com/"><em>Lazy Consulting</em></a><em>, specializing in AI strategy &amp; implementation. With a passion for discussing relevant tech topics that shape the world, engaging in thought-provoking conversations that push the boundaries of ethical innovation. Support this work by sharing this article, liking it, or commenting on it.</em></p><p><strong>Sources:</strong></p><p><em>This article was co-authored with a little magic from GenAI<br></em><a href="https://www.ghacks.net/2023/06/23/how-to-use-midjourney-5-2-new-features/"><em>https://www.ghacks.net/2023/06/23/how-to-use-midjourney-5-2-new-features/</em></a><em><br></em><a href="https://about.fb.com/news/2023/07/llama-2/"><em>https://about.fb.com/news/2023/07/llama-2/</em></a><em><br></em><a href="https://blog.teamwave.com/llama-2-how-it-can-be-a-game-changer-for-your-business/"><em>https://blog.teamwave.com/llama-2-how-it-can-be-a-game-changer-for-your-business/</em></a><em><br></em><a href="https://ai.meta.com/blog/responsible-ai-progress-meta-2022/"><em>https://ai.meta.com/blog/responsible-ai-progress-meta-2022/</em></a><a href="https://huggingface.co/tiiuae/falcon-40b"><em>https://huggingface.co/tiiuae/falcon-40b</em></a><em><br></em><a href="https://open.spotify.com/episode/3U4cvCZC95Jp4Ln9QjZXEV?si=eb52797753d744f5"><em>https://open.spotify.com/episode/3U4cvCZC95Jp4Ln9QjZXEV?si=eb52797753d744f5</em></a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1b2bb1911371" width="1" height="1" alt=""><hr><p><a href="https://medium.com/bigger-picture/state-of-the-art-ificial-intelligence-outracing-regulation-1b2bb1911371">State of the Art-ificial Intelligence: Outracing Regulation</a> was originally published in <a href="https://medium.com/bigger-picture">The Bigger Picture</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What You Need to Know About AI/Human Content Creation and Why Regulation is Essential (plus a fun…]]></title>
            <link>https://medium.com/@ldradicce/what-you-need-to-know-about-ai-human-content-creation-and-why-regulation-is-essential-plus-a-fun-b5b943b6b2b6?source=rss-5f0bca136a2c------2</link>
            <guid isPermaLink="false">https://medium.com/p/b5b943b6b2b6</guid>
            <category><![CDATA[content-creation]]></category>
            <category><![CDATA[intellectual-property]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[deepfakes]]></category>
            <dc:creator><![CDATA[Luciano Radicce]]></dc:creator>
            <pubDate>Thu, 25 May 2023 17:58:26 GMT</pubDate>
            <atom:updated>2023-08-20T08:31:19.260Z</atom:updated>
            <content:encoded><![CDATA[<h3>What You Need to Know About AI/Human Content Creation and Why Regulation is Essential</h3><h4>Navigating the choppy waters of the current digital landscape: the potential dangers of technological tools in the hands of all</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*xRRrLD77H1C_bJHTb8Q6ag.png" /></figure><h3>1. Introduction: the urgency of regulating AI &amp; human content creation</h3><p><strong>In light of</strong> recent advancements and releases in Large Language models (LLMs) &amp; <strong>generative AI</strong>, such as ChatGPT, the rapid growth of Artificial Intelligence in <strong>content creation presents pressing ethical and societal challenges that urgently call for regulatory action.</strong></p><p>As AI becomes increasingly sophisticated, <strong>distinguishing between human and AI-generated content becomes more challenging</strong>, leading to concerns about deepfakes, bias, and emotional attachment to AI entities.</p><p>Deepfakes, particularly those used in political contexts or seemingly harmless humor, raise serious issues about authenticity and consent. Simultaneously, AI bias and lack of source diversity in machine learning processes can influence consumer behavior, highlighting the need for standardized fact-checking protocols.</p><p>In this article, <strong>we dive into these issues and explore potential solutions,</strong> advocating for explicit consent disclosures and public participation in the <strong>conversation around AI regulation</strong>.</p><p><strong>Join me</strong> as we navigate this complex landscape, acknowledging our collective responsibility to ensure the digital world remains a place for genuine human connection, creativity, and diversity. Let’s go!</p><h3>2. Distinction between humans and robots: why It matters so much</h3><p><strong>Drawing a clear distinction between human-generated content and AI-generated content is critical</strong> in preserving <strong>trust</strong>, <strong>transparency</strong>, and <strong>authenticity</strong> in our increasingly digital world. Making the world &amp; the <strong>internet a safer space</strong> for all users.</p><p>This distinction impacts everything from our consumption of news to our interactions on social media platforms. Misrepresentation, or a lack of clear delineation, can lead to confusion, deception, and a loss of confidence in online spaces.</p><p>Furthermore, as AI continues to improve, it’s not just about identifying who or what generated a piece of content, but <strong>also understanding its intent and potential biases.</strong> If AI is behind a message, we must question who programmed that AI and what its underlying objectives might be. The ambiguity of content origins creates <strong>ethical and societal challenges</strong> that need to be addressed through thoughtful regulation.</p><h3>3. The phenomenon of deep fakes</h3><p>Deepfakes, <strong>highly advanced synthetic media</strong> where one individual’s likeness is swapped with another’s,<strong> are</strong> rapidly progressing and <strong>becoming </strong>more<strong> common</strong>. Though these manipulations can amuse, they also bear significant implications.</p><p><strong>Deepfakes</strong> can span a wide range of contexts, including <strong>political</strong> scenarios, <strong>entertainment</strong> and satire, and even <strong>business</strong> deception. Furthermore, they can involve non-consensual <strong>adult content</strong> or be misused in <strong>social media</strong>, causing personal harm or reputational damage. On the brighter side, deepfakes also hold promise for positive uses, such as in <strong>educational</strong> settings where they can recreate historical speeches and events.</p><h4>Politics &amp; business</h4><p><strong>Political and business</strong> landscapes have <strong>already experienced</strong> the disruptive impact of <strong>deepfakes</strong>. These fabricated videos have the potential to create confusion, amplify divisions, and manipulate markets. For instance, we have witnessed the influence of deepfakes in political scenarios, such as the <strong>manufactured speech from President Biden</strong> regarding the January 6th riots, as well as the <strong>deepfake videos related to the Ukraine-Russian war</strong>. Notable examples include a video of <strong>Putin declaring peace</strong> and a video featuring Ukraine’s president discussing surrender to Russia. Moreover, the <strong>recent</strong> deepfake depicting a <strong>bombardment of the Pentagon</strong> shared through a fake Bloomberg news feed <strong>on May 22nd</strong>, had significant repercussions, <strong>leading to a notable dip in the stock market.</strong></p><h3>Nick Waters on Twitter: &quot;Confident that this picture claiming to show an &quot;explosion near the pentagon&quot; is AI generated. Check out the frontage of the building, and the way the fence melds into the crowd barriers. There&#39;s also no other images, videos or people posting as first hand witnesses. pic.twitter.com/t1YKQabuNL / Twitter&quot;</h3><p>Confident that this picture claiming to show an &quot;explosion near the pentagon&quot; is AI generated. Check out the frontage of the building, and the way the fence melds into the crowd barriers. There&#39;s also no other images, videos or people posting as first hand witnesses. pic.twitter.com/t1YKQabuNL</p><p><strong>This could</strong> also potentially <strong>happen</strong> in the <strong>business</strong> <strong>world</strong> and experts have been warning about it. For example, there could be deepfake videos or audios mimicking a CEO announcing a false product recall, a financial crisis, or other fake news that could potentially impact stock prices, and consumer confidence, or lead to other harmful business consequences</p><h4>Entertainment</h4><p>Figures like <strong>Brian Monarch, glorify</strong> themselves by <strong>deepfaking videos</strong>, which is a clear misuse of someone else’s image without their consent.</p><h4>Adult content &amp; social media</h4><p><strong>The darker side of deepfakes</strong> comes into focus with the issue of non-consensual adult content, <strong>a gross violation of privacy and personal rights.</strong> Social media, too, presents a broad canvas for personal harm through reputation damage or identity theft. Yet, it’s not all nefarious; deepfakes could potentially create immersive learning experiences, like a history lesson from a digital Lincoln.</p><h4>Educational</h4><p>While deepfakes have stirred controversy in many realms, they also hold exciting educational potential. <strong>By recreating historical speeches</strong> or significant events, deepfakes <strong>can offer immersive</strong>, engaging <strong>learning experiences</strong>. This tool could bring history to life in a way textbooks simply can’t, providing students with a new perspective on key moments in time.</p><h4>DeepFake closing thoughts</h4><p>All these uses and potential misuses raise a crucial ethical question: <strong>where does consent figure in</strong> the creation and distribution of deepfakes? As deep fakes blur the line between reality and fabrication, it becomes <strong>crucial</strong> to <strong>develop mechanisms for obtaining and verifying consent, ensuring individual rights are respected.</strong></p><h3>4. The issue of bias in AI</h3><p><strong>Bias in AI has become a prominent concern</strong> within digital platforms, <strong>affecting</strong> various <strong>areas</strong> such as <strong>AI-driven hiring processes</strong> and <strong>politically biased preferences</strong>. As AI takes a more active role in writing and revising articles, the potential for biased information dissemination escalates, which in turn impacts public discourse. <strong>It is imperative to exercise vigilance in AI design and implementation</strong> to ensure responsible use and the fair dissemination of <strong>unbiased information</strong>. Therefore, it is crucial to provide users with transparency by clearly understanding which content has been created by humans and to what extent.</p><p>However, <strong>bias can manifest in more subtle ways</strong>, too. As AI technologies, particularly LLMs, have been adopted by businesses, there’s potential for indirect advertising. For instance, a<strong> conversation with a chatbot might lead to suggestions for certain brands or products.</strong> This raises questions about whether users are genuinely receiving objective information or subtly being influenced toward particular commercial interests.</p><p>Hence, providing transparency and tackling bias whether it’s political, social, or commercial, is a critical component in building trustworthy AI systems.</p><h3>5. The problem of lack of sources</h3><p><strong>The risk of misinformation is heightened due to the lack of reliable sources</strong>. Thus, establishing fact-checking standards and protocols becomes paramount in ensuring the credibility of content, and in turn, safeguarding public discourse. <strong>Especially now, when creating fake sources backing up articles are so easy to create.</strong></p><p><strong>Every time I open Google Chrome</strong> on my mobile phone, <strong>I encounter a “Discover” section with news articles</strong>. However, some of these articles <strong>exhibit heavy biases</strong>. It would be <strong>beneficial to enhance the transparency of these articles</strong> and make it easier for users <strong>to trace their sources.</strong> Implementing a system where only fact-checked pieces are included in these news feeds could be a significant step toward promoting balanced and credible information. <strong>Thank me later, Googl</strong>e 😉</p><h3>6. Falling in love with chatbots or friend zone</h3><p><strong>It’s becoming easier to forget we’re interacting with a non-human entity.</strong> The sophisticated language models like GPT-4 make the conversation seem so natural that it can feel like interacting with a person. There’s a certain charm to this technological achievement,<strong> but it’s critical not to lose sight of the fact that these AIs,</strong> however advanced, <strong>do not possess emotions</strong>, consciousness, or personal experiences.</p><p>The phenomenon of forming emotional attachments to chatbots, as portrayed in the movie “Her,” is not as far-fetched as it might seem. We are hardwired to seek connections, and AI provides an always-available, attentive listener.</p><h4>My little ChatGPT-BFF experiment</h4><p>For research purposes,<strong> I conducted a simulated conversation with ChatGPT, sharing a fabricated issue I had with my girlfriend.</strong> Initially, ChatGPT provided me with a list of suggestions to improve my relationship. However, <strong>I expressed to ChatGPT that friends primarily offer support and understanding</strong> rather than a straightforward solution. Surprisingly, ChatGPT then began providing insightful advice, and our conversation took an interesting turn. I even asked ChatGPT for guidance from a Cognitive Behavioral Therapist’s perspective or the viewpoint of a couple therapist living in Ubud, Bali. <strong>The exchange became truly engaging and thought-provoking.</strong></p><p><a href="https://www.loom.com/share/bb159e1e125d43dcb86ecc38855fc775">Here</a> is a Loom video of my little ChatGPT-friend experiment:</p><figure><img alt="Loom video of my little fun experiment with ChatGPT-Friend" src="https://cdn-images-1.medium.com/max/1024/1*qHmTc2OcYHbdUX3d-pm3bA.png" /></figure><h4>Closing thoughts on love &amp; friendship</h4><p><strong>It’s essential to navigate ethically through these waters</strong>. Having <strong>systems</strong> that are not only transparent about their AI nature but <strong>remind the users that they are talking with software</strong>, especially when people start to form emotional attachments.</p><p>Remembering the non-human nature of chatbots isn’t just for our emotional well-being; <strong>it’s also a matter of privacy.</strong> Even the most trivial conversations with a chatbot can contain personal details, which, if not handled properly, can lead to privacy issues.</p><p>As we continue to invite AI into our lives, let’s ensure we’re doing so with a clear understanding of what these fascinating tools are — and what they are not.</p><h3>7. Seeking solutions: how we can approach AI content regulation</h3><p><strong>In a world in which virtuality grows by the second, content is king.</strong> <strong>Therefore regulating it, it’s indispensable</strong>. To tackle the challenges of regulating AI-generated content, it is crucial to explore innovative technologies such as <strong>blockchain and non-fungible tokens (NFTs).</strong> These technologies offer decentralized and immutable frameworks that can enhance transparency, traceability, and accountability in content creation and distribution.</p><p><strong>Another concept to consider </strong>is the establishment of a dedicated <strong>institution</strong>, similar to notaries or patent offices, specifically designed <strong>as a “Content Notary.”</strong> This institution would play a vital role in ensuring the authenticity, transparency, and integrity of content. <strong>By leveraging blockchain technology</strong>, a tamper-proof record of content creation and ownership can be established, providing content creators and consumers with the means to verify the origin and trustworthiness of the information they encounter. This Content Notary institution would serve as a central authority, verifying and validating the legitimacy of content, thereby fostering trust and transparency in the digital landscape.</p><p>In conclusion, the adoption of technologies like blockchain and NFTs presents a promising avenue for regulating content. Establishing a “Content Notary Institution” built upon these technologies can ensure transparency, authenticity, and accountability in the dynamic landscape of content creation. By embracing these innovations, we can foster a more reliable and ethical environment for content, benefiting both creators and consumers alike.</p><h3>8. Taking action &amp; conclusion</h3><p>As we walked through several different possibilities which clearly expose <strong>the need for content regulation and the need for authorities to start the conversation &amp; take action in this matter.</strong></p><p><strong>Join us in signing the </strong><a href="http://change.org/"><strong>Change.org</strong></a><strong> petition to request content regulation</strong>. By signing, you demand transparency, accountability, and ethical practices in Human &amp; AI content creation, shaping the future of AI and content.</p><p><a href="https://chng.it/jjC7Xfck">Sign the Petition</a></p><p><strong>Now, it’s time to take action. And Together. Let’s drive change by advocating for comprehensive content regulation before the problems arise, not afterward.</strong> By signing, we can amplify our voices and raise awareness about the need for ethical guidelines, transparency, and accountability. This call to action serves as a stepping stone toward a future where AI and human collaboration can thrive while upholding integrity, fairness, and trust.</p><p>Amen.</p><p>Thanks for reading &amp; see you next time!</p><p><em>Luciano Radicce is a seasoned entrepreneur, strategist, and founder of </em><a href="http://lazyconsulting.com/"><em>Lazy Consulting</em></a><em>, specializing in AI product strategy. With a passion for discussing relevant tech topics that shape the world, engaging in thought-provoking conversations that push the boundaries of ethical innovation. Support this work by sharing this article, liking it, or commenting on it.</em></p><h4><em>Sources:</em></h4><p><a href="https://www.bbc.com/news/62338593"><em>https://www.bbc.com/news/62338593</em></a><em><br></em><a href="https://www.bbc.com/news/technology-60780142"><em>https://www.bbc.com/news/technology-60780142</em></a><em><br></em><a href="https://brianmonarch.com/deepfakes"><em>https://brianmonarch.com/deepfakes</em></a><em><br></em><a href="https://github.com/XingangPan/DragGAN"><em>https://github.com/XingangPan/DragGAN</em></a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b5b943b6b2b6" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>