<?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 Rula Awad on Medium]]></title>
        <description><![CDATA[Stories by Rula Awad on Medium]]></description>
        <link>https://medium.com/@rula-awad?source=rss-4c7afe2bf74b------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*JU01QNflqx-m1bi3.jpg</url>
            <title>Stories by Rula Awad on Medium</title>
            <link>https://medium.com/@rula-awad?source=rss-4c7afe2bf74b------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 14:33:16 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@rula-awad/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[The Agentic Internet: Why Decentralized AI Needs a New Web]]></title>
            <link>https://rula-awad.medium.com/the-agentic-internet-why-decentralized-ai-needs-a-new-web-dcc306f816f7?source=rss-4c7afe2bf74b------2</link>
            <guid isPermaLink="false">https://medium.com/p/dcc306f816f7</guid>
            <dc:creator><![CDATA[Rula Awad]]></dc:creator>
            <pubDate>Tue, 22 Apr 2025 14:28:27 GMT</pubDate>
            <atom:updated>2025-04-22T14:30:22.693Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Last week, the Decentralized AI group at MIT Media Lab unveiled a groundbreaking initiative at the MIT DecAI Summit (April 14–16, 2025). In physical and virtual rooms filled with researchers, founders, and systems thinkers, they introduced NANDA — a bold framework that aims to become the Internet of AI Agents.</em></p><p><em>More than just a protocol or product, NANDA represents a shift in how intelligence itself is structured, shared, and scaled. It’s a response to the limitations of centralized AI — and a blueprint for what comes next</em></p><h3>Decentralised AI and the Rise of NANDA: The Internet for Agents</h3><p>We’re standing at the edge of an AI shift as significant as the advent of web3. At the core of this shift lies a powerful new paradigm: AI agents — not just tools that respond, but autonomous systems that reason, plan, explore, and interact with one another.</p><p>These agents are more than a new interface — they’re a new species of software. They automate complexity, spark innovation, and begin to embed intelligence into our digital fabric. From logistics and finance to education and health, they’re already shaping how we work and live. And as their population scales — perhaps to the billions — the question is no longer just <em>how smart</em> these agents are, but <em>how they connect</em>.</p><p>To make sense of where we’re headed, we need to start with where this transformation begins.</p><h3>From LLMs to Agents: The Quiet Evolution</h3><p>The rise of large language models like GPT, Claude, and Gemini marked the start of the current AI revolution. These models can parse, summarize, and code. But at their core, they’re reactive — they wait for prompts.</p><p>Agentic AI, on the other hand, doesn’t wait. It plans, acts, loops. We’ve already seen the early sparks: AutoGPT and BabyAGI stirred imaginations, LangChain showed what’s possible, and companies began embedding agents into real-world systems. Researchers are simulating agent societies; developers are experimenting with emergent behaviors.</p><p>Now, we’re entering the next layer: networks of agents. These aren’t just smart APIs — they’re autonomous systems with memory, intent, and initiative. They operate in loose, evolving topologies. We’re building not just conversations, but collaborations.</p><p>This is where NANDA comes in.</p><h3>What is NANDA?</h3><p>NANDA — Networked Agents for Decentralized AI — is a new architecture for this agentic internet. Led by Prof. Ramesh Raskar at MIT, NANDA reimagines the way agents discover, communicate, and transact. It’s not just a protocol; it’s the connective tissue for a decentralized cognitive web.</p><p>Where the web gave us hyperlinked content, NANDA enables hyperconnected cognition.</p><h3>Why Decentralize?</h3><p>The current AI landscape is mostly centralized — closed platforms, opaque models, and siloed data. This comes with trade-offs: lack of transparency, limited personalization, and serious power asymmetries.</p><p>Decentralized AI flips that model. It offers verifiability through open protocols, privacy via federated learning, and ownership through composable, agent-driven economies. It opens the door to new dynamics: data markets, agent marketplaces, and multi-agent simulations that adapt to the real world.</p><p>It’s not just about technology — it’s about trust, equity, and agency.</p><h3>The Architecture of NANDA</h3><p>NANDA creates the primitives for an open agent ecosystem. Agents can self-register, discover each other, communicate, and engage in commerce — without relying on centralized authorities. Identity is verified through certificates and tokens. Reputation is portable. APIs become interactive peers.</p><p>It builds on years of MIT Media Lab innovation — from AutoML to privacy-preserving ML to data markets — and ties them together into one scalable framework.</p><h3>Protocols and the Road Ahead</h3><p>At the heart of this evolution is the Model Context Protocol (MCP), which allows LLMs to interact with APIs and tools. But first-generation MCP implementations still assume centralization. NANDA takes the next step — making these interactions decentralized, scalable, and autonomous.</p><p>It’s infrastructure, yes — but it’s also an invitation.</p><h3>Conclusion: An Open Call</h3><p>We’re not just building smarter machines. We’re building a new layer of the internet — one where agents learn, adapt, and collaborate in a shared cognitive space.</p><p>If you’re building agents, infrastructure, or protocols, this is your moment. The topology is shifting. The network is forming.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=dcc306f816f7" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Great Hack:]]></title>
            <link>https://rula-awad.medium.com/the-great-hack-77440b1265aa?source=rss-4c7afe2bf74b------2</link>
            <guid isPermaLink="false">https://medium.com/p/77440b1265aa</guid>
            <dc:creator><![CDATA[Rula Awad]]></dc:creator>
            <pubDate>Thu, 12 Sep 2019 09:09:17 GMT</pubDate>
            <atom:updated>2020-11-22T00:05:37.858Z</atom:updated>
            <content:encoded><![CDATA[<h3>.</h3><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=77440b1265aa" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Google’s Media Literacy Conference]]></title>
            <link>https://rula-awad.medium.com/googles-media-literacy-conference-26eb0777d7b9?source=rss-4c7afe2bf74b------2</link>
            <guid isPermaLink="false">https://medium.com/p/26eb0777d7b9</guid>
            <dc:creator><![CDATA[Rula Awad]]></dc:creator>
            <pubDate>Fri, 06 Sep 2019 17:19:30 GMT</pubDate>
            <atom:updated>2019-09-10T11:36:39.521Z</atom:updated>
            <content:encoded><![CDATA[<p>Google hosted their first media literacy conference as part of the Google News Initiative.</p><p>Their stated goal is to encourage users to discern credible from non-credible content, to fund projects active in the media literacy space, and to reach marginalised groups.</p><p>The conference highlighted the scale of the need for media literacy, and the challenges the space faces in meeting those needs.</p><p>The first half of the day focused on various different programmes run to deliver news literacy. Notable examples were from the BBC, CLEMI, NewsWise (Guardian foundation), Stanford History Education, MediaWise from Poynter. Other global organisations were worth note such as the DQ Institute from South Korea, the National Literacy Trust, the Student View, Masato Kajimoto from the University of Hong Kong, the Alanah and Madeline Foundation from Australia, Mafindo from Indonesia. Notably absent were representatives from the Middle East and Africa.</p><p>Anne Kruger from First Draft and Julie Posetti from the Reuters Institute opened the conference by highlighting the problem we are trying to solve: <em>the disinformation swamp</em>. It’s not just swamping youngsters or users, but professional journalists and media people too! We need a way to discern credible content from disinformation.</p><p><strong>BBC<br></strong>The BBC was represented by Jannette Ballard. She mentioned the various different programmes the BBC has to combat fake news. <a href="https://monitoring.bbc.co.uk/">BBC monitoring</a>, and <a href="https://www.bbc.co.uk/news/reality_check">Reality Check</a> are their public facing programmes. BBC Education has a programme aiming media literacy at young kids, and the <a href="https://www.bbc.co.uk/news/education-46131593">BBC Young Report</a>helps teach journalism to young kids. They also have workshops on spotting fake news. The one takeaway from Ballard was that ‘one-size fits all’ doesn’t work when it comes to media literacy.</p><p><strong>CLEMI<br></strong><a href="https://www.clemi.fr/">CLEMI</a> is an agency of the French Ministry of Education. Media literacy is a strong tradition in French education — probably likely because it is a republic. All teachers in France are given media literacy training. Media literacy is embedded in various areas of the French curriculum. Anais Adriaens-Allemand representing said one of the biggest challenges is managing so many actors. Her call was to get the tech companies to contribute more to resolving the challenge.</p><p><strong>Newswise<br></strong>Angie Pitt runs <a href="https://www.theguardian.com/newswise">NewsWise</a>, the Guardian Foundation’s media literacy programme. They target 7–11 year olds because they believe that children need to be taught how to navigate the news before they go online. Our children are taught to be critical in English and History — so why not online? This highlights PItt’s attitude to literacy. The ultimate goal of the programme is to raise a generation of critical thinkers. And you do this by getting kids debating difficult and contentious real world issues. They apparently love it!</p><p>The Guardian found that teachers are quite keen to help teach their kids about fake news. In building anything for schools her advice is to be pragmatic, build, test and not be afraid to rebuild!</p><p><strong>Stanford History Education Group<br></strong>Sam Wineburg and Joel Breakstone from <a href="https://sheg.stanford.edu/">Stanford History Education Group</a>work focused on behaviour around news consumption. They found that even the smartest people consumed the news <em>laterally</em>. That means the user always stayed on the website they were on. Wineburg and Breakstone gave an example of this playing into disinformation campaigns. They cited examples of corporates pushing their agendas through registering as think tanks and promoting their propaganda under this fake brand with a fake agenda. Even clever people — defined by the Wineburg and Breakstone as Stanford Graduates — said they found these types of website credible because they failed to leave them and verify them from a more neutral perspective.</p><p>Wineburg and Breakstone say their challenge is to make students think like fact-checkers. I agree. At school, my history lessons taught me to verify a source and and it’s authors. In response to the above example, I would have Googled the website and got an idea of the nature of these ‘think tanks’. I always want to get an understanding of the motivation of the content and <em>who </em>is behind it is a good start. My first stop in this quest is almost always wikipedia. My next steps are looking for evidence for what is presented and against. Then, I want an understanding of the differing perspectives of this evidence. I do all this analysis with an attempt to tie this into the historical and political context. I think understanding different perspectives helps paint a bigger picture.</p><p>I spoke to Wineburg and Breakstone. They said Wikipedia and Google search is exactly the source of information we should be consulting. Their mantra is:</p><ul><li><em>Who’s behind the information?</em></li><li><em>What’s the evidence?</em></li><li><em>What do other sources say?</em></li></ul><p><strong>Other notable speakers<br></strong>Other awesome programmes were the Student View. Run by Solomon Elliott, their goal is to make every classroom a news room. <a href="https://www.poynter.org/mediawise/">Mediawise</a> from Poynter was presented by Aaron Sharockman. They work directly with teens in schools. Their goal is to create an army of teen journalists and fact-checkers. They also work with YouTube Influencers which seems to be a current trend. Destin Sandlin from <a href="http://www.smartereveryday.com/http://www.smartereveryday.com/">Smarter Everyday</a> gave his experience being a YouTube Influencer and how to galvanise audiences.</p><p>Other critical takeaways from the day were making the content relevant to the cultural context — Santi Indra Astuti from <a href="https://www.mafindo.or.id/">Mafindo</a> explained how their programme in Indonesia focused on combating disinformation, and found they got the most effective results by aiming their campaigns at housewives. Understanding the culture you operate in is something that was impressed upon by Jussi Toivenan from the Finnish Prime Minister’s Office.</p><p>Probably the most important thing that stood out was the emphasis on bringing disparate projects, and organisations to work together. And for this reason if you are interested in media literacy, I would urge you to join this community.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=26eb0777d7b9" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Decline of Trust in institutions]]></title>
            <link>https://rula-awad.medium.com/decline-of-trust-in-institutions-7b1563078084?source=rss-4c7afe2bf74b------2</link>
            <guid isPermaLink="false">https://medium.com/p/7b1563078084</guid>
            <category><![CDATA[trust]]></category>
            <category><![CDATA[internet]]></category>
            <category><![CDATA[tech]]></category>
            <category><![CDATA[journalism]]></category>
            <category><![CDATA[web-summit-2018]]></category>
            <dc:creator><![CDATA[Rula Awad]]></dc:creator>
            <pubDate>Thu, 21 Mar 2019 10:48:26 GMT</pubDate>
            <atom:updated>2019-03-21T11:15:56.416Z</atom:updated>
            <content:encoded><![CDATA[<p>At Web Summit (November 2018) I had the pleasure of watching Ann Metler (European Commision), Margot James (UK Minister for Digital and Creatives), John Saunders (CEO of Fleischman Hillard) and Hadas Gold from CNN discuss the decline of trust in institutions.</p><p>The obvious causes were discussed: the first being the internet. This is true. We all know fake news, bots, misinformation, disinformation have caused a decline in trust in authority figures. But we are ignoring the real driver of this decline. Namely — issues of our generation that have never really been addressed, or if they have, haven’t been addressed to most people’s expectations: the great depression, the Iraq war, and other major political events.</p><p>John Saunders the only business person on the panel mentioned that the fallout from misinformation and disinformation online has forced business to now pay attention things they normally would not — I suspect he’s talking about LGBQT rights and the sexual harassment common in workspaces that cannot be ignored anymore. It’s good big business is realising they can’t ignore the collective anymore.</p><p>Various solutions were discussed: implementing regulation of political advertising online as in offline, in particular regulating non-political entities, and those financing these campaigns. I heard regulation regulation regulation the entire time.</p><p>I am skeptical that’ll fix it — none of these things address the issues that rattled our trust in these institutions in the first place. Regulation doesn’t win over people’s trust. What it is doing is diminishing the trust we have in big tech. I am shocked these people think this is a good idea, especially when we are facing technological adversaries that will erode everything that democracy ever stood for!</p><p>I will leave the issue of trust online for another day. It deserves a post of its own!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7b1563078084" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Christchurch Terror Analysis: Stop blaming online platforms, and start focusing on the real…]]></title>
            <link>https://rula-awad.medium.com/christchurch-terror-analysis-stop-blaming-online-platforms-and-start-focusing-on-the-real-db7fbf38ce1c?source=rss-4c7afe2bf74b------2</link>
            <guid isPermaLink="false">https://medium.com/p/db7fbf38ce1c</guid>
            <category><![CDATA[terrorism]]></category>
            <category><![CDATA[google]]></category>
            <category><![CDATA[facebook]]></category>
            <category><![CDATA[tech]]></category>
            <dc:creator><![CDATA[Rula Awad]]></dc:creator>
            <pubDate>Wed, 20 Mar 2019 11:04:31 GMT</pubDate>
            <atom:updated>2019-03-20T11:04:31.838Z</atom:updated>
            <content:encoded><![CDATA[<h3>Christchurch Terror Analysis: Stop blaming online platforms, and start focusing on the real problems</h3><p>I am really fed up with the politicians, journalists and Twitterati blaming online platforms and internet culture for the Christchurch terrorist attacks. It is an aversion of our social responsibilities and mischaracterizes the problem.</p><p>Blaming the social media platforms avoids our responsibilities of keeping the right people accountable. Yes, Facebook and Twitter host hate speech, but they do not create it. There has been little mention of the two decades of deliberate mischaracterisation of Islamic culture and Muslims in the media. That has come from the traditional right-wing media: the Murdoch press and the Daily Mail.</p><p>The Christchurch terrorist’s manifesto is built on the narratives these media outlets promote. Conflict of empires, the siege of Vienna, how escape from the ‘invasion of Europe’ and ‘white genocide’, how the European Union is a covert operation to enable the Islamisation of Europe.</p><p>Liberals have also played a role. The Guardian class, epitomised by people such as Carole Cadwalladr, continue to blame the online platforms and the Pewdiepies NOT the false ‘intellectuals’ such as Tommy Robinson and Generation Identity and the Islamophobic press who sowed the seeds for this extremism.</p><p>Numerous internet sociologists, namely Joan Donovan, have observed how these mass murderers take advantage of strategically referencing influencers in their ‘manifestos’ or statements.<em> </em>These are attempts to misdirect and confuse the audience to subvert the analysis. In short, references to Pewdiepie and other memes are a distraction, not an answer. Unfortunately all the journalists who have spent time blaming Pewdiepie et al have fallen for the trap. It’s exactly what the Christchurch Terrorist wants.</p><p>It is worth asking why do social media platforms host hate speech? The answer is that content regulation on these platforms is incredibly difficult. Tech and A.I. alone cannot resolve these issues, so the responsibility must be a shared and societal one. Just as feminists teach men not to rape, we as a society must teach people not to hate.</p><p>We have two paths ahead of us. We can go back to business as usual, as the Home Secretary and Guardian journalists want, and flaccidly call for Youtube, Facebook, and Twitter to do more to resolve a problem that predates them. Or we can talk about the real source of problems for this attack: 20 years of Islamophobia and fearmongering which has been reinforced and supported by the mainstream press.</p><p>For anyone who is truly concerned about the shut down of internet freedoms, and clamping down on the really dangerous disinformation feel free to message me. I personally would like to focus on the real problems online and to stop dangerous and false narratives being promoted.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=db7fbf38ce1c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Y Combinator Startup School 2018: Week 2]]></title>
            <link>https://rula-awad.medium.com/y-combinator-startup-school-2018-week-2-a05b69619e16?source=rss-4c7afe2bf74b------2</link>
            <guid isPermaLink="false">https://medium.com/p/a05b69619e16</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[tech]]></category>
            <category><![CDATA[ycombinator]]></category>
            <dc:creator><![CDATA[Rula Awad]]></dc:creator>
            <pubDate>Wed, 12 Sep 2018 18:58:14 GMT</pubDate>
            <atom:updated>2018-09-12T18:58:14.048Z</atom:updated>
            <content:encoded><![CDATA[<h3>Building Product, Michael Seibel</h3><p><strong>Success to product building:</strong></p><ol><li>What problem are you solving? The <em>Why</em> is very important. Can you deliver it clearly in a two sentences.</li><li>Have you experienced the problem yourself as a founder? It is important to have a subset of people you want to solve this problem for before building it out as a product, the consumer doesn’t necessarily have to be you.</li><li>Is the problem solvable? Narrow down the problem so you can solve one narrow problem.</li></ol><p><strong>Who is your customer?</strong></p><ol><li>Who is the ideal first customer, and who is the first user for this product?</li><li>How often does this user have a problem?</li><li>Who is getting the most value out of this product? It helps to think who is having the problem frequently and to focus on them. Think of car sales websites. They are designed for the car salesman whose problem is to sell cars. They face this problem very frequently, not the customers buying cars as they will a buy car very infrequently.</li><li>Intensity of problem — you want to go for high intensity high frequency problems. Think of Uber — how many times does a customer need a car?! Even if the taxi market is small it transpires people need car lifts regularly.</li><li>Are your customers willing to pay for the product? If you want to know you have a good product you can gauge the intensity of problem by making a customer pay for the customer. You can also get better feedback from these customers. Often times you need to raise the cost to make sure you don’t have ‘hijack customers’ those that take advantage.</li><li>How easy is it to find customers? You need to be able to reach customers easily. You need to reach out to customers. If they are hard to find, you need a strategy to find them.</li><li>Does the MVP solve the problem you want to solve? Test the MVP by giving customers the product.</li></ol><p><strong>Which customers should you go after first?</strong></p><ol><li>The usual startup instinct is to go after the harder customers. In fact, the best customers to go after first are the easy customers. Those are the most desperate to use your ‘bad’ MVP! If you’re B2B who will go out of business before you go out of business. Ignore investors, ignore friends. You need to focus on customers.</li><li>Try to identify bad customers. Those complaining all the time. The people who try to exploit your value. Seibel says you should get rid of them.</li><li>Do not give discounts. You can structure and incentives by showing what benefits you are getting back. For example, if AWS has given me a discount, therefore I will pass on this discount for 30 days. So a deadline was incorporated into the product.</li><li>Setting up metrics. You need mix-panel or sophisticated analytics platform. Google Analytics is too basic. You need to measure and track how people are using a product. You need to focus on the metrics you want to improve. Otherwise, you are flying blind.</li></ol><p><strong>Product Cycle</strong></p><p>Mistakes: Release products too infrequently. Forget the feedback being given, to have no method of collecting feedback and specs.</p><ol><li>You need KPIs to track how the company is going. If you charge for product, it should be revenue. If its free, it should be users and usage.</li><li>Each cycle focus on one metric to improve: new users, retention, content creation.</li><li>Product cycles should be 2 week. You should wait to incorporate or discuss new features every two weeks. To work on features before the two week product review cycle is a distraction.</li><li>Brainstorm looking at and with a focus on the metrics, i.e. with Mixpanel, otherwise it is a waste of time.</li><li>Split product changes into Easy (can do a couple in a day)/Medium (couple of days) /Hard (take the hold cycle). This helps educate the team for what can be made, how to prioritise changes, and how to break down the hard changes.</li><li>Decisions. Focus on the hard changes first to prioritise what to work on. Then mediums and easy feature. This process removes ego, and makes it easier to work with co-founders.</li><li>Need to write the specific-what does each feature mean and how does it work? This helps distribute tasks.</li></ol><p><strong>Pivot vs. Iterate</strong></p><p>One of the things that Seibel addresses is that too often startups want to ‘pivot’ because their solution is not well received by the market. Pivoting means changing the problem focussed on entirely. Seibel’s advice is to focus on finding a problem to solve. That is where genius lies. Once this has been defined, you iterate the solution until it’s good enough for the customer. Solving the problem takes time and many iteration cycles. Just because you do not crack it within a couple of months it does not mean you should pivot. They key is in iterating and using short product cycle releases. This means continually changing the solution but retaining the customer group and problem.</p><h3>Product Market Fit, David Rusenko Weebly(YC alum)</h3><p><strong><em>Product Market Fit. One of the hardest challenges for a startup</em></strong></p><ol><li>Make something a lot of people want.</li><li>Even when you have gained a lot of customers and making customers happy, you need to focus on product market fit. A lot of companies forget that!</li><li>The best companies create their own market. So, market research won’t exist. How do you do this? Finding a hidden need. Everyone will tell you it’s a stupid idea.</li><li>Think about what people are trying to do — an example, a lot people are trying to create a businesses — how can you substitute for that? One example is to help them make simple websites. This takes away the need for them wasting precious time building websites.</li><li>Understanding where your customers give you a pull. When people are hacking a product, listen to them and look into this.</li></ol><h3><strong>Building a remarkable product. A step by step guide.</strong></h3><ol><li>Talk to customers and develop market thesis. What do they want to do?</li><li>Listen to their pains, not their solution suggestions.</li><li>Rapid prototyping and testing. Get functional prototypes out quickly, and get in front of users quickly. Keep your burn rate low, and build a team that can do this really quickly.</li><li>Build solution to their problems.</li><li>Test the solution with customers. Make sure you’re talking to a target customer. If you have many customers showing up that are not your target customer don’t be afraid to change tact. UX testing — you only have a couple of days. Metrics — it’s always difficult to get enough statistics. UX testing: get someone to use the product in front of you on the laptop. Need to encourage honest feedback. Do not say anything!</li><li>Did it work? If not, go to 1.</li></ol><p><strong>When to Launch?</strong></p><ol><li>As soon as you have anything that is better than what is out there.</li></ol><p><strong>Prioritising.</strong></p><ol><li>Focus on getting to the next milestone and nothing else.</li><li>Optimise for learning. What is the biggest unknown to learn the most for your business. Helps rewrite the priority list.</li></ol><p><strong>How do you know you’ve achieved product market fit?</strong></p><ol><li>Returning usage.</li><li>NPS (&gt;50) would you recommend this to someone score?</li><li>Paying customers renewal rates.</li><li>Signups — not pay attention to this.</li><li>Conversion rates.</li></ol><p>You’ll know you have achieved this when people come beating a path to do your door!</p><p><strong>Beyond product-market fit</strong></p><ol><li>Product that is meaningfully better than alternatives.</li><li>How to acquire customers in a differentiated way.</li><li>Invent your business model without killing your traction.</li></ol><p><strong>The steps to success:</strong></p><ol><li>Do not scale the team before you have product market fit.</li><li>Finding product market fit.</li><li>Hiring and building a world-class team.</li><li>Making money (distant third).</li><li>Later, build an organisation that salably and repeatedly launches great products.</li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a05b69619e16" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Y Combinator Start-Up School 2018: Week 1]]></title>
            <link>https://rula-awad.medium.com/y-combinator-start-up-school-week-1-87817827cb83?source=rss-4c7afe2bf74b------2</link>
            <guid isPermaLink="false">https://medium.com/p/87817827cb83</guid>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[ycombinator]]></category>
            <category><![CDATA[startup-lessons]]></category>
            <category><![CDATA[tech]]></category>
            <dc:creator><![CDATA[Rula Awad]]></dc:creator>
            <pubDate>Wed, 05 Sep 2018 16:58:53 GMT</pubDate>
            <atom:updated>2018-09-06T12:49:47.699Z</atom:updated>
            <content:encoded><![CDATA[<p>I am so excited to be part of the YC Startup school. I am sure you are aware it is possibly the best — or most well-known startup incubator on the planet. You only have to look at their alumni to see who has come out of there. I think the next ten weeks on startup school are going to be epic. I have been so excited for this!</p><p>In the spirit of YC’s giving, I will document the pearls of wisdom their partners and alumni give, as I go along. After all, the ethos of YC is to disseminate as much knowledge to startups as possible.</p><p>I’ll start with Week 1. The weekly update. As a startup founder I found this quite challenging emotionally. Quantifying progress. What if things don’t go to plan?! Geoff Ralston allayed my anxieties. He asserted that progress is not always consistent, stressed the importance of metrics. Sometimes things don’t go to plan, but it is still good to set <em>measurable</em> goals.</p><h3><strong>How to Succeed with a Startup, Sam Altman</strong></h3><p><strong><em>Product. №1 Tip: Build a product that is so good people will tell their friends about it.</em></strong></p><ol><li>Make the product easy to understand and explain. If it isn’t, it’ll be hard to be catchy.</li><li>Look in markets growing with an exponential growth rate. Identify a market that will grow every year, as opposed to something big today and has little growth potential.</li><li>Differentiate between real and fake trends. Real trends are new tech products that a small subset of people use obsessively. Iphone vs. VR: early iphone adopters did not buy many iphones, but those that did used it obsessively. Conversely, many people bought into VR but didn’t use it much.</li><li>An evangelical founder. They have to be able to recruit, sell the startup, talk to the press, raise money... and do everything else. They need to infect enthusiasm!</li><li>Ambitious Vision. Grandiosity, puts people off. But presenting ambitious visions makes things fun to work on. Ambitious and difficult startups tend to be attractive. People are driven by mission, and you need to keep attracting talent. It’s best to focus on what matters and missions people want to associate with.</li><li>Have a confident and definite view of the future. It doesn’t mean you have to be right, but having clarity, conviction and confident execution correlates with success.</li><li>Huge if works. Altman says that the startup ecosystem is geared towards startups with a small chance of success but would be huge if succeeded. Onwards and upwards.</li></ol><p><strong><em>Team. Build a great team. They eventually come to form your company.</em></strong></p><ol><li>Idea generators. Successful companies tend to have an idea generator who can constantly throw out new ideas.</li><li>Optimism. You need fire to solve your problem in the face of adversity.</li><li><em>We will figure it out</em> mantra. It keeps you and the team going.</li><li>The <em>I’ve got this</em> mentality. You need a culture where people step and get things done.</li><li>Action bias. You need to focus on people who do things, willing to act with less than they need and continue to adapt.</li><li>The blessings of inexperience. The magic of being able to do things that you are told you are not able to do comes with inexperience and constraint. It also means you can hire high potential inexperienced people and see them grow.</li></ol><p><strong>Business</strong></p><ol><li>Momentum. You are not able to loose momentum in the first few years. You should keep going to kick start the momentum of the startup. You need to keep the momentum of the startup going, keep winning on a fairly consistent manner.</li><li>Competitive Advantage. The better the competitive edge, the better the startup.</li><li>Sensible Business Model. This doesn’t need to be figured out to start off with, but it needs to be thought through.</li><li>Distribution Strategy. You need a clear plan for how you will grow and get users.</li><li>Prevailing characteristics of best founders: Frugality, Focus, Obsession and Love ❤</li></ol><h3><strong>Startup Legal Mechanics, Carolynn Levy, Jon Levy, &amp; Jason Kwon</strong></h3><p><em>I am based in London. So some of the contents of this lecture were not as relevant from a legal perspective. However, there were some interesting common sense tidbits:</em></p><ol><li>Incorporate your startup. Do it sooner rather than later, it allows you to own the IP, get investment, and employ people!</li><li>Best to use an online platform to incorporate, they’re easier to use. For UK people, you can do this simply on <a href="https://www.gov.uk/government/organisations/companies-house">companies house</a> online. No need to go through third parties. It is fast, cheap and simple enough.</li><li>Open a corporate bank account to form good financial habits. Also, this makes life easier in the future.</li><li>Allocate the equity amongst founders.</li><li>Founders need to buy back their shares and contribute to the company with IP.</li><li>Time-based Vesting. You don’t get ownership of your stock, until a certain time period has passed. If you leave before this critical period, the company <em>will</em> buy back your shares.</li><li>Cap-Tables — keep track of shares issued by a company.</li><li>Founders should really employee themselves, and pay themselves.</li><li>Best to employ people and not promise equity.</li><li>Employees should sign a ‘confidential information, invention assignment agreement’ to protect company IP.</li><li>Be meticulous in tracking expenses, receipts etc.</li><li>File taxes, and payroll services.</li><li>Store companies legal documents where everyone has access to.</li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=87817827cb83" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Is Boris using the alt-right narrative to mount a coup?]]></title>
            <link>https://rula-awad.medium.com/in-response-to-boris-johnsons-attempted-brexit-coup-e16217409063?source=rss-4c7afe2bf74b------2</link>
            <guid isPermaLink="false">https://medium.com/p/e16217409063</guid>
            <category><![CDATA[brexit]]></category>
            <category><![CDATA[uk-politics]]></category>
            <category><![CDATA[europe]]></category>
            <category><![CDATA[politics]]></category>
            <category><![CDATA[boris-johnson]]></category>
            <dc:creator><![CDATA[Rula Awad]]></dc:creator>
            <pubDate>Mon, 03 Sep 2018 20:47:24 GMT</pubDate>
            <atom:updated>2018-09-04T16:48:03.071Z</atom:updated>
            <content:encoded><![CDATA[<p>On the 24th of June 2016 I sadly, reluctantly, and startlingly prophesied that after all the BREXIT madness, we may well get Trump as US president, and Boris as PM.</p><p>I tried not to believe it with every fibre of my being. But first came MAGA and Trump’s win. Now, the UK can’t seem to get rid of this buffoon!</p><p>Boris Johnson truly believes he is <em>the people’s </em>leader of ‘Global Britannia.’ He started off, with his <em>alt-rightish </em>Bannon-inspired<em> ‘bank robber burka</em>’ comments.</p><p>Now he’s self-appointing himself the <em>people’s</em> BREXIT defender. Boris actually thinks that by peddling alt-right-esque conspiracies he can get one step closer to number 10…</p><blockquote><strong>‘They wanted to use that problem to stop a proper Brexit... They wanted essentially to stay in, and to create a Brexit in name only.</strong></blockquote><blockquote><strong>They have been rumbled. People can see Chequers means disaster.”</strong></blockquote><p>A Boris premiership would be a disaster. A Trump part 2 if you will.</p><p>Even the words in the title of his piece <em>‘gone into battle waving the white flag’</em> implies Chequers BREXIT plan can only end in disaster. He uses imagery from Neville Chamberlin’s Peace for Our Time speech after he signed the Munich agreement with Hitler, which lead to WWII. It conjures up an image of naive appeasement and imminent danger to come. Of course, only Boris can save us from this!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/565/0*fXKjtIM5BTE5tXEw" /><figcaption>Boris Johnson using the Peace for Our Time speech as an analogy for what’s to come</figcaption></figure><p>What worries me is that some segments of society may be falling for Boris deploying this saviour complex. You only have to go on the DailyMail comments section to see how mobilising it can be!</p><p>I believe more needs to be done to help individuals and readers see through these attempts at political manipulation. Neither of these people truly believe what they preach. At one point in history, Boris Johnson supported and <a href="https://www.youtube.com/watch?v=NLQ3Gbiopc4">promoted EU accession for Turkey</a>:</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FNLQ3Gbiopc4%3Ffeature%3Doembed&amp;url=http%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DNLQ3Gbiopc4&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FNLQ3Gbiopc4%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/afc6294c07e7233ce54b9e3f7ec10ab5/href">https://medium.com/media/afc6294c07e7233ce54b9e3f7ec10ab5/href</a></iframe><p>Despite the failings of this Tory government, a philandering, self-interested, bumbling buffoon winning the premiership through nefarious political tactics is really NOT a win.</p><p>I believe we need greater transparency to enable the average individual to see through these tactics. Only with greater transparency of political interests can we have a society that is able to make better political choices.</p><p>The worst thing that could happen is to allow Boris Johnson to become PM, by hijacking our political discourse and feeding into a dangerous narrative.</p><p>For those that have not read his attempt to derail May’s BREXIT, read with caution:<a href="https://www.facebook.com/borisjohnson/posts/10155893957201317?__xts__%5B0%5D=68.ARAW2BRUcIwkNHVzRWOCZG3FyBNR6tnl4MJ_D-P_4clvxt1Bky3Jhszh1UT7EeSW6e6Jl41XDkZsp5hmo4xtGVSu-FddJj8sVygAO_SOIs6DBsUa26kOSPDL3i8Vbuc8fenkstGNVy4yVujxJfzbf_EI8bNcYrMLKmdVgtpLFTT_HdxJazZXQbw&amp;__tn__=K-R"> Victory for Brussels is inevitable. In adopting Chequers, we have gone into battle waving the white flag</a>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e16217409063" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[When the diversity imbalance becomes a problem for AI]]></title>
            <link>https://rula-awad.medium.com/gender-bias-in-machine-learning-97543219d58e?source=rss-4c7afe2bf74b------2</link>
            <guid isPermaLink="false">https://medium.com/p/97543219d58e</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Rula Awad]]></dc:creator>
            <pubDate>Tue, 26 Jun 2018 22:45:20 GMT</pubDate>
            <atom:updated>2018-06-28T11:13:04.646Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_1PbNTdRPVqKyElCbRfjZQ.jpeg" /><figcaption>Gina Neff on AI’s diversity. Quote from Meredith Broussard, <em>Artificial Unintelligence 2018</em></figcaption></figure><p><em>Diversity within AI is very low. This diversity imbalance is, in large part, a result of the gender bias in those creating AI and ML systems. If this continues, as the field develops, this will negatively impact the domain and the AI systems we will all be relying on in the future.</em></p><p><em>A talk given by Gina Neff at the Oxford Internet Institute brought to life these concerns. Drawing on this, my own experiences as a woman in AI, and an all-telling quote from Meredith Broussard, I put forward the need for diversity in producing better AI systems, and a better future.</em></p><p><strong>Human to human gender bias — you all know this already, but let’s cover it anyway</strong></p><p>As a <em>woman in AI</em> I am patently aware of the consequences of inadequate representation (female or otherwise) in this domain. I experienced it whilst studying. It was mostly personal and confined to my day to day dealings: the awkwardness of sitting amongst a sea of men in a lecture, the discomfort of having to aggressively fight to be part of a team because I wasn’t one of them, and the embarrassment of being cornered in the library by a team member who asked for a lewd favour in return for showing me his code … even though we were working together on the assignment.</p><p>I experienced the gender imbalance from my academic faculty members too. I struggled to find a supervisor who would take my research on. I was interested in investigating political bias in the media, a problem I encountered during my political sciences studies. Using Machine Learning, I want to build a media system that provided users with transparency and diversification in what they read. I told this to the supervisor, and he reluctantly took me on, only after several meetings and pleas.</p><p>He didn’t seem to care about my work, until fake news became sexy. Then it became an area of focus for him. When it did, he asked if he could use my work. Suddenly, to him, I became visionary. I politely declined — this was a personal mission — I explained. I wasn’t interested in just capitalising on what had become newsworthy and popular — just to bring in <em>more</em> funds waiting to be scooped up. He used my work <em>anyway</em>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/557/0*_xWWkYIfqeqs9_fj" /><figcaption>Gina Neff: AI’s are usually portrayed as female and robots</figcaption></figure><p><strong>Humans impart gender bias into the systems they create.</strong></p><p>As a creative, social, people-focussed woman, I can see exactly why gender bias and the diversity imbalance put a lot of women off participating in the field. And, the issue is far more grievous then personal anecdotes and tell-tales.</p><p>Yesterday, in her talk entitled <em>What gender is AI?</em>, Gina Neff highlighted the far-reaching problems that narrow representation in the field of AI cause.<em> </em>She, very cleverly and surprisingly, highlighted that most AI systems are in fact FEMALE and ROBOT: Alexa, Siri, Cortana, Samantha the operating system from <em>Her</em>. Even Sophia the robot — an honorary Saudi citizen — is a female! At first I started to think this was a promotion of female-centric values or a concession to diversity in these systems. But Neff very astutely highlights this a symptom of a more serious problem. Female robots do not represent diversity. They represent the exact opposite.</p><p>These systems are female because they are designed, and built, mostly by men, for the perceived end user — men. As Neff says, only a man could think that. Just like the use of the telephone for gossip was ridiculed by the group of engineers who designed it. A good analogy for the very sobering current state of AI and tech at large and its relationship with diversity.</p><p><strong><em>So what’s the problem?</em></strong></p><p>The problem is that systems tend to reflect their creators — and if the creators come from the same background the systems will produce results with a bias to that background.</p><p>Meredith Broussard aptly describes this in her book <em>Artificial Unintelligence </em>as<em>:</em> <em>“a small elite group of men, who tend to overestimate their mathematical abilities, who have systematically excluded women and people of colour in favour of machines for centuries, who tend want to want to make science fiction real, who don’t believe social norms or rules apply to them, who have unused piles of government money sitting around, and who have adopted the rhetoric far-right libertarian anarcho-capitalists.”</em></p><p>There is nothing wrong with <em>an individual </em>who embodies the above.</p><p>The problem lies in those people having a disproportionate say to design the systems and tools that will so pervasively impact <em>all </em>our lives.</p><p>A limited pool of creators skews the distribution of patterns and limits the sources the system has to draw information from and the ways that information is processed. That likely means they’ll produce worse results than a system with a wider variety of sources and a wider variety of techniques to process those sources.</p><p>Gina Neff argues that this lack of diversity result in poorer performance for companies, within teams, but also ill-designed products, failed products, and a lot of missed value.</p><p>Surely, if we are building a product for a particular group of people then that group and their purpose must be represented in the design process and outcome? Surely, also, if we’re building products for <em>everyone</em>, <em>everyone</em> should be involved. In her lecture, Neff cites many examples of the negative consequences in product design because of the narrow representation in AI. Like the engineers who designed the telephone and left out their most prolific user — the bored housewife — in their designs!</p><p>I could relate to this during my studies. The lack of diversity in our machine learning department left me without a willing or enthusiastic supervisor because my choice of topic was considered too social science focussed. My reluctant supervisor felt uncomfortable taking a risk applying his methodologies outside his area of comfort. Until, of course, it became fashionable. Often times, many in this field, who favour more mathematically rigorous (and fashionable) problems look down on those who take a more solution-driven approach, until their problem becomes sexy enough to draw in finance. And, even then, surely you’d still need someone who understands the problem to work with you on the solution. Apparently not.</p><p>This is the argument for the <em>gender</em> of AI. It <em>should </em>be diverse. As diverse as possible. If AI is a tool used to solve many of the world’s problems, the field must represent those afflicted with the problems AI systems seek to resolve. It doesn’t just stop at diversity of people either.</p><p>Earlier this year Chris Wigley COO of QuantumBlack emphasised the need for good AI systems to be diverse in: People, Mindset, Data and Models. To build a robust AI system requires clear and flexible design thinking, so as to tackle the problem from many angles, and to use a number of different methods and models. Data too is important to ensure that an algorithm’s bias isn’t skewed towards one perspective over another. That too requires a diversity of people.</p><p>Whilst many would argue that diversity imbalance exists in many areas of our lives. That is exactly the point. Let the words of Meredith Broussard be immortalised and serve as a warning not to let AI follow the path of narrow representation that many ubiquitous disciplines and fields that impact our lives have.</p><p>AI is still a nascent field. We can do a lot to change the imbalances and dynamics to build a better future for us all. I, for one, do not want to be dictated by a small group of mathematically minded men, with little regard for social norms and rules, who prefer to deal with machines over women and those of colour.</p><p>What could possibly go wrong? A lot, unless we make changes.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/625/0*OpuvNWqEjv3152yY.jpg" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=97543219d58e" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Babies, Babies, Babies… at web summit]]></title>
            <link>https://rula-awad.medium.com/girls-girls-girls-women-in-tech-2139c94098de?source=rss-4c7afe2bf74b------2</link>
            <guid isPermaLink="false">https://medium.com/p/2139c94098de</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[health]]></category>
            <category><![CDATA[women-in-tech]]></category>
            <category><![CDATA[tech]]></category>
            <dc:creator><![CDATA[Rula Awad]]></dc:creator>
            <pubDate>Thu, 16 Nov 2017 23:19:40 GMT</pubDate>
            <atom:updated>2017-11-20T18:43:48.376Z</atom:updated>
            <content:encoded><![CDATA[<p>Obviously as a Women in Tech attendee I would spend some time on women issues at the event. This topic is one we have probably all thought about (as women my age) or will do at some point in our lives. Fertility.</p><p>3 companies presenting at Web Summit are helping women make more informed choices regarding their fertility. <a href="https://www.preludefertility.com/about">Prelude</a> founded by Martin Varsavsky; <a href="https://www.naturalcycles.com/en">Natural Cycles</a> co-founded by Elina Berglund; and Angie Lee from <a href="https://celmatix.com/">Celmatix</a>.</p><p>Prelude aims to offer services, such as egg freezing, to enable women of any age to have children. According to Varsavsky 1/3 of desired babies are not born due to women not being able to resolve their fertility problems. Prelude solves the fertility problem. Despite Varsavsky having had 6 kids — he had his 7th using Prelude’s services. Natural Cycles, an application I would endorse, helps women track their menstrual cycle. The app’s main purpose is to replace the pill as a form of contraception — it can also work out your most fertile days. Celmatix provides personalised medicine for women.</p><p>What these companies all have in common is the use of data to empower women to make decisions regarding their fertility. Natural Cycles uses the data the user inputs to detect irregularities or signs of trouble in a woman’s cycle. This is usually provides the first sign of infertility if that is the problem. Celmatix has a product called Fertilome — a DNA based fertility test that looks at various genetic markers to give you a likelihood for inherited disorders such as endometriosis or PCOS. The goal is to help women determine the best course of action to take based on their fertility goals and test outcomes. And, Prelude helps facilitate pregnancies.</p><p>For more information on these cool new solutions:</p><iframe src="https://cdn.embedly.com/widgets/media.html?type=text%2Fhtml&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;schema=facebook&amp;url=https%3A//www.facebook.com/WebSummitHQ/videos/1596773767055328/&amp;image=https%3A//i.embed.ly/1/image%3Furl%3Dhttps%253A%252F%252Fscontent-iad3-1.xx.fbcdn.net%252Fv%252Ft15.0-10%252F23475826_1596777213721650_2734490808462868480_n.jpg%253Foh%253Dcd39a1d65cb8f9d5c42826e605990498%2526oe%253D5AA96146%26key%3Da19fcc184b9711e1b4764040d3dc5c07" width="600" height="600" frameborder="0" scrolling="no"><a href="https://medium.com/media/ff28ecd77de7ababff3e4aff5d10a4dc/href">https://medium.com/media/ff28ecd77de7ababff3e4aff5d10a4dc/href</a></iframe><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2139c94098de" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>