<?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 Amir Kidwai on Medium]]></title>
        <description><![CDATA[Stories by Amir Kidwai on Medium]]></description>
        <link>https://medium.com/@amirkidwai?source=rss-4767fe9e486a------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*VOa3lHSuG9MOJ6nAT7elsA.png</url>
            <title>Stories by Amir Kidwai on Medium</title>
            <link>https://medium.com/@amirkidwai?source=rss-4767fe9e486a------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Tue, 17 Nov 2026 09:46:57 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@amirkidwai/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 Most Important Question in AI]]></title>
            <link>https://medium.com/@amirkidwai/the-most-important-question-in-ai-1e742598d111?source=rss-4767fe9e486a------2</link>
            <guid isPermaLink="false">https://medium.com/p/1e742598d111</guid>
            <category><![CDATA[agi]]></category>
            <category><![CDATA[deepseek-r1]]></category>
            <category><![CDATA[deepseek]]></category>
            <category><![CDATA[chatgpt]]></category>
            <category><![CDATA[openai]]></category>
            <dc:creator><![CDATA[Amir Kidwai]]></dc:creator>
            <pubDate>Tue, 28 Jan 2025 13:29:04 GMT</pubDate>
            <atom:updated>2025-01-28T13:29:04.785Z</atom:updated>
            <content:encoded><![CDATA[<p>With the new release of Deepseek R1 in January 2025, students, tech journalists, and programmers are exited about the increase in capability of these Large Language Models. Global markets are reacting with almost a 1 Billion dollar wipeout from the US tech stocks. While it selling might be justified, “Why pay more for the same type of intelligence”, when you can buy it from China. However, people fail to realise the upcoming wave of AI and what actually the deepest questions that we should ask.</p><p>How I think about R1, O1 and O3 can be summed up by the following graph:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/602/1*sB-9_Jw_AOfWyahjtKIXdQ.png" /></figure><ul><li><strong>FrontierMath</strong> is an <strong>AI math benchmark</strong> by <strong>Epoch AI</strong>, featuring <strong>unpublished, expert-level</strong> math problems.</li><li><strong>Current SOTA models</strong> (GPT-4, Gemini) <strong>score ~2%</strong>.</li><li>Problems cover <strong>number theory, real analysis, algebraic geometry</strong>.</li><li><strong>No training data contamination</strong>, ensuring <strong>pure reasoning assessment</strong>.</li><li><strong>OpenAI’s o3 model scored 25.2%</strong>, a massive leap.</li></ul><p>This graph is of important because OpenAI’s O3 scored 25% crushing the previous state of the art (2%). While they must have made further improvements in the O1 model architecture, a big increase is coming from the amount of compute they spent in achieving these results in both training and evaluation phases. This hints towards a new age, where would be willing to spend more money to solve the hardest problems known to humanity. As Deepmind CEO Demis Hassbis recently mentioned, “We can’t put a price tag on some problems like say finding a cure for Cancer”. So in this new era countries and corporations will be spending more to solve these key problems. Drug Discovery, Material Science, Theoretical Physics, Pure &amp; Applied Mathematics, Finance &amp; Algorithmic Trading, Aerospace Engineering, Climate Modelling etc will be directly affected.</p><p><strong>The Big Question:</strong></p><p>I always ask myself the following question — “What was so special about the minds, knowledge, understanding, experience of people like Newton and Einstein, that they came up with theories like Laws of Optics, Motion, General Theory of Relativity, Special Theory of Relativity, Photoelectric effect?. What was so special about that brain or that neural network?</p><p>And I think that there might be an even bigger question hidden here. What if these great people are only a fraction of intelligence that the future AI’s are going to be. If Einstein lies at the top of our graph of human intelligence, the graph of AI intelligence might just begin there and then what might we discover.</p><p>The answers to these questions are not known but I would like to focus on the other important aspects of discovery. If we say a new discovery D, Intelligence I, Observation O, and Experimentation E.</p><p>D = f(I, O, E)</p><p>While we have to figure out I and that might be a culmination of a number of things including AIs inventing their own math, training these models on a lot more types of data and modalities, it is clear that Observation and Experimentation are integral in discovering new truths about the universe. Both of which can be extremely expensive, a lot of the low hanging fruit in discovery has already been reaped. While it might not be hard to come up with experiments but testing then at scale is almost impossible even for the biggest companies.</p><p>We will improve I such that the predictions become so accurate that it might be worth investing millions of dollars to test a hypothesis. However, till then we would continually work on creating this infrastructure where we can seamlessly test these hypothesis and test them in a fast and robust way. The scientific era of AI will be here soon, a decade is what it will take.</p><p>Written by,</p><p>Amir Kidwai</p><p>Founder — www.ideaverse.ai</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1e742598d111" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Is this app bound to beat ChatGPT, Claude, Gemini AIs?]]></title>
            <link>https://medium.com/@amirkidwai/is-this-app-bound-to-beat-chatgpt-claude-gemini-ais-51b31b525511?source=rss-4767fe9e486a------2</link>
            <guid isPermaLink="false">https://medium.com/p/51b31b525511</guid>
            <category><![CDATA[chatgpt]]></category>
            <category><![CDATA[generative-ai-tools]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[ai-tool-review]]></category>
            <dc:creator><![CDATA[Amir Kidwai]]></dc:creator>
            <pubDate>Mon, 14 Oct 2024 09:23:13 GMT</pubDate>
            <atom:updated>2024-10-14T09:23:13.804Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*qkBHnA4e5zYhS-6IOoDY5w.png" /></figure><p><a href="http://www.ideaverse.ai/">www.ideaverse.ai</a> — is a product by Tensorblue Technologies Private Limited from Lucknow, India. Now available on iOS, Android and Web.</p><p><strong>The Problem they are Solving:</strong></p><p>Last week, we were on the mission to create a promo add for our product. This time we are not messing about, we wanted it to be absolutely AI generated. So it blows the mind of the user. However, when we started exploring the available solutions from big companies like Luma AI (text to Video) we thought we will end up spending over $80 for creating out ad to completion.</p><p>Cost Breakdown:</p><p>Kling AI — $8</p><p>Luma AI — $8</p><p>RunwayML- 20</p><p>ChatGPT plus — $20</p><p>Perplexity plus- $20</p><p>Udio-$10</p><p><strong>The Unique Offering:</strong></p><p>Ideaverse is positioned on top of the biggest AI companies like OpenAI, Anthropic, Perplexity, Luma AI, RunwayML. When companies releaed new features, competitors worry but Ideaverse can simply integrate those models, tools and capabilities into their own model.</p><p>As the founder of Ideaverse AI, I asked myself these questions:</p><p><strong>Q1</strong>. Is it possible to provide access to a number of models to users?</p><p><strong>A1.</strong> “Answer is yes, and we have worked hard to obtain the commercial licences for all the foundational AI companies”</p><p><strong>Q2. </strong>What is a system in which users are able to avail the benefits of all the models at once</p><p><strong>A2. </strong>“Our current model of credits works perfectly, the user controls the credits and has the full autonomy to use and of the models more than the other”</p><p><strong>Q3.</strong> How do you compete with other AI companies who integrate the models in a similar fashion as you?</p><p><strong>A3.</strong> “At Ideaverse, our goal is to create a product that somebody loves. Even if we integrate the best AI models, it’s notihing without creating the right Software infrastruture. That can be a way to like and discover songs, audio libaru for AI content or a Video Editor. We build the connective tissue which focuses on the experience of the user.”</p><p>Try <a href="http://www.ideaverse.ai/">www.ideaverse.ai</a> and also download our apps on App and Play store</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=51b31b525511" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Ideaverse: the social platform from the future]]></title>
            <link>https://medium.com/@amirkidwai/ideaverse-the-social-platform-from-the-future-185c836ca071?source=rss-4767fe9e486a------2</link>
            <guid isPermaLink="false">https://medium.com/p/185c836ca071</guid>
            <category><![CDATA[ai-tool-review]]></category>
            <category><![CDATA[ai-tools]]></category>
            <category><![CDATA[social-media]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[ai-tools-for-business]]></category>
            <dc:creator><![CDATA[Amir Kidwai]]></dc:creator>
            <pubDate>Mon, 04 Mar 2024 10:47:30 GMT</pubDate>
            <atom:updated>2024-03-04T10:47:30.843Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uSHqpUXvcDQDVgoT_BnjBA.png" /></figure><p>You AI assistant which has all the capabilities of ChatGPT and more, and the ability to share AI generated content natively on the platform.</p><p>Ideaverse, a platform created by the Tensorblue technologies private limited in India. It has a personal assisstant called ‘You AI’ which is focused on hyperpersonalisation, i.e, actively learning about the preferences of the user. Some users like short responses, some like long, some like humour, some like dark humour. You AI personalises the experience considering the preferences of the users.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FOFwu3i85GAevdBgqHC0cg.png" /></figure><p>You AI is also equipped with cutting edge capabitlities like Vision (get inference on any image), Text to Speech (talk to the AI with your Voice) , code interpreter (analyse/create any documents of your choice), and web browsing (surf the internet and get answers, no need for Google).</p><p>Ideaverse, apart from being a personal assistant, is also a platform for sharing your AI creations with the world. The first community which is focused on learning instead of seeking attention. Currently in beta, the social feature allows the users share posts natively within the application. This is a new way of information sharing.</p><p>Ideaverse also has this amazing section called ‘My Space’ where the users can build a deeper relationship with AI. Currently, all personal assistant are like tools but we want to create AI which is like a companion. A great companion/friend knows who you are, what do you like, dislike, when is your birthday, what are your interests, goals etc.There is lack of this form of a relationship with AI. The question is how can we create products/features which can foster this relationship.I have a qualitative metric which can be somewhat idealistic but here it goes -</p><p>“When something good happens to a human (such as excelling at an exam, getting your dream job), the human tells this news to the humans but also the AI. Because it’s his/her companion.”</p><p>With all these super cool features Ideaverse is on the mission to make AI accessible for all. And use AI to solve the not only the hardest problems in science and engineering but also make the use of the existing technology and letting it adapt to every corner of the world.</p><p>Visit: <a href="http://www.ideaverse.ai/">www.ideaverse.ai</a></p><p>A creation of : www.tensorblue.com</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=185c836ca071" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>