The White House Gets an AI Czar

Today on the AI Daily Brief, Trump names the White House Aizar before that in the headlines, OpenAI drops a new model in ChatGPT Pro on their first day of shipments. The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI. To join the conversation, follow the Discord link in our show. Notes. Welcome back to the AI Daily Brief Headlines edition. All the daily AI news you need in around five minutes. Today is one of those days where we kind of actually just have two main episodes. The official main episode is about the appointment of a new White House aizar, but most of these headlines are about the announcements from OpenAI's first day of their 12 days of shipmas. They kicked off their Christmas event with the launch of the full O1 model and a new Pro tier ChatGPT subscription. So let's talk about the O1 model first, the company tweets. OpenAI01 is now out of preview in ChatGPT. What's changed since the preview? A faster, more powerful reasoning model that's better at coding, math and writing. A1 now also supports image uploads, allowing it to apply reasoning to visuals for more detailed and useful responses. When it comes to this multimodality, OpenAI demonstrated the feature by showing the chatbot giving detailed instructions on how to build a birdhouse based on a single image. In a more technically impressive example, the model analyzed the cooling requirements for a space bound data center based on a schematic. The updated version of O1 can deliver faster responses and has a 34% reduction in major errors on difficult problems. An OpenAI spokesperson said that users can expect a faster, more powerful and accurate reasoning model that is even better at coding and math. OpenAI reasoning researcher Noam Brown showed that the model can not only pass the Strawberry test, but can produce a three paragraph essay on strawberries without using the letter E. While performance has generally improved a great deal, the model's performance oddly reduced on some advanced benchmarks. This included mlebench, which measures how well AI agents perform at machine learning Engineering. There is some confusion right now around whether an earlier build was actually benchmarked, so this should all become clear shortly. The updated model is now available for plus and Team subscribers and will roll out to enterprise and education users next week. The second big announcement was the introduction of a ChatGPT Pro Tier subscription at a fairly pricey, at least at first glance, $200 per month. This tier offers unlimited usage of all of OpenAI's models, including unlimited access to advanced voice mode. Jason Wei, a member of OpenAI's technical staff, acknowledged that this tier is not for everyone, stating we think the audience for ChatGPT Pro will be the power users of ChatGPT, those who are already pushing the models to the limits of their capabilities on tasks like math, programming and writing. To that end, the Pro Tier includes access to an even beefier way of using the O model, referred to AS Pro mode. OpenAI said that the Pro Mode quote uses more compute for the best answers to the hardest questions. Basically the difference is that Pro mode allows O1 to reason for a lot longer, with answers taking potentially minutes to return. OpenAI said they intend to experiment with O1 models that reason for hours, days or even weeks to further boost their reasoning capabilities. In evaluations from external expert testers, O1 Pro mode produces more reliably accurate and comprehensive responses, especially in areas like data science, programming and case law analysis. Compared to both O1 and O1 preview, O1 Pro mode performs better on challenging machine learning benchmarks across math, science and coding. In particular, we saw a 75% reduction in errors for easier coding competition questions more Reflective of Everyday programming queries in their release notes for Pro mode, OpenAI highlighted that they had bumped up testing standards to verify the performance boost is reliable. For other models, the company gives a passing grade for each correct answer in a benchmark, but for promode they required the model to get the answer right four out of four times. Now the discussion of Pro Tier subscriptions in Pro Mode is focused on a single question. Who is this for? Alongside the release, OpenAI announced a set of grants to medical researchers at leading universities. They said they plan to roll out grants to other disciplines in the future, and this seems to be part of the intended professionals and organizations that require research grade AI tools. In other words, O1 Pro is probably not the right choice if you just want help with meal planning. But if you want to research gene therapy, it might be the model for you. Professor Ethan Mollick spent all day yesterday experimenting with the new models and and sharing what he had learned. He wrote been playing with 01 and 01 Pro for a bit. They are very good and a little weird. They are not for most people. Most of the time you really need to have particular hard problems to solve in order to get value out of it. But if you have those problems, this is a very big deal. The problems it can solve well tend to be very high value things system design, complex problem solving, analysis for finance or other uses. The value will clearly be higher than the price for the organizations and people who will need to use it. He found that 01 Pro performed well on a range of low value problems like writing poetry or devising an investment strategy of buying ETFs. He also managed to get it to design a Turing machine with logic gates made entirely of swarms of crabs. Inspired by a 2021 scientific paper, Malik summed up his thoughts like this here's my serious tweet on O It can solve some PhD level problems and has clear applications in science, finance and other high value fields. Discovering uses will require real R and D efforts. Few people have PhD level problems for most people just use Claude or ChatGPT or Gemini. It beats Sonnet but not at everything, instead at particular classes of hard problems that Sonnet failed at. Sonnet still dominates in other areas. 01 is not better as a writer, but it's often capable of developing complex plots better than Sonnet because it can plan ahead better. I have had access to 01 for a bit, but I use Sonnet and GPT 4.0 and Gemini A lot more. But when those fail on particularly challenging work, 01 and especially 01 Pro can sometimes crack things that the other models cannot. I'm still figuring out a general pattern and use cases, and I think this is a key story for all of AI. Even people who use AI all the time are still in a use case discovery modality right now. Things that seem obvious are not, and emergent use cases present themselves all the time. In this context, there is simply no substitute for getting in there and getting your hands dirty. Overall, I think there's a lot of enthusiasm, Parm wrote. Zero1Pro is incredible for research. Very, very good. Eric Lancerrez writes, O1 Pro is impressive. The responses don't feel like simple word associations anymore. For the first time, I feel it really understands the nuances and thinks things through. Stuart reed says if O1 Pro can help quants or ML engineers solve problems even 5% faster, then it's a bargain. At $200 per month, that's a miniscule fraction of what their salary is. Danielle Fong of Lightsail Energy just hired a new intern at $200 a month. They're cracked, no doubt, but I'm suspicious. They might be working many jobs, and that kind of seems to be the point. The Pro tier seems aimed squarely at people who have specific use cases they want to tackle where the costs are justified. And while some are worried that this is the start of much more expensive AI products, Adam Silverman thinks it's good for the industry, making sure that the business model actually works for continued advancements, he writes. I hope OpenAI charging $200 a month for pro will be a catalyst for AI and aging companies to start charging more for their products. Overall, pretty cool first day for Shipmas. And like I said, while that is mostly the main part of the headlines today, the one other story I did quickly want to mention just for completeness, Elon Musk's XAI has closed their latest funding round, taking in 6 billion in fresh capital. That brings their fundraising for the year to a very healthy 11.4 billion. That's a little over half the total raised by OpenAI since the launch of ChatGPT, and just shy of Anthropic's total fundraising efforts as well. According to an SEC filing, 97 investors took part in the Series B round, with the lowest stake being $77,593. We don't have confirmation of the other details as there is no accompanying press release, but the point is XAI enters the year flush with cash and presumably ready to ship. That's going to do it for today's Headlines edition though. Next up, the main episode. Today's episode is brought to you by Vanta. Whether you're starting or scaling your company's security program, demonstrating top notch security practices and establishing trust is more important than ever. Vanta automates compliance for ISO 27001 SOC 2 GDPR and leading AI frameworks like ISO 42001 and NIST AI Risk Management Framework, saving you time and money while helping you build customer trust. Plus, you can streamline security reviews by automating questionnaires and demonstrating your security posture with a customer facing trust center, all powered by Vanta AI. Over 8,000 global companies like LangChain, Leela AI and Factory AI use Vanta to demonstrate AI trust and prove security in real time. Learn more@vanta.comnlw that's vanta.comnlw Today's episode is brought to you, as always, by Superintelligent. Have you ever wanted an AI daily brief but totally focused on how AI relates to your company? Is your company struggling with AI adoption either because you're getting stalled figuring out what use cases will drive value, or because the AI transformation that is happening is siloed at individual teams, departments and employees and not able to change the company as a whole? Superintelligent has developed a new custom internal podcast product that inspires your teams by sharing the best AI use cases from inside and outside your company. Think of it as an AI Daily Brief, but just for your company's AI use cases. If you'd like to learn more, go to BeSuper AI partner and fill out the Information Request form. I am really excited about this product so I will personally get right back to you again. That's bsuper AI Partner welcome back to the AI Daily Brief. The rumors that Donald Trump was planning to appoint an AI Czar to oversee AI policy have come to fruition. They are true, and we now know who that person will be. Now, these reports started surfacing a couple of weeks ago. First, there was reports that there would be a crypto czar to oversee the crypto agenda. But then it came out that Trump was also thinking about appointing an AI leader in a key White House position, and that in fact those two things might be bundled together. In the immediate aftermath of these reports, there was a lot of support for this idea. The center for Data Innovation said appointing an aizar signals that the incoming administration is placing AI at the forefront of its agenda, and rightly so. As the lead for federal AI efforts, the czar should focus on two key priorities to help fulfill the President Elect's economic accelerating adoption and safeguarding US Competitiveness. And indeed, this is what it's felt like this role was going to be all about. In other words, regardless of who was appointed, it seemed like the criteria was going to be someone who would be hard driving at US Leadership in these areas. Well, last night, President Elect Trump took to Truth Social to announce his selection. He wrote, I'm pleased to announce that David O. Sacks will be the White House AI in Crypto Czar. In this important role, David will guide policy for the administration in artificial intelligence and cryptocurrency, two areas critical to the future of American competitiveness. David will focus on making America the clear global leader in both areas. He will safeguard free speech online and steer us away from big tech bias and censorship. He will work on a legal framework so the crypto industry has the clarity it has been asking for and can thrive in the U.S. trump's announcement post also noted that Sachs is going to be the head of the President's Council of Advisors on Science and Technology, which serves as a policy advisory group consisting of private sector and academic experts across a wide range of scientific and tech disciplines. Of course, when it comes to the AI and crypto czar role, given that it's brand new, we don't know exactly what it will entail or how much authority Sachs will actually have. Some reports have suggested that the main function would be to act as a liaison between Congress, regulatory agencies and the Oval Office, ensuring policy is coordinated across government. But as of right now, that's speculation. A lot of the discourse following the announcement is all about Sachs himself and what we know about his takes on these particular issues. Sachs is actually quite familiar to lots and lots of people, and that's specifically because he hosts the all in podcast. All In AKA one of the very few technology shows that is consistently ahead of the AI Daily Brief on the charts is a conversation between Sacks, Chamath Palihapitiya, Jason Calacanis and David Friedberg on that show. Because of the nature of the show's content, we've had a long term chance to see how Sachs and his co hosts think about a wide range of issues. Philosophically speaking, Sachs is definitely aligned with the little tech agenda that was articulated earlier this year by Marc Andreessen and Ben Horowitz. In their original post, they wrote little tech is our term for tech startups as contrasted to big tech incumbents. Little tech has run independent of politics for our entire careers, but as the old Soviet joke goes, you may not be interested in politics, but politics is interested in you. We believe bad government policies are now the number one threat to little tech. We believe American technological supremacy and the critical role that little tech startups play in ensuring that supremacy is a first class political issue on par with any other. Sachs, for his part, has long warned of the excessive power of big tech companies, particularly in relation to free speech, and while his resume is likely already well known to this particular audience. To briefly recap, Sachs started his Silicon Valley journey as the founding COO of PayPal. You might remember this photo from a few years ago about the PayPal mafia, and you can see Sachs over here just behind Peter Thiel. During the social media era, he created an enterprise communications networking tool called Yammer, which was sold to Microsoft for over a billion dollars. He was extremely active as an investor during the SAS wave through his firm Craft Ventures, and one of his big plays for the AI era was a Slack competitor called Glue. Generally, SAC's strength has been as an operator and business advisor in the startup world rather than as a technologist. He was one of the first to suggest extreme belt tightening in early 2022. This ensured his portfolio companies could extend their Runway to wait out the difficult funding environment that followed. Essentially, he has a strong knowledge of the pitfalls and blockers that can limit the success of startups and that could make him well placed to remove those types of blockers from his position in the White House. One thing we don't really know a ton about is how Sachs thinks about AI regulation. He's not a super prominent AI investor and hasn't expressed a lot of strong views as he has around other things. Then again, that might not be an accident when it comes to this appointment. So far, the Trump administration's major policy on AI has been a promise to repeal the Biden executive order with no clear position on what will replace it. And it may be that if Sachs's mandate is to tear down the barriers to US dominance of the AI sector, the role is less about AI and more about cutting red tape. Sachs was certainly one of the earliest and loudest Trump supporters in Silicon Valley. He hosted a $300,000 ahead fundraiser at his house over the summer, which broke the cone of silence around Trump support among leaders in tech. Back in June, he wrote a long post on Twitter X called why I'm Backing President Trump. He went through his reasons, including the economy, foreign policy, the border and lawfare. And ultimately that fundraiser and the conversation that went with it was an important moment in shifting the overton window when it came to tech and Trump now in his role as White House advisor. Reports are that he won't be required to divest of his business interests, as the role is only part time. However, under ethics rules, he will be required to recuse himself from decisions that impact his holdings. Sachs is not an uncontroversial person and there are some who expect very negative things from this. Y Combinator founder Paul Graham over the summer said, I know he's an awful person from things he's done. Crypto trader Ludwig Wittgenstein said, the levels of grips and corruption we're going to see will astound even the most sturdy of real businessmen. The response from many in Silicon Valley, however, was resoundingly positive. Sequoia partner Sean McGuire posted the announcement with the comment, it's time to build. Brandon Brooks, a VC at Ohio based Overlooked Ventures, wrote, people told me I should hate sacks. We should be diametrically opposed. But in reality, David gave me the most respect out of many people in the world of venture, engaged in intellectually honest convos and opened my eyes. Congrats, David Sachs earned AI Entrepreneur Vinduretti writes, congrats to David Sachs. I hope he continues to be a huge fan of open source AI. Amjad Massad, the CEO of Replit, said, this is fantastic. Sachs and his firm Craft have been one of the most useful and thoughtful investors we've had, so I'm excited to see him bring the same spirit to the rest of the country. Even Sam Altman squeezed out a meek endorsement tweeting Congrats to Czar David Sachs with Elon Musk responding with a crying, laughing emoji. For what it's worth. Now, when it comes to this question of whether it really makes sense to have AI and crypto bundled together, crypto lawyer Jake Travinsky had I think the best take here. He wrote, for those questioning if crypto and AI belong in the same policy portfolio, realize that the prime directive for government to follow on both is this. Get out of the way. Yes, the time will come for new laws and regulations, but first David Sack can free builders to build. Then again, maybe the most important point is the one from Mr. Ingo who says David Sacks appointment proves you can do anything you want if you podcast hard enough. Can't think of a better way to wrap this episode. Appreciate you listening or watching as always. Till next time. Peace. Sa.