A Manhattan Project for AI Infrastructure

Today on the AI Daily Brief, we are talking about what appears to be a potential Manhattan Project for AI infrastructure buildout. 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. Quick note. Once again, I am traveling this week so the episodes are going to be a little bit different, at least from a format perspective. And today there is so much news that we're going to do an extended headlines edition where we touch briefly on just about a dozen different stories. For those of you wondering if the summer lull is over, boy is it over. We kick off today with the latest report about Sam Altman's big ambitions on an infrastructure buildout for artificial intelligence. Bloomberg reports that OpenAI pitched the White House on an unprecedented their Word data center buildout. Basically, it appears as though OpenAI has been pitching the Biden White House on the fact that we are going to need a huge amount of power to develop advanced against AI. You might remember in Altman's piece that we read recently, the Intelligence Age, he talked about how if we don't do this sort of build out, wars will ultimately be fought over access to AI and it will just be a tool for the rich to get richer. Now apparently after the meeting, OpenAI shared with government officials a paper that outlined their plans and the benefits of building 5 gigawatt data centers in various US states. Now if you are like many of us and don't know what a gigawatt actually represents, 5 gigawatts is roughly the equivalent of five nuclear reactors, or put another way, produces enough power to power 3 million homes. The benefits that OpenAI listed from this weren't just about maintaining the lead in AI development, but included things like creating thousands of new jobs, boosting GDP and more. For months now we have been hearing About Altman and OpenAI's efforts to build a global coalition around this sort of massive infrastructure buildout. Way back at the beginning of this year, the number 7 trillion had been thrown around in terms of the amount that Altman was looking to raise, and recently it's appeared that the efforts have been more focused on starting in the United States. A spokesperson for OpenAI somewhat reinforced this message, providing a statement to Bloomberg that said OpenAI is actively working to strengthen AI infrastructure in the US which we believe is critical to keeping America at the forefront of global innovation, boosting reindustrialization across the country and making AIs benefits accessible to everyone. In many cases. Old analogies like moonshots or the Manhattan Project have become cliches, but this really is one of those areas that warrants the cliche. The rest of the Bloomberg article basically reads with experts explaining why this isn't possible, which if anything, is just likely to make Silicon Valley want to do it even more. Now for Those of us ChatGPT users, perhaps the more exciting OpenAI announcement was that Advanced Voice Mode has finally rolled out to all plus users. I have been beating my head against the wall for months waiting for this thing and I just noticed as I pulled up ChatGPT yesterday that finally it was here. I have not yet had a chance to play around with it, but I'm excited to dig in now. One little note in his tweet late last night, Altman said rollout completed early. Amazing work by the team, except for jurisdictions that require additional external review. Referencing that Dean W. Ball wrote under a strict reading of the AI Act, ChatGPT Advanced voice is illegal in EU workplaces and schools because the system can recognize a user's emotions. That's prohibited by the AI Act. Many jumped in to point out that this is a great example of how regulations can have unintended consequences, because it seems very unlikely that EU regulators were thinking about something like ChatGPT advanced voice mode when they wrote these rules. Yesterday, we also got some model updates. Google DeepMind announced that they were releasing two new production ready versions of Gemini 1.5 Pro and Flash. They continue for tasks like summarization, question answering and extraction. The default output length of the updated models is now around 5 to 20% shorter than previous versions, making them easier to use now. Unlike the recent 01 announcement from OpenAI, which was really about a sea change in how artificial intelligence was delivered, these seem to be much more practical updates. Some weren't thrilled. AI entrepreneur Bindu Reddy writes Sai OpenAI releases 01 that aces IQ tests and Google releases some minor update to Gemini 1.5. How is this possible when they have 100x the resources, 10x the talent and 10x everything? However, not everyone saw it in that same way. Another AI entrepreneur, Saliomar, wrote, been using the new Gemini Flash models and it's good. Like really good. Probably the best low cost model by a wide margin with the right prompts. It's better than GPT4O and Sonnet at large context reasoning but not coding. And it turns out that this low cost thing may be exactly what this update is really about. In fact, Bindu Reddy came back once again later and said Gemini's real superpower. It's 10x cheaper than 01. Predictably, the new version is better than the old version, and Gemini now trails behind Zero1, Sonnet and GPT4.0, but is significantly better than Llama 405B. The real story, however, is the cost. Gemini is 10x cheaper than Zero1 Preview. The price to performance ratio makes it pretty competitive to 4.0, and so perhaps this update was not about catching up to 01 from a capacity standpoint, but but about being a practical option for developers who are conscious of cost Today's episode is brought to you by Venice. Venice is a private, uncensored, generative AI app. It accesses open source models to enable text, image and code generation without the fear of being spied on or having your data exploited. Discuss anything with Venice without concern about it being monitored, sold or given to advertisers and governments. Venice is different because your conversations and creations are kept securely within the browser, never stored or accessible by Venice. Unlike other AI apps, Venice won't tell you what's okay to say or not. Venice won't patronize you. It simply provides direct access to machine intelligence. No topics or off limits, no ideas are taboo. With Venice, you're in control of the AI as you should be. Pro subscriptions are available for $49 a year or $8 per month. AI Daily Brief listeners receive a 20% discount on Venice Pro. Visit Venice AI NLW and enter the discount code NLW Daily Brief. That's NLW Daily Brief, all one word. Today's episode is brought to you by Superintelligent, which is of course our platform that helps you learn how to use AI tools and perhaps even more importantly, gives you ideas on the best use cases that are actually going to help you achieve whatever it is you want to achieve. To recognize the end of summer and back to school, back to work, we are running our best promotion ever. When you sign up for Super Intelligent Using Code So back, your first month will be 100% free. The platform features over 600 fun, highly practical AI tutorials that get you using AI fast and with an eye to actually transforming how you get things done. We've just launched super for Teams. So if you have a group of people at your company that want to figure out how to use AI together, I highly suggest you check it out. But for those of you who are using superintelligent as an individual once again, if you sign up for Super Intelligent between now and the end of the month using Code Soback, you will get your first month 100% free. Go to be super AI and check it out today. Speaking of cost developers and revenue, CNBC reported that Anthropic is expected to hit $1 billion in revenue this year, which would be 1000% increase year over year. Of that, around 60 to 75% of their revenue is coming from third party API. Another rumor swirling around Anthropic are reports that the company is in talks about a new round with investors in a deal that would value the company between 30 and 40 billion. That would be roughly double the valuation from funding that closed earlier this year. This seemed pretty inevitable given that OpenAI is also out there fundraising, although for now this remains just reports. And speaking of just reports, there are rumors swirling everywhere around right now that we are about to get a new anthropic model, perhaps 3.5 opus. I'm sure that will be the subject of the AI Daily Brief as soon as it happens. Now another big story from yesterday had to do with AI and Hollywood and this was kind of multi dimensional. On the one hand you definitely see an integration. Titanic and Avatar director James Cameron, for example, has joined Stability AI's board of directors. This is a huge coup given how much Stability has struggled recently. In a statement, Stability said that Cameron's Coming on was a, quote, monumental statement for the AI industry as a whole, describing Cameron as someone who, quote, lives in the future and waits for the rest of us to catch up. Now obviously this follows recent news that Lionsgate had signed a deal with Runway, and so the trend line seems pretty clear of a Hollywood integration with AI rather than a dismissal of AI at the same time. The other Hollywood AI news yesterday was that a group of famous actors and directors wrote an open letter urging Governor Gavin Newsom to sign SB 1047. More than 125 actors, directors, producers and others signed the letter, one of whom actually wrote it as well, although we don't know exactly who it was. The group basically said that they did not believe that threats from AI were just the stuff of science fiction anymore, and of course tried to dismiss the people who don't want SB 1047 as just billionaire opponents. A fairly standard reaction, at least from the AI entrepreneur sphere, was summed up by this tweet from Melinda B. Chu, who wrote, joseph Gordon Levitt is an effective altruist like SBF and also Caroline Ellison and a child actor from 3rd Rock from the Sun. He's not an authority on AI. LA Times should also cover techie's opinions on the Oscars or Emmys. This L3 engineer at Google thought Barbie should have won for Best Picture. Barbie was robbed. Print it. The argument, of course, being that this set of people don't really have a technical basis to make factual judgments about these risks. However, ultimately politics is a game of influence, and so to the extent that Governor Newsom views this as a political headache, it could be meaningful in how this all plays out. Just to round out the entertainment industry engagement Sandwich Variety also reports that Deepak Chopra has signed on with 11 labs to lend his voice to their Reader app. Again, whatever Hollywood thinks about AI regulation, it's very clear that these spaces are going to be intertwined sooner rather than later. Over in Chipland, intel has launched its latest AI chip. These are the Xeon 6 CPU and the Gadi 3 AI accelerator. Now this comes hot on the heels of a Wall Street Journal report that Qualcomm is potentially looking to take over intel as a way to bolster its own chip business. At the same time, Apollo Global Management might make a multibillion dollar investment to back Intel CEO Pat Gelsinger's big, bold turnaround plan. Basically, it seems like something is going to change at intel, although it's not exactly clear what right now. A few more small stories before we get out of here. The steady integration of AI into every service that you touch continues. Reddit is now bringing AI powered automatic translation to new languages, with a goal, of course, to make it more accessible to a broader number of users. Five months ago, Reddit introduced sitewide translation for French speakers, and one of the really cool things here is that the way that this is deployed is users from two totally different language groups can have a conversation in the comments or in Reddit threads, really without having to manually translate everything. They can just toggle the setting and be off. Having a conversation without linguistic borders. This, I think, is something that lots of people are excited about, is a huge benefit of AI, and so it's exciting to see it actually come to market. Lastly today, a little one from Aravan Srinivas, the CEO of Perplexity. He noticed that in an interview with famous investor Stanley Druckenmiller, he was talking with Squawkbox's Joe Kernan and said, do you want to hear how I invested in Argentina? It's a funny story. I wasn't at Davos, but I saw the speech in Davos and it was about one in the afternoon in my office. I dialed up Perplexity and I said, give me the five most liquid ADRs in Argentina. In other words, friends, it is not just techie early adopters. These tools are starting to get everywhere. However, that is going to do it today for the AI Daily Brief. Even while traveling, I will continue to bring you the most pertinent and interesting AI news. Appreciate you listening or watching as always and until next time, peace.