The 5 Most Important Stories in AI This Week

Today on the AI Breakdown, we are counting down the five most important stories in AI this week. The AI Breakdown is a daily podcast and video about the most important news and discussions in AI. Go to breakdown.network for more information about our Discord, our newsletter and our YouTube. Welcome back to the AI Breakdown. Today we are doing something a little bit different. I want to see if this is a format that you guys like, maybe for some Friday or Saturday shows. What we're going to do is we're going to count down subjectively, just this is based on my list, the five most important stories in AI this week. I'll tell you what the story was and then give a little bit of context about why I think it was so significant and what the takeaway is. We kick off with an honorable mention, which is the introduction of the O1 light by open Interpreter. This is another physical hardware device that's meant to allow people to interact with AI in a physical, real world sort of way. We've of course seen the rabbit R1, or at least we've seen it previewed. And then there are things like the Rewind pendant, which again is just theoretical so far, and the tab which has been demoed. And all of these are trying to create a different type of relationship between us and our devices effectively. The excitement for these products is that in a world where AI agents can control our computers, we we can theoretically just control those agents by speaking to them. That could free us from the tyranny of screens and allow us a more natural way of interfacing with the world that's not mediated by these phones in our pockets. Open Interpreter has been doing some really interesting things with the actual software side of their product, which runs locally on your computer and acts as an agent for interacting with your computer. And the O1 Lite is, as they call it, a portable voice interface for that device. Now they say this product came together in the last six weeks and was driven by wanting to have an open source version of these type of physical hardware devices. And there was definite a lot of interest. Killian, the founder of Open Interpreter, wrote yesterday that Batch one had sold out in two and a half hours and the profits would be redistributed to open source contributors. So this is an honorable mention because although the device doesn't exist yet, it certainly shows the excitement around this particular trend of AI hardware. Now moving on to our main list, at number five, we have Reddit's ipo. Reddit is obviously not an AI company and so why is this on this list. Well a big part of Reddit's pitch to investors had to do with its ability to sell its data to AI companies. It's been a classic conundrum for big social media companies how best to monetize. While Reddit has done well to build a robust ad business, one of the things that they're pushing as a really valuable asset, a unique asset, is all of their data, which is now increasingly valuable to frontier model developers. As evidence of the value of their data, Reddit has been pointing to a deal with Google that will give Google non exclusive access to Reddit's Data for training AI models for $66 million for the year. That'll amount to something like 7 or 8% of their revenue, with the pitch being that there's more to come. Now before the IPO actually happened, the Information had published a piece suggesting that investors were a little bit skeptical of this pitch. But then things went off and Reddit was up almost 50% on the day. And as Bloomberg writes, much of investors enthusiasm is attributed to that AI business. Writes Bloomberg, the AI revolution was at the center of Reddit's proposed value proposition to investors as companies eye the record setting rallies in stocks like Chipmaker and Video. Now whether this all plays out exactly the way that Reddit hopes remains to be seen. Obviously there are some big questions around how data, sales and access are going to go. That is not just going to be a business question, but a legal question. But in the meantime, one of the big beneficiaries of Reddit's pop is Sam Altman. When Reddit was being bought back from Conde Nast, Altman was one of the investors who helped facilitate that deal. He now controls 7.6% of the company, which based on the IPO is now worth $613 million. Number four on our list of the top stories today are deals with Apple. The week kicked off with reports that Apple was in deals with Google to potentially bring Google's Gemini AI to the iPhone. Now we've seen over the last couple weeks that Apple has been getting more serious about its AI strategy. It ditched its car project, Project Titan, and apparently shifted a bunch of those resources over into the AI space. However, people were pretty surprised to see them talking with Google about handing over premium real estate to Google's AI. There was lots of speculation around what it meant. Was it reflective of the fact that Apple was maybe taking specific bets on AI that it wanted to control, such as more narrowly defined use case based on device AI, perhaps for Siri, while letting everything else be handled by Gemini or was there something else going on? This news was greeted with enthusiasm by Wall street, but a lot of skepticism among the AI set. The end of the week saw another similar report suggesting that Apple has held talks with China's Baidu over AI for its devices in that country. The Wall Street Journal writes, in China, Apple has been looking for a local generative AI model provider, mainly because China requires such models to be vetted by its cyberspace regulator before being launched to the public. Now this one, I think makes a little bit more sense based on the particular regulatory situation in China. But once again, it could also be reflective of a larger strategy which is just coming into focus, where Apple really does use partnerships to bring cutting edge AI to the iPhone. Number three on our list. Today, intel and the US government have announced a major deal by which intel will get $8.5 billion in grants and have access to up to another $11 million in debt funding for chip plant construction and renovation across four locations in the United States. A couple years ago, the US passed the CHIPS act, which is an attempt to bring semiconductor manufacturing back to the US this had been a growing point of discussion for a while, but Covid era supply chain disruptions really put a fine point on the fact that that it may not have been the best idea to totally outsource the supply of some of the most important components of modern life to countries in China's sphere of influence. The CHIPS act has a number of different provisions, but a big part of it was $53 billion that was earmarked for actually facilitating this type of construction. And so far, this $8.5 billion grant is the biggest one we've seen yet. Intel will use it to expand projects in Arizona, New Mexico, Ohio and Oregon. And the goal of this, roughly $20 billion combined between the grants and loans is, is to catalyze more than $100 billion in private investment into these facilities. Intel is very ambitious with this project. They've made it clear that their goal is to see half of global chip production come back to US shores by 2030. It's the type of very ambitious goal that actually gets people excited. And the enthusiasm that I'm seeing around intel is something very different from the past. Now, we talked about this a little bit when this was first announced, but part of what makes Intel's strategy interesting is that they're moving towards a foundry model similar to tsmc, where instead of just producing their own chips, they're producing chips to other specifications. The Biden administration is touting this one mightily. And so I'm interested to see what sort of momentum it puts around these types of chip reshoring efforts. Number two on our list this week, obviously the biggest planned event was Nvidia's GTC conference. Tens of thousands of people came to this event. Indeed, it had more of the energy of, if not a rock concert than certainly an Apple consumer type keynote than a chip and infrastructure event. But I think that's reflective of how much Nvidia has become the absolute epicenter of the AI revolution. At the event, Nvidia announced its new Blackwell system, including the B200 tensor core chip, which has 208 billion transistors and is significantly more powerful than their last generation. A lot of the discourse around this has been around how significant the battle for COMPUTE is going to be and how quickly companies are moving to accelerate our compute access. Now, there were a ton of other really interesting announcements at Nvidia. GTC as well as Project Groot is a moonshot initiative to build AI that can be used for a variety of humanoid robots. They have an Omniverse digital twin system that they were demoing. Really, it seems like it was quite the bonanza and we'll be talking next week with at least one person who was there, so we'll have a little bit more of a direct perspective. Despite the fact that everyone was sort of anticipating all of the announcements that Nvidia made, Wall street was still excited, writes Daniel Howley, the technology editor at Yahoo. Finance. I've been in my fair share of tech events over the years and this was the first time in quite a while that attendees gave off an overwhelming sense of excitement and anticip anticipation for what was to come. Nvidia is currently trading around $928 a share, a gain of about 1.5% on the week. Finally, we move on to our number one news story for the week. One that if Nvidia was expected news, this was very much the opposite. Mustafa Suleyman, the co founder of DeepMind and then co founder of Inflection AI, which of course manages the PI personal AI chatbot, has been recruited to go lead a new consumer AI division at Microsoft. He brings with him another co founder at Inflection, along with a significant portion of the team there. Now this was very unexpected. Inflection was widely seen as one of the very small number of companies that had access to enough resources to actually compete, and only a couple weeks ago they were touting Inflection 2.5, their latest model. The company had raised more than $1.3 billion less than a year ago. And so people have been trying to figure out what's going on ever since. I would say that by and large, people's assessment of the situation is that it's a reflection of just how hard to compete in the frontier model space it is that even with billions of dollars of access to compute, it still might not be enough. Now, in addition to whatever it says about inflection, AI and startups competing in that area, people also think that there are implications here for Microsoft's relationship with OpenAI. Broadly speaking, it seems like people are viewing this as another indicator that Microsoft is diversifying its approach to AI and relying less on whatever OpenAI does. To be honest, this feels like a tectonic plate shift. And so it's a situation that I'm going to be watching very closely. All right, guys, there you have it. That is your top five AI events of the week. Let me know what you think about this format. Appreciate you listening as always. And until next time, peace.