The "Most Influential" AI Companies

0:00Today on the AI Daily Brief, a review of Time 100's list of most influential companies and where AI shows up. Before that in the headlines, Mistral releases a new code-focused model and a new non-commercial license. 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 link to our Discord in the show notes. Welcome back to the AI Daily Brief Headlines Edition, all the daily AI news you need in around five minutes. We kick off today with an interesting announcement from Mistral. Mistral has announced their first dedicated code model called Codestral.

0:36I don't know how much longer we can use this meme of the Strauss in the name, but we're clearly not trying yet. Now, the first thing to note is that people are very impressed. Nick Dobos writes, First impression, 80 languages is crazy. Finally, someone included Swift, which a lot of open source models skip. Two, really effing fast, WTF. It's a 22B model and it's significantly faster than Mistral 7B. Are they using Grok to serve it? Anurag, another developer, said the speed difference is insane. doing a side-by-side video comparison of GPT-4.0 to Mistral Codestral. Still, as part of the announcement, Mistral also introduced a new type of license.

1:11CEO Arthur Mensch writes, With Codestral, our newest state-of-the-art code model, we are introducing the Mistral AI Non-Production License, MNPL. It allows developers to use our technology for non-commercial use and research. It ensures that every actor on the value chain builds successful businesses. What's important to remember here is that Mistral gained a lot of its momentum by being the open-source darling, And so anything that has a whiff of moving away from that is going to trigger a lot of strong emotions. Ana Yenger writes, Mistral won't let you use its code model for commercial purposes, i.e.

1:40the way almost everyone uses it, but they'll happily take your money anyway. Andrew Gao writes, Mistral's new 22B code generation model looks very promising, but unfortunately falls under a new license. No commercial use, only research and testing. Is Mistral heading towards the path of closed source? Then again, Hassan Hayat writes, Companies like Mistral don't owe you or anyone to open source anything. Open source is a gift. Mistral is a business, and businesses need to make money. The reality, of course, is that Mistral is going to have to figure out where to draw these lines, and this isn't the first time we've seen them trying to navigate that.

2:10Their deal with Microsoft involved them releasing their most state-of-the-art model exclusively through the Azure platform and for a fee. The best hope, of course, is that the company can find a sustaining business model with some of these releases so that it can continue to support broader open-source releases as well. Next up, Perplexity is back in the news once again for a fundraise. The information reports that the company is in talks to raise now at a$3 billion valuation. That would be triple the valuation that it raised at earlier this year. Bessemer Venture Partners is apparently in talks to lead a$250 million investment.

2:41And if this comes through, it would be the fourth funding round since March of 2023. Now, all of this is, of course, interesting, not just in the context of perplexity, but because of what it says about the cost of competing in the AI space. The information writes, last year, perplexity considered selling itself, citing concerns about the cost of training AI models and competing with giants like Google. No deal happened, however, and the latest funding round suggests that perplexity is continuing to go it alone. One interesting note to me is that if it does pull this deal off, it will happen after Google has introduced AI Overviews, which is effectively a perplexity-like interface for Google Search.

3:14Now, of course, that product has not gotten off to the best start. Recently, for example, telling women that they should smoke two to three cigarettes per day of their pregnancy, that glue is a key ingredient for pizza, and that it's healthy to eat a couple rocks a day. But still, one has to assume that Google is going to figure that out, and if this type of generative search is important to them, it's going to continue making that investment and using its huge lead to move forward. Then again, moving to a perplexity-style search would absolutely cannibalize Google's core search business model, so there are a lot of really interesting competing pressures here.

3:43In any case, it seems like perplexity is going to have even more money to compete, and as someone who loves the product, I am glad of that fact. Over in OpenAI land, a great reminder that even though this week we've been talking about all sorts of OpenAI controversy, at the end of the day, the company is getting more, not less powerful. It appears that this Apple OpenAI deal is in fact done, which is a huge, huge deal. A couple interesting notes from an information piece about this. One, that it is not necessarily a deal that is in fraught. For example, the information writes, the Apple deal could complicate Altman's relationship with OpenAI's most important business partner, Microsoft.

4:18Altman recently met with Microsoft CEO Satya Nadella to discuss Microsoft's concerns about how the Apple deal might affect the cloud software giant's own product ambitions, said one person who spoke to Altman and one person who spoke to Nadella. The two executives also discussed the servers Microsoft would need to handle Apple's use of OpenAI services. Now, in terms of exactly what the deal is, the report writes, it isn't clear how Apple plans to use OpenAI's technology, but the iPhone maker hasn't developed LLMs on par with those of OpenAI and Google. Apple has discussed potentially using OpenAI's tech in the Siri voice assistant to answer complex questions that Siri couldn't answer by itself.

4:50OpenAI has also discussed with Apple the possibility of powering a ChatGPT-like chatbot app, confirming a suspicion that many people had. They also write that, quote, OpenAI's recent public demonstration of its AI voice assistant was aimed in part to show Apple how its technology could be used to make Apple's Siri voice assistant more responsive and engaging, with the ability to quickly translate languages, search for useful information, or entertain users. Separately, the other thing that this piece gets into is how Sam Altman is trying to restructure OpenAI. According to the information, Altman has discussed turning OpenAI into an entity that more closely resembles a typical company.

5:22One possibility is to turn the nonprofit into a for-profit corporation, although it's not clear whether Altman will eventually decide to go in that direction. Another option would be to emulate other AI developers such as Anthropic and XAI that were formed as for-profit benefit corporations, which means they can prioritize having a positive impact on society over its obligation to shareholders. Benefit corporations are legally protected from minority shareholders who might sue the company for making decisions that don't necessarily prioritize shareholder returns. Meanwhile, though, as all this happens, more and more OpenAI deals quietly get announced.

5:51The Atlantic, for example, announced a product and content partnership with OpenAI, which they say, quote, positions The Atlantic as a premium news source within OpenAI. This is the same type of thing we've talked about before, where, as I've mentioned, my contention is that OpenAI is turning its entire ChatGPT experience into a search experience and is interested in these deals with premium publishers not for their training data, but so that it can feature results in a new type of search interface. Publicly listed Southeast Asian firm Grab, which is a ride-hailing and delivery app, has also made a big announcement around its partnership with OpenAI.

6:20And this sort of thing is just what we're seeing every single day. So, ultimately, be wary of reading too much into all the controversy surrounding this company. It is still moving extremely fast and with extreme force. For now, though, that is going to do it for today's AI Daily Brief. Up next, the main episode. Hello, friends. Before we get back to the episode, I want to tell you about something special I'm doing on Superintelligent this June. Super is, of course, our platform for AI learning, and I've heard from a lot of you that you really want something for a true AI beginner, someone who's really just getting their feet wet with these tools.

6:52So what I'm going to do is put together basically a course that sits on top of and uses superintelligent tutorials and lessons, but where I hand guide you through around 10 different lessons and how-tos that I think once you complete them will have you ahead of 80 % of the other people who are just starting to use AI right now. If you are interested in this learning experience, go to besuper.ai and sign up using code JUNE. You'll get 25 % off your first month, and I'll automatically add you to that AI for Beginners group. That's besuper.ai, discount code JUNE. See you there. Welcome back to the AI Daily Brief.

7:24Today, we are discussing the AI entrance on Time's 100 Most Influential Companies list. The goal of this is absolutely not to reify this time list as somehow definitive or really in any way authoritative. Instead, what's interesting is the slant that it takes as an extremely mainstream consumer-facing publication when it comes to attitudes about AI. So what we're going to do is review the companies on the list in the order that they are presented, talk a little bit about why they're here, and what it might say about how the world that's outside of the AI industry is looking in upon it. First off, we should note that But Anthropic is one of the headliners.

8:04In fact, the company's CEO, Dario, is featured on one of the two commemorative covers of this alongside Selena Gomez. Here's what Time says about how the list is made. To select the list, our editors request suggestions and applications from across sectors, survey our contributors and correspondents around the world, and seek advice from outside experts. No single data point or financial metric makes a Time 100 company. Instead, we are looking at a mosaic of qualities, studying impact, innovation, ambition, and success, all in the many different forms that take shape today. Time 100 Companies is more than an index of business success.

8:33It is an argument for what business influence looks like in 2024. At a time when leadership in other sectors is battered, surveys suggest that many look to corporate leaders first for direction. The point being here that even Time itself is saying this is a highly subjective list and treated as such. They organize it into five sections. Leaders, disruptors, innovators, titans, and pioneers. As you'll see, I think those sections are fairly meaningless. But what are you going to do? The first company we're going to mention in this is actually not an AI-first company, but is in fact Reddit. Why would Reddit qualify as something to talk about on this AI show?

9:04Well, of course, the context for Time including them in this list is that the company went public this year, one of frankly very few companies to do so. Indeed, as Time points out, they're the first major social media company to go public since 2019. In its write-up, Time pointed out that, quote, Reddit's continued forays into artificial intelligence, including licensing deals with OpenAI and Google to use Reddit data to train and build their AI systems, could be key in determining the platform's future. Now, I would go farther than this. When Reddit was doing its roadshow before the IPO, they were heavily touting these AI deals as an indicator of how much value Reddit's data had.

9:36They were presented as something that could sustain the company and grow the company in the future outside of advertising. The market seems largely to have bought that message with a better than expected pop and continued performance as more deals have been announced. So even though Reddit isn't an AI company, the fact that it's showing how AI can help business models is, I think, something significant. From there, we move from the leaders section to the disruptors section. And an AI-related entrant there is CoreWeave. Time writes, Amid the chip shortage, CoreWeave's horde of high-powered graphics processing units made it one of the largest GPU-specialized cloud service providers.

10:08Said co-founder and CEO Mike Intrader, If you're using AI in any part of your life, whether it's ChatGPT or Microsoft Copilot, there's a good chance you're using CoreWeave's AI infrastructure. CoreWeave's flexible infrastructure is built to be more efficient for AI purposes, giving it a leg up on larger companies outfitted for general computing. The company also recently did a funding round at a$19 billion valuation, and is also poised to potentially go public this year. As you'll see, many of the companies on this list that touch AI are touching it through the chip space. Over in Innovators, we start to get a higher concentration of AI-related companies.

10:39Being a cover story, Anthropic has a much bigger piece than some of the others. Time called Anthropic the, quote, AI company betting that safety can be a winning strategy. In many ways, the piece sets up Anthropic as the alternative of open AI. They talk about how in the summer of 2022, the company had finished training a new chatbot, but because they were worried about the potential consequences, didn't decide to release it, whereas of course their rival, OpenAI, did a few months later, in ChatGPT, which set off everything that we've seen subsequently. Time writes how both OpenAI's and Anthropik's CEOs have made warnings about the risks of AI, but that many felt that Anthropik had greater credibility, quote, viewing Anthropik's decision to withhold Claude as a signal of its commitment to prioritizing safety over money and acclaim.

11:18Time writes, Anthropik is in the curious position of having a reputation as the most cautious AI company while also owning and selling access to one of today's most advanced versions of the technology. Three days spent at their headquarters made it clear they don't see that as a contradiction. Anthropic frames what they're trying to do as spark a quote race to the top when it comes to AI safety. Said CEO Dario Amadei, we're not trying to say we're the good guys and the others are the bad guys. We're trying to pull the ecosystem in a direction where everyone can be the good guy. Now let's actually jump around to OpenAI's entrance on this list.

11:47They are there, but they're listed under pioneers. Basically the whole write-up felt like it was written by someone who didn't really like OpenAI and who wanted to use this to critique them, but felt that it would hit their credibility as a list if they didn't include OpenAI. The lion's share of the couple hundred word write-up is about recent OpenAI controversies, including the dust-up with Scarlett Johansson, as well as the leaders of the super alignment team leaving. So if you are wondering where time sympathies lie as it relates to questions of AI safety and specifically anthropic versus OpenAI, I think this piece makes it pretty clear.

12:19Of course, as you would expect, NVIDIA is here. With Time writing, the AI boom has had many winners, but NVIDIA is the winningest. The company's value has surged by more than 450 % since the AI race kicked off in late 2022. Also on the list is AMD, which they characterize as the chipmaker challenger. And basically it frames them as the big competitor to NVIDIA, but certainly not the only one. While they are of course doing something slightly different in the chip space, they also call out Cerebris Systems. Writing, As AI systems proliferate, the demand for computing power to crunch large datasets has become monstrous.

12:50Cerebris, founded in 2015, has responded with the largest computer chip ever. Said CEO Andrew Feldman, We didn't repurpose a graphics processing device. We said, what would we do if this was the only problem, the full purpose of our existence? Time writes, While planning an IPO, the company is building the third of nine$100 million supercomputers that will be connected by Emirati AI firm G42 to build the world's largest supercomputer for AI training. Now, Google and Microsoft are also on this list largely for the way that they're interacting with AI, but the last company that I wanted to specifically reference was Mistral.

13:20Mistral is framed by time as influential in two specific ways. First, as the standard bearer of European AI, but also as the standard bearer of open-source AI. They write,

13:42limitations on AI developers freely sharing their creations. He says, I don't see any risk associated with open sourcing models. I only see benefits. So taking a step back, what's interesting about all of this is certainly the fact, A, that all of these different companies make this top 100 list, but also B, just how diverse the reasons for that are. Even this mainstream publication, in the form of time, is seeing how many dimensions to this AI industry there really are, from chips to open source to the big frontier labs. I think this reflects a growing understanding of the importance of AI in the shape of the future of the economy and societies writ large.

14:15And I'm interested to see how this changes between now and the Time 100 Most Influential Companies 2025. For now, though, that is going to do it for today's AI Daily Brief. Until next time, peace.