The AI Daily Brief: Artificial Intelligence News and Analysis - Will AI Acqui-hires Avoid Antitrust Scrutiny?

Episode Date: July 2, 2024

Amazon bought Adept...sort of. Just like Microsoft soft of bought Inflect. NLW explores the new big tech strategy which seems designed to avoid antitrust scrutiny. But will it work? Concerned about being spied on? Tired of censored responses? AI Daily Brief listeners receive a 20% discount on Venice Pro. Visit ⁠⁠https://venice.ai/nlw⁠⁠ and enter the discount code NLWDAILYBRIEF. Learn how to use AI with the world's biggest library of fun and useful tutorials: https://besuper.ai/ Use code 'youtube' for 50% off your first month. The AI Daily Brief helps you understand the most important news and discussions in AI. Subscribe to the podcast version of The AI Daily Brief wherever you listen: https://pod.link/1680633614 Subscribe to the newsletter: https://aidailybrief.beehiiv.com/ Join our Discord: https://bit.ly/aibreakdown

Today on the AI Daily Brief, we're looking at adept in Amazon's deal and what it says about the state of AIMNA. Before that, in Headlines, 11 Labs, has released a new iconic voices feature that is sure to be controversial. 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. One big question, which neither AI companies nor society has exactly resolved, yet is what feels correct when it comes to bringing back people who have passed on in the form of AI avatars of their entire presence or just their voice. Well, we have a little bit more context

to have that conversation now as 11 Labs has announced their new iconic voices. This is a follow-up to the reader app that they launched on iOS last week and today posted on Twitter, listen to your favorite books and articles voiced by Judy Garland, James Dean, Bert Reynolds, and Sir Lawrence Olivier on our reader app. Now, 11 Labs made sure to be clear. that they have partnered with the estates of all of these people, of course. This is not just them begging for forgiveness rather than permission. Said Liza Minnelli, Judy Garland's daughter, it's exciting to see our mother's voice available to the countless millions of people who love her.

Through the spectacular new technology offered by 11 Labs, our family believes that this will bring new fans to Mama and be exciting to those who already cherish the unparalleled legacy that Mama gave and continues to give to the world. Said Dustin Blank, the head of partnerships at 11 Labs, Judy Garland, James Dean, Bert Reynolds, and Sir Lawrence Olivier are some of the most celebrated actors in history. We deeply respect their legacy and are honored to have their voices as part of our platform, adding them to our growing list of narrators marks a major step forward in our mission of making content accessible in any language and voice.

Now, this announcement came in just before I started recording, and so there hasn't been all that much chatter yet. Morgan tweets appropriately licensed necromancy, and mostly the comments are sort of banal. Ada Pye says, interesting, presumably licensed from the dead actor's estate. I think that this question is one that people are going to have very strong feelings about. The people who are going to be against bringing back the dead in audio or video form are going to be extraordinarily against it. However, I think ultimately that with this capacity existing, it is almost entirely inevitable that it will come to market. From a cynical perspective, it represents an entirely new category of revenue for the estates of famous people.

I find it very hard to believe that that won't be monetized in the same way that every opportunity is monetized. Less cynically, I don't think, for example, Liza Minle is wrong to note that this will be, incredibly beloved by many of Judy Garland's fans. And I think there will be many who argue that you don't have to participate in it if you don't like it. But that also doesn't mean that you have to deny fans who do like it the chance to participate in their way. Where it might get even thornier is the decisions that people make about their own friends and family. Will it become normal to let people interact in some way with a digital ghost of people from their past? It is certainly going to be a capacity that we have, and so I anticipate a fierce debate about whether it's

actually good for us or not. Next up, another product launch announcement, Suno, the music generation AI, is now available as an iOS app. The company tweets make any song you can imagine anytime and anywhere and share with your friends all from your phone. Right now, the app is only available in the US, and it's only available on iOS, but they say that Android and Worldwide is coming soon. Suno is, of course, itself very controversial with the people who hate it, really hating it, and a big lawsuit being announced last week, but at the same time, the people who love it really love it. Riley Brown, for example, a great AI content creator writes, okay, this is lit A. F. Moving over into the world of finance, some interesting comments from Ken Griffin.

The Citadel founder has walked back some AI hype, saying that he's not sure that LLMs will actually be able to replace employees in the next few years. He said, for a number of reasons, I am not convinced that these models will achieve that type of breakthrough in the near future. As evidence for his argument, he cited self-driving cars, which, as Fortune writes, quote, struggle when confronted with so-called edge cases that fall beyond the boundaries of statistically common traffic occurrences. Staying in finance but moving to another area, Robin Hood has announced that it has acquired Pluto Capital. The Wall Street Journal writes, according to the deal announcement, Pluto Capital offers investment research powered by AI,

and quote, delivers highly customized investment strategies based on customer needs and financial goals. Importantly, Pluto doesn't do the same work of robo advisors. In other words, it doesn't actually manage assets or make trades for users. Instead, it's just focused on research. But that hasn't stopped speculation that this is maybe a first step for Robin Hood to get into that sort of investment advisor game. AI in finance is a fascinating and large topic, and one that we will definitely be getting more into here in the months to come. Speaking of finance and AI, big money for a startup at the intersection of AI and decentralization.

Sentient Labs has raised $85 million led by Peter Thiel's founders fund and joined by other Web 3 investors, including Pantara and Framework. Bloomberg writes, Singapore-based Sentient Labs, a month-old AI startup co-founded by an executive behind the Polygon blockchain is building an open-source platform through which developers can get paid for their contributions to emergent AI tools, such as chatbots and enterprise software. The fundraising comes ahead of the Sentient Platform's planned TestNet launch scheduled for the third quarter. Said Sandeep Nailwall, co-founder of Sentient Labs, there's a lot of anxiety around centralized forces running away with AI. Sentient aims to compete with those centralized behemoths, adding

that crypto can, quote, completely disrupt the way AI will play out. Now, so far, this AI crypto overlap has been mostly a discussion point for crypto companies rather than a real force in the AI space, but seeing bigger deal sizes could signal that that might be changing in the future. For now, though, that will remain something that we have to watch. That's going to do it for the AI Daily Brief Headlines edition. Up next, the main episode. Today's episode is brought to you by Super Intelligent, the platform for fun, fast AI learning.

Super has a ton of new things going on. We recently announced our partnership with Spotify, through which users of that app can now access Super Intelligent content directly from their mobile apps. We've also just launched the AI learning feed. In addition to seeing the tutorials that we're dropping, there are polls, news items with related lessons, and a chance for people to show off the projects and use cases that are making AI come alive for them.

We've also just kicked off the Super Summer Challenge, where each week we'll share a new challenge that you can use to discover new AI tools and use cases. Go to be super.ai and use code superfund for 50% off your first two months. That's besuper.a.i. Today's episode is brought to you by Venice. The leading AI companies store your entire conversation history and attach it to your identity forever. Every question you ask, every answer you receive, every image you generate, every thought you share with

the machine, it's all being spied on. If you trust all the companies, hackers, and NSA board members that will ever have access to your AI conversations, then rejoice, for you are well served. For the rest of us, Venice is an alternative. Venice is a powerful AI app for text, image, and cogeneration that respects you as a sovereign individual and believes privacy and free speech are not only human rights, but are necessary for civilizational advancement. Private, permissionless, and uncensored. You can try it for free without an account at venice.a.I. Welcome back to the AI Daily Brief. Today we are talking about a new deal between AI agent company Adept and Amazon. Now, this is the time in the generative AI industry that we would expect to start to see a fair bit of consolidation.

We're now 19, 20 months after ChatchipT launched and got the whole party started. and right around the time that a lot of companies are running out of money from their last round. Some of them will, of course, be able to raise another round and keep building independently. However, many companies will be forced to either get themselves bought or shut down. Interestingly, however, there is increasingly, apparently a third path that is sort of like an acquisition, but not exactly. And so that's what we're going to be discussing today. Before we talk about the Adept deal, let's talk about the first big deal where people got a sense

that this might be the new approach for big tech. In March, Microsoft surprised lots of folks in the AI world when they announced that Mustafa Sullyman, the co-founder of not only Google DeepMind, but more recently, Inflection, was joining the company to effectively lead all of their AI efforts. There were a bunch of things that were notable about this announcement. The first had to do with Microsoft's relationship with OpenAI. Effectively, Mustafa Sullyman was being brought in to do the job that they had discussed Sam Altman doing when it looked like the entire OpenAI team was going to come join Microsoft

after the board kicked Sam Altman out. Some of the speculation then was that as they had started to plan for that, Microsoft actually started to really like the idea. And when Altman went back to OpenAI keeping the team intact, they wanted to pursue the same concept just with different personnel. Another line of discussion was whether it reflected Microsoft hedging against OpenAI as their only AI investment. And while that has been discussed in somewhat conspiratorial terms, it seems to me pretty obvious that in the wake of the Open AI board acting capriciously, Microsoft simply had to start hedging a little bit more than they had been in the past. But that wasn't the only thing that made the deal interesting. In addition to Mustafa Sullyman, the deal was bringing a big portion of Infliction's team and technology over into Microsoft.

In fact, it sort of seemed like it was an acquisition in all but name. We would ultimately learn that Microsoft was paying something like $650 million to Infliction for access to their technology, which was conveniently enough to make their existing investor base whole. Inflection had raised around $1.3 billion less than a year before, so that in and of itself was a big dramatic moment for the AI space. It seemed to reflect just how difficult and high cost it is to compete in this space that even a company funded on that sort of level might decide that it made more sense to go inside Microsoft than from outside.

But for our purposes today, the paradigm that's interesting is this idea of an acquisition in all but name. It seems like that's what happened with Adept as well. Ali Miller wrote, one of the bigger AI texting topics in my circles this weekend was Adept in Amazon. Adept co-founder and CEO David Luan, former VP at OpenAI and director at Google AI will join Amazon. Other Adept co-founders and employees will also move to Amazon. Amazon will license Adept Tech, and Adept will continue its product-focused operations with a smaller team. Luan, the

CEO of Adept, will oversee Amazon's AGI autonomy division and will have the Automations team report into him as well. Reports say that Adept had 100 employees and that 20 will remain at adept. Ali added a few of her own notes. She writes, it's unclear if the 80 Delta are all going to Amazon, but knowing Amazon's high level of ownership per person that seems pretty high, but lower than other acquisitions I witness are not a crazy number. Feels like either high-quality agents are farther off than we think, or the money to get there is only at OpenAI Anthropic CSP levels. The Adept team themselves put up a note called an update to Adept, announcing some updates to our strategy and company. They write, our mission at Adept since we started two and a half years ago has been to build

useful general intelligence that enables people and computers to work together. Our plan has been to train progressively larger and smarter multimodal foundation models fine-tune them into agents and then build products around them that help people do their day-to-day work better. From there, they talk about the progress they've made, but they say, continuing with Adep's initial plan of building both useful general intelligence and an enterprise agent product would have required spending significant attention on fundraising for our foundation models rather than bringing to life our agent vision. Therefore, to focus on maintaining the strong momentum and potential of our agent tech stack, we are announcing some updates to our strategy and the company. A Depp, they say,

will now focus entirely on solutions that enable agentic AI, which will continue to be powered by a combination of our existing state-of-the-art in-house models, agentic data, web interaction software, and custom infrastructure. In addition, the Adept co-founders and some of the team are joining Amazon's AGI organization to continue to pursue the mission of building useful general intelligence. Effectively Adept is saying that they were doing two things. Those two things run at cross-purposes, at least when it comes to the time available to them and what they would have had to do next, and so they're splitting it up with a big chunk of the folks who are working on general intelligence going to Amazon. The information writes the deal implies Amazon could be looking

to develop more AI in-house rather than relying on third-party providers. Amazon's licensing deal with Adept, which includes its agent technology models and datasets, is non-exclusive. Adept hadn't raised as much as inflection, which, like I said, was over a billion dollars, but it had raised more than $400 million, so it's not like this company was without resources. Almost immediately, people started to see the pattern between the Microsoft inflection deal and the Adept Amazon deal. Stephanie Palazolo and the information writes, I never thought I'd say this about a nearly 50-year-old company, but Microsoft is turning out to be a trendsetter.

First, Microsoft's $10 billion investment in partnership with OpenAI early last year prompted rival cloud providers to team up with conversational AI developers, namely Amazon Web Services similar packed with Anthropic. Then Microsoft's co-pilot AI products set the roadmap of Google. Now Microsoft seems to be teaching Amazon how to do an acquisition-like deal with a well-funded AI startup without making a formal acquisition, in theory, helping avoid regulatory scrutiny. Stephanie goes on, what Adep's statement really means is that launching AI for enterprises or consumers is hard if you don't already have a lot of customers, because companies that have a lot customers are implementing cutting-edge AI rather quickly. We don't know the terms of the Amazon

adept deal yet, but Microsoft's quasi-acquisition of inflection AI gives us a hint. In that arrangement, Microsoft agreed to pay inflection approximately $650 million, mostly through an LLM licensing deal that would help provide its investors a modest return on their capital. I wouldn't be surprised if adept in Amazon came to a similar agreement. That would be a passable outcome for the Silicon Valley investors, like Greylock Partners and General Catalysts that have poured over 400 million into the AI agent startup over the past two years. These deals are a win for the big tech firms which get AI product expertise. I'm sure there are plenty of other struggling but talent-rich AI startups like Character AI, Cohere, or AI21 Labs, and big tech companies waiting,

watching and learning from Microsoft. And don't be surprised if the Federal Trade Commission pokes around, if only to annoy the acquirers or licensees of these deals. So where is the FTC right now when it comes to AI? Throughout last year, in the beginning of this year, there had been growing bluster. For example, at the Wall Street Journal's Future of Everything Festival in May, FTC chair Lena Khan said that AI model training could be a violation of antitrust laws. She said, The FTC Act prohibits unfair methods of competition and unfair or deceptive acts or practices. So you can imagine if somebody's content or information is being scraped that they have produced

and then is being used in ways to compete with them and to dislodge them from the market and divert businesses. In some cases, that could be an unfair method of competition. Then at the beginning of June, the Wall Street Journal and others reported that the FTC had indeed opened up an antitrust probe surrounding Microsoft. And guess what? This one was specifically about the inflection deal. The WSJ writes, Companies are required to report acquisitions valued at more than $119 million

to federal antitrust enforcement agencies, which have the option to investigate a deal's impact on competition. The FTC or the Justice Department which share antitrust authority can sue or block mergers or other investments if an investigation finds the deal substantially reduced competition or lead to a monopoly. The FTC is drilling down on Microsoft's deal with inflection seeking information about how and why they negotiated their partnership.

The agency is trying to determine whether Microsoft crafted a deal, that would give it control of inflection, but also dodge FTC review of the transaction. So TLDR, to the extent that that deal was meant to ward off FTC interest, wildly unsuccessful. However, of course, the big question is whether the deal actually passes legal muster, even if the FTC isn't particularly happy about it. In general, there is definitely a sense that antitrust issues in the era of AI are getting bigger, not smaller. At the end of last month, the top German antitrust official argued in the pages of Bloomberg that AI would intensify competitive abuses by big tech.

said Andreas Munt, there's a great danger that will get an even deeper concentration of digital markets and power increases at various levels, from chips to front end. Then just yesterday, we found that French antitrust regulators are preparing to charge NVIDIA with anti-competitive practices. The report came from Reuters, and what is known is that back last year, in September, quote, French antitrust enforcers rated the offices of a business suspected in engaging in anti-competitive practices in the graphics card sector. They didn't identify the company as NVIDIA at the time, but the chipmaker has since acknowledged that France and other entities are examining its business practices. It is far from clear how all these antitrust cases are going to shake out, but it is likely that they have one way or another a fairly significant

impact on how AI develops. I will, of course, keep you posted as both antitrust actions, as well as these acquisitions that aren't really acquisitions, continue to become the norm in AI land. For now, though, that is going to do it for today's AI Daily Brief. Until next time, peace.