The AI Daily Brief: Artificial Intelligence News and Analysis - Inflection Raises Monster $1.3 Billion Round to Build One of World's Biggest AI Clusters

Episode Date: June 30, 2023

Inflection has raised $1.3B for its "personal intelligence" and to build one of the world's largest GPU clusters. OpenAI sets up shop in London, updates their mobile iOS ChatGPT app with Browse with Bing, and gets sued for data scraping. Runway raises $141M. MLB experimenting with AI scouting. 150+ corporate executives write an open letter slamming EU AI Act. Today's Sponsor: Supermanage - AI for 1-on-1's - https://supermanage.ai/breakdown   NOTE: While NLW is traveling this week, The AI Breakdown will only be releasing The Brief each morning. We'll be back to our regular content at the end of the week.    The AI Breakdown helps you understand the most important news and discussions in AI.    Subscribe to The AI Breakdown newsletter: https://theaibreakdown.beehiiv.com/subscribe   Subscribe to The AI Breakdown on YouTube: https://www.youtube.com/@TheAIBreakdown   Join the community: bit.ly/aibreakdown   Learn more: http://breakdown.network/

Today on the AI Breakdown Brief, we're discussing inflections monster $1.3 million fundraising round, as well as a set of news from OpenAI. The AI breakdown is a daily podcast and video about the most important news and discussions in AI. Check out Breakdown.network for more information. Hello, friends, welcome back to the AI breakdown. Some good news. I am finally back in the U.S., so today will be the last day of the Brief-only shows. Tomorrow, I'm planning on a little bit of a research recap, then some. Sunday, we'll have a very cool long read Sunday, and then next week we will be back to our regularly scheduled brief plus main topic format. Now, one more quick thing, though, before we get to today's

episode, I'm really excited to share a new sponsor with you guys, Supermanage. I'm of the strongly held belief that many of the areas where AI is going to be most impactful the soonest is in improving the things that are already important parts of how we work, for example, the one-on-one. one-on-one meetings should be a chance to celebrate wins, solve problems, and deepen the relationships between coworkers. Supermanage helps those meetings live up to that potential by tapping the power of AI to distill all of the communications in public Slack channels into a one-on-one briefing that saves time and allows managers and their team members to get right into the most important discussions.

I'd love it if you would go check them out, supermanage.a.i. forward slash breakdown. That's supermanage.com slash breakdown. And with that, let's go to the AI breakdown brief. Welcome back to the AI breakdown brief, all the AI headline news you need in five-ish minutes or less. Today we kick off with some major funding news. Inflection, the company behind the personal AI platform Pi has raised a $1.3 billion financing round,

led by notables including Microsoft and Nvidia, and also included Reid Hoffman, who's one of the co-founders of not only Inflection, but of LinkedIn. Now, on the one hand, it would be tempting to see this as just an example of AI hype going crazy, a company that only has about 35 employees raising a $1.3 billion round seems nuts, but you have to remember that when it comes to foundation models, data and compute are the name of the game. A huge part of this funding is going to building what they're calling one of the largest AI clusters in the world that is comprised of 22,000 Nvidia H1 TensorFlowGPUs.

Now, Inflection's whole thesis is around personal AI. As Mustafa Silliman, the CEO and co-founder of Inflection puts it, Personal AI is going to be the most transformational tool of our lifetimes. And in the announcement post, they really reinforce those themes. They say that Pye, your personal AI, was created to give people a new way to express themselves, share their curiosities, explore new ideas, and experience a trusted personal AI. Now, that wasn't the only major funding round announced in the last couple of days. Runway, which is the Text 2 Video Company, has announced a $141 million fundraise at a $1.5 billion valuation.

This investment, like so many in AI, includes participation from the big companies, including Google and InVideo once again. Now, if you're watching the AI breakdown regularly, you'll know that Runway's Gen 2 started rolling out two participants far and wide over the last month and is already getting rave reviews. Many people have compared it to Mid Journey's moment last year and are wondering if that in the same way that that democratized access to text to image creation, runway can bring text to video to the masses. Next up on the AI breakdown brief, we turn to a slew of news around OpenAI. First, from a product perspective, chat GPT has gotten a major upgrade on its iOS mobile app. The iOS mobile chat GPT application now supports the browse with Bing feature. That means that people who are using the iOS app can now access the web directly from their

chat GPT app. Currently, this is only available for iPhone users, but they say that it will be coming to Android soon. OpenAI has also announced their first international location with the company. creating a new headquarters in London. Now, of course, the UK has been in the news quite a bit recently for wanting to take a leadership role not only on the development of AI technology, but on the development of AI policy and regulation. Sam Altman has recently met with Rishi Sunak, the British Prime Minister, and Rishi has made his intentions around creating London and the UK more broadly as an AI hub very clear. You have to think this is a boon for that strategy. OpenAI CEO Sam Altman said in

their press release, we see this expansion as an opportunity to attract world-class talent and drive innovation in AGI development and policy. Now that said, not all of OpenAI's news is particularly grand. This week, the Washington Post reported about a lawsuit against the company over how it used people's data in its training. A California-based law firm is launching a class action lawsuit against OpenAI, saying that the way that the company built ChatGPT violated the copyrights and privacy of countless people because OpenAI scraped the data from the public web. As the post puts it, the lawsuit seeks to test out a novel legal theory, that OpenAI violated the rights of millions of internet users when it used their social media comments, blog posts, Wikipedia articles, and

family recipes. Said one of the partners, all of that information is being taken at scale when it was never intended to be utilized by a large language model. Now, whatever one thinks of this particular lawsuit, the issues underlying it seem destined to have their day in court, and sooner rather than later. Says Catherine Gardner, an IP lawyer at Gunderson Detmar, the question of fair use is an open issue that we will be seeing play out in the courts in the months and years to come. Investor Chimath Palahapatia wrote, I only read parts of this and lightly at that, but there are some arguments here that probably need an answer from the courts or Congress. Without commenting on the merits of this specific lawsuit, as large-scale training

becomes an arms race, third-party companies should probably be mandated to explain how they are training their models and should get consent from users to collect their data and train on it, especially if it comes via scraping. Now, of course, these sort of questions are exactly what new policies and regulations, such as the EU's AI Act, are trying to sort it out from a legal perspective. While some were enthusiastic about the EU AI Act's draft text, this week over 150 executives from some of Europe's leading companies signed an open letter saying that this act, as written, could seriously hamper European competitiveness when it comes to AI. The letter which was sent on Friday to the European Commission, the European Parliament,

and to member states said, in our assessment, the draft legislation would jeopardize Europe's competitiveness and technological sovereignty without effectively tackling the challenges we are and will be facing. From the Financial Times who saw the letter, quote, the signatories said their concerns were particularly acute over genitive AI. The proposed rules would quote heavily regulate foundation models, which is the technology underpinning chatbots regardless of their use cases. Now, what's interesting about this, if you listen to my coverage of the EUAI Act, is that when the European Parliament started debating this, it was pre-chatGBT. Many of the core provisions that

have been argued over the last couple years were focused more. on things like how the government was going to use AI and specific use cases that might contain particular risk, such as criminal profiling. Many of the provisions around generative AI and what to do with foundation models were sort of bolted on in the last six months. So all in all, despite the fact that a draft of the EU AI Act has been passed, it seems like the political battles over it have just begun. Even as regulators around the world race to catch up with AI, however, companies are plowing ahead with new implementations. Korea's SK Telecom, one of the country's biggest mobile carriers announced Friday that it had added chat GPT tech to its AI chatbot,

saying that you could now talk to it like a quote-unquote close friend. Now, SKT telecom has been playing around with similar features for over a year now. However, the new tools add a chat room feature and the ability for people to interact with an AI character. SKT said in a press release through these conversations, users can share information about their daily life like they are talking to a close friend. To the extent that you're looking for patterns and trends, this certainly seems to add evidence in the favor of Benflection's personal intelligence idea. Finally, an interesting, if controversial one from the world of sports, Major League Baseball has begun experimenting with AI-powered scouting tools.

The Wall Street Journal writes, The most important thing a baseball team can do ahead of the amateur draft is predict future success for developing young ball players. Scouting has changed drastically since an explosion of data and technology entered the game a decade ago, and now Major League Baseball is introducing an even more high-tech tool, analysis of player potential via artificial intelligence. So basically what's going on is MLB has partnered with Uplift Labs, which is a biomechanics company that documents using iPhone cameras.

Using AI, Uplift says it can translate the images that are captured by those cameras to quantify different elements of a player's potential. That includes on the downside things like risk for injury, but on the upside, skill sets that may not be immediately apparent, particularly with younger players. Uplift founder, Sukimasa Kabayamo said, We have metrics on things like kinematic sequence, stride-length, ball contact timing. At the same time, we also have this new kind of very early injury warning detection. Let's say if you have too much of an arm flare, you know that there may be potential overload on the elbow, which can unfortunately lead to Tommy John's surgery. Now, optimistically, the Wall Street Journal writes, some player agents believe that movement data could help players who are likely to go in later rounds of the draft, showing deeper skill sets that may not be immediately apparent in the player's high school or college results. Then again, as they write, the same information that could be used to turn you into a better player could also be used to devalue you.

said Tom Collar, a former Major League pitcher and now player agent, I think whenever something is introduced, it's natural to wonder why and to proceed with caution. The real question becomes, why is it being used? Is it being used as a barometer to see how these guys develop throughout their career and how to prevent injuries moving forward? Or is it being used to hurt draft position based on how somebody's body works? And like so much else with AI, whether this is ultimately good or bad, the one thing that's clear is it's coming.

That's it for today's AI breakdown brief. If you're enjoying it, please subscribe to the podcast, the newsletter version, and until next time, peace.