Tech Brew Ride Home - Thu. 06/29 – Big Interesting (AI) Raises

Episode Date: June 29, 2023

Satya Nadella had his day in court. Looks like the FTC is finalizing its big case against Amazon. YouTube has advertisers asking for refunds. Big new interesting raises in the AI space. And ahead of the Fourth of July travel weekend, we may not have the autonomous driving we want, but maybe we can get the autonomous rental car delivery we deserve. Links: Microsoft and Activision Chiefs Testify Merger Will Benefit Consumers (NYTimes) Lina Khan Is Coming for Amazon, Armed With an FTC Antitrust Suit (Bloomberg) Advertising industry seeks refunds over YouTube skippable ads (FT) ChatGPT maker OpenAI faces a lawsuit over how it used people’s data (Washington Post) Inflection AI Raises $1 Billion (Forbes) Halo Car launches remotely piloted rental car deliveries in Las Vegas (TechCrunch) Learn more about your ad choices. Visit megaphone.fm/adchoices

Welcome to the Tech meme right home for Thursday, June 29th, 2023. I'm Brian McCullough today. Satcha Nadella had his day in court. Looks like the FTC is finalizing its big case against Amazon. YouTube has advertisers asking for refunds, big, new, interesting raises in the AI space. And ahead of the 4th of July travel weekend, we may not have the autonomous driving we want, but maybe we can get the autonomous rental car delivery we deserve. Here's what you miss today in the world of tech. More from the whole Microsoft FDFRAV.

T.C. Courthouse. Sacha Nadella himself took the stand yesterday saying he would love to end exclusives on consoles. Bobby Kotik, in his testimony, said call of duty being exclusive to the Xbox would cause revolt and reputational damage. So, they're making their case in court, quoting the New York Times. If it was up to me, I would love to get rid of the entire exclusives on consoles. Mr. Nadella testified, rebutting claims from tech regulators that Microsoft's deal for the video game giant would curtail competition and restrict Activision's games only to players on Microsoft's Xbox console. I would have no love for that world, he said, end quote. The fourth day of a hearing in U.S. District Court in San Francisco that could determine the deal's outcome was the highest profile session with appearances by Mr.

Nadella and Activision's chief executive Bobby Kotick. You would have a revolt if you were to remove the game from one platform, Mr. Kotick said. It would cause reputational data. damage to the company. Mr. Nadella likewise said he would not withhold call of duty. Court arguments have focused on the practice of exclusivity, releasing a highly anticipated game only on one console. Microsoft has repeatedly promised it will not make Call of Duty exclusive to Xbox if it acquires Activision and offered Sony a contract putting that guarantee in writing. But the FTC argued in court last week that Microsoft had moved swiftly to buy Xenamax and its slate of gaming studios for $7.5 billion in 2020 when it realized that

Sony might pay to make one of Xenamax's important upcoming games, Starfield, exclusive to PlayStation. New Xenemax titles, including Starfield, are now exclusive to the Xbox. Jim Ryan, the chief executive of Sony, testified in a recorded video deposition that he thought that even if Call of Duty remained on PlayStation, Microsoft would try to, quote, drive PlayStation gamers to the Xbox platforms by somehow degrading the Call of Duty experience on PlayStation. I believe that they're going to use Call of Duty somehow to damage us, Mr. Sir Ryan said. But Mr. Nadella testified that he opposed a Waldorf approach to gaming, quote. I grew up in a company that always believed that software should run on as many platforms as

possible, he said, and I believe in that, end quote. Speaking of the FTC, seems like after nearly two years of prep, Lina Khan is getting ready to start cooking with gas. According to various documents and sources, the FTC plans to file a far-reaching case focused on Amazon's core marketplace in the coming weeks after months of honing the complaint. Quoting Bloomberg. The main allegation is expected to be that Amazon leverages its power to reward online merchants that use its logistic services and punish those who don't. FTC investigators and Khan's office have been honing the complaint for several months, two of the people said, and finalizing key details such as where to file suit. Con and her colleagues are keen to file

before personnel changes in August, according to the people who cautioned that the timing could slip. Based on her public comments, Con is unlikely to accept compromises from Amazon and could seek to restructure the company, a dramatic outcome that Amazon would surely appeal. Taking on Amazon promises to be a career-defining moment for the 34-year-old Khan, who rose to prominence articulating fresh analysis of how the Seattle-based company abuses its market power. In a prominent Law Review article, Con argued that the existing antitrust enforcement framework was ill-equipped to tackle the potential harm Amazon poses to competition. Amazon says Khan should recuse herself owing to her public statements about the company.

But when Meta made a similar demand recently, Khan continued to lead that case against Meta, ignoring a non-binding recommendation from the agency's top ethics official that she step aside. A footnote in a memo from the Ethics Council earlier reported by Bloomberg said Khan had been authorized to serve as a prosecutor on Amazon's cases, end quote. Advertisers are demanding significant refunds from YouTube, after a report suggested millions of ads served via YouTube's True View system are hidden from users. Quoting the FT, Adeletics, a digital ads analysis group has conducted research into YouTube's true view system through which the platforms more than 2 billion users can skip watching an advert

after five seconds. It found, quote, hundreds of thousands of websites and apps in which these ads play imperceptibly in the background without sound and on automatic loop. This appears to be a way to avoid viewers noticing the videos altogether, so ads were not skipped, but the strategy violates Google's terms. Google tells advertisers that a key selling point for its choice-based ad format is that they are only charged if a user watches the full clip or at least 30 seconds of it. If skipped, an advertiser pays nothing. TrueView ads are a core product for YouTube's $30 billion a year business. Equibity, a media investment, Analysis Group in London said its global clients typically allocate 40 to 50% of their YouTube budgets

on the skippable ads. They are supposed to play in-stream, meaning viewers see them, quote, before, during, or after other videos on YouTube, or through its Google Video Partners Network of, quote, high-quality publisher websites and mobile apps where you can show your video ads to viewers beyond YouTube, end quote. But analytics, which used web crawler data to scour the internet and also worked with dozens of media buyers from brands and agencies, found thousands of true-view ads were placed, quote, outstream. Tucked away on portions of a website where viewers had little to know interaction with them, the Wall Street Journal previously reported on some of these same findings.

Joshua Locock, a global chief media officer at UM, a New York-based ad agency, said he expected YouTube to investigate the issue and refund the affected advertisers. Adeletics created a list of those hit, which included dozens of leading brands, such as J.P. Morgan Chase and Johnson and Johnson, as well as the U.S. Department of Health and Human Services, and quote. It has begun, and by it, I mean this. A California-based law firm has launched a class action suit against OpenAI, claiming that company violated millions of internet users' rights by scraping their data. Quoting the Washington Post. The lawsuit seeks to test 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. Clarkson, the law firm behind the suit, has previous. brought large-scale class action lawsuits on issues ranging from data breaches to false advertising. The firm wants to represent, quote, real people whose information was stolen and commercially

misappropriated to create this very powerful technology, said Ryan Clarkson, the firm's managing partner. The case was filed in federal court in the Northern District of California Wednesday morning. A spokesman for OpenAI did not respond to a request for comment. The lawsuit goes to the heart of a major unresolved question hanging over the surge in generative AI tools, such as chatbots, and image generators. The technology works by ingesting billions of words from the open internet and learning to build inferences between them. After consuming enough data, the resulting large language models can predict what to say in response to a prompt, giving them the ability to write poetry, have complex conversations and past professional exams. But the humans who wrote

those billions of words never signed off on having a company such as Open AI use them for its own profit. All of that information is being taken at scale when it was never intended to be utilized by a large language model, Clarkson said. He said he hopes to get a court to institute some guardrails on how AI algorithms are trained and how people are compensated when their data is used. The firm already has a group of plaintiffs and is actively looking for more. The legality of using data pulled from public internets to train tools that could prove highly lucrative to their developers is still unclear. Some AI developers have argued that the use of data from the internet should be considered fair use, a concept in copyright law that creates an exception if the material is

changed in a transformative way. The question of fair use is, quote, an open issue that we will be seeing play out in the courts in the months and years to come, said Catherine Gardner, an intellectual property lawyer at Gunderson Detmer, a firm that mostly represents tech startups, end quote. Metric ton of interesting raises in the AI space this morning. I'm not even going to mention half of them, but I'll give you one. New Year's. York City-based runway AI, which lets customers generate videos from text prompts, raised $141 million from Google, Nvidia, and others taking its total funding to $237 million. Runway AI is another one of those companies that's been around for a while, since 2018, in fact,

this values the company at $1.5 billion. The much bigger headline comes from Inflection AI, a startup led by ex-deep-mind leader Mustafa Suleiman that released its Pi chatbot just back in May of 2023. It has raised $1.3 billion, a source says, at a $4 billion valuation. Quoting Forbes, less than two months after the launch of their first chatbot pie, artificial intelligence startup inflection AI and CEO Mustafa Suleiman, have raised $1.3 billion in new funding. Microsoft, Nvidia, and three of tech's most influential billionaires led the investment in the Palo Alto-based startup launch in early 2022. LinkedIn co-founder Reid Hoffman, Microsoft co-founder Bill Gates, and former Google

CEO Eric Schmidt, all personally invested with Nvidia, the sole new investor among the group. The new funding values inflection at $4 billion, according to a source with knowledge of the transaction. Infliction said the company and Suleiman remained majority shareholders and declined to further comment. In an interview, Suleiman said that the group of mostly insiders proposed the additional investment after inflection was, quote, overwhelmed with offers following the launch of Pye, its conversational chat bot launched in May. I think people can see that it's just the tip of the iceberg, Suleiman told Forbes. There's so much further to go after Pai validates the core thesis, which is that conversation is the new interface, end quote. Some details of inflection's new deal with Microsoft and

NVIDIA are, like Suleiman's iceberg, still largely out of view. He declined to provide a breakdown of how much of the $1.3 billion raised included cash equivalents, such as computing credits, but said that a very, very large chunk was in dollars. We have all the cash. we need to run and operate, he added. Inflection also declined to comment on how much equity Microsoft and Nvidia now held in the business, but Soleiman said neither company commands ownership like control over it or other preferential rights. In practice, it was a very traditional round, he said. There's no IP movement, and we still are entirely independent and at liberty to do whatever we want

on the commercial front and partner with whomever we want. So there are no real restrictions, he said. What is clear, the round significantly deepens inflection's ties with Microsoft and NVIDIA, two key partners in the AI race. Invidia worked closely with Inflection and service provider Corweave to co-develop inflection's current H-100 cluster. Inflection paused its own work for NVIDIA to run a recent test that NVIDIA announced this week, had set records on eight tests of current AI model training benchmarks,

completing a benchmark based on GPT3 in less than 11 minutes. That test, which matched the computational power of training a model that took an estimated three to six months to develop, ran on inflection's 3,584 H-100 GPUs already in service, Suleiman noted. But in the wake of this funding round and partnership inflection's growing horsepower is about to get a turbocharge. NVIDIA and CorWeave, which helps physically deploy the GPUs, are now in the process of helping Inflection install many thousands more.

Once fully operational, Inflection's new cluster will run 22,000 H-100s. Inflection believes that to be the largest GPU cluster for AI applications in the world, ahead of META's 16,000 GPU cluster announced in May. Just how many open AI is using is currently unknown. InVIDIA announced last November, it planned to incorporate tens of thousands of GPUs into Microsoft's Azure Cloud Service. Against the world's largest clusters overall, inflection said it estimated that it would trail only frontier, the supercomputer maintained by the Oak Ridge National Laboratory in Tennessee, end quote. After all that, this is eyebrow raising, according to sources after

struggling to raise at a $4 billion valuation, Stability AI secured a convertible note worth less than $25 million. It being a convertible note is what's eyebrow raising, also that dollar figure being small. Stability AI denies fundraising difficulties, but earlier this year, the talk was that stability was looking to raise at a $4 billion valuation that never materialized. Their last valuation last year was $1 billion. Finally, today with the big hot.

holiday week here in the U.S. on our doorstep, lots of people will be renting cars over the next couple weeks. Renting cars is a bit of a nightmare, you probably know. You have to go to out-of-the-way places like airports just to get the things and also to return them here in Brooklyn to call the Hertz rental place on Union Street. A nightmare hellscape is frankly being generous. But also, think about it from the rental car company's perspective. They have to have all of this overhead, essentially giant parking lots or garages just to make the cars available to people. Think of how the costs could be lowered if cars could be delivered to people right to their door literally on demand. Well, quoting TechCrunch. Halo Car, a startup that uses remote operators

to deliver rental cars to customers' doors, has launched driverless operations in Las Vegas. Driverless operations means something different for Halo than they do for autonomous vehicle companies like Cruise or Waymo because Halo's vehicles aren't capable of self-driving. The startup's fleet is kidded out with a suite of six cameras, modems, antennas, and other components to send data back to remote pilots at a Halo operation center. Those pilots then use the video and sensor data that streamed in to remotely drive the vehicles. Once a remote driver completes a car delivery, they hand over control of the vehicle to the customer and move on to the next vehicle awaiting remote delivery or collection. Halo has been delivering

vehicles to customers in Las Vegas using teleoperations for around a year, but a human driver has always been present in the front seat for safety reasons. Now Halo cars will be delivered to customers with no drivers in the vehicle. This is an important step towards achieving Halo's vision of on-demand vehicles being economically viable, according to a Nandakamar founder and CEO of Halo. That said, Halo isn't at the stage where it's hitting positive unit economics just yet. The company will still use a remote chase car that tails behind the remotely piloted vehicles initially. The driver of the tail car can stop the remotely piloted vehicle and takeover if needed. The tail car also acts as a buffer vehicle in case the Halo car needs to stop, thus preventing a potential rear-end accident

with other road users. Halo's cars will come to a stop in lane if the system detects an anomaly, which means they meet Nevada's minimum risk condition for AVs that says vehicles must be able to stop if there is a malfunction in the system. Halo says it will ditch the tail car, over the next year based on how the current operations perform. That'll happen in phases across operation zones and depending on the times of day, according to Nanda Kumar. While Halo might be the first company to successfully deliver remote-pilated EVs to customers in Las Vegas, it's not the only one attempting such a feat.

In December 2022, R.C. Moto, the maker of the three-wheeled electric fun utility vehicles, teamed up with faction to develop EVs that can be delivered to a customer's hotel through a combination of low-level autonomy and teleassist technology, end quote. Nothing for you today. Talk to you tomorrow.