Tech Brew Ride Home - Wed. 12/27 – NYT Sues OpenAI

Starting point is 00:00:00 On April 4th, 2023, around 2 in the morning, a man was found stabbed multiple times on a sidewalk in downtown San Francisco. Hey, who did this to you? What happened next turned the story into a political firestorm. Reports have identified the victim as Bob Lee, the founder of Cash App. From Bloomberg Podcasts, this is Foundering, the Killing of Bob Lee, beginning April 16. Welcome to the Tech meme right home for Wednesday, December 27th, 2023. I'm Brian McCullough today. The New York Times has broken the seal and sued OpenA. and Microsoft. Ads are coming to prime video in mere days. More rumors about Johnny I founding an AI hardware startup with Sam Altman. Anthropics seems to be proving open AI isn't the only one that can make money in the AI space. And when chips go beyond the one kilowatt barrier, you need to liquid cool them. Here's what you miss today in the world of tech. Well, we knew something like this was coming. We just didn't know who would fire the first shot. Turns out it's the New York Times, which has sued Open AI and Microsoft for copyright infringement,

Starting point is 00:01:14 alleging millions of articles were used in training their AI products. This is the first major U.S. Media Organization to sue on these grounds, quoting, well, the New York Times. The lawsuit filed in Federal District Court in Manhattan contends that millions of articles published by the Times were used to train automated chatbots that now compete with the news outlet as a source of reliable information. The suit does not include an exact monetary demand, but it says the defendants should be held responsible for, quote, billions of dollars in statutory and actual damages related to the, quote, unlawful copying and use of the Times' uniquely valuable works, end quote. It also calls for the companies to destroy any chatbot models and training data that use copyrighted material from the Times. Defendants seek to free ride on the Times'

Starting point is 00:01:59 is massive investment in its journalism, the complaint says, accusing OpenAI and Microsoft of using the Times' content without payment to create products that substitute for the Times and steal audiences away from it, end quote. So apparently this lawsuit came after extensive negotiations between the Times and the two companies hit a wall. The newspaper apparently proposed a potential commercial partnership along with the implementation of technological guardrails on AI-driven products like chat GPT. The lawsuit goes beyond intellectual property, though. It positions AI technologies such as chat GPT as emerging rivals in the realm of journalism. This is particularly relevant when AI systems upon querying them produce content on current events and topics of public interest, often drawing

Starting point is 00:02:45 upon the Times' past reporting. The Times is worried that consumers might opt for AI-generated summaries instead of accessing its own website, potentially diminishing its web traffic. In the legal complaint, the Times highlighted instances where AI chatbots like ChatGBTGPT replicated content from its articles almost word for word. Such content is usually behind a paywall and accessible only to its subscribers. The lawsuit underscores the emphasis by OpenAI and Microsoft on incorporating content from the New York Times in their AI training processes. Remember that recent story we did about Apple wanting to pay to get content from, say, the New Yorker? Increasingly, the idea is that training content from, say, the New York Times and the New Yorker,

Starting point is 00:03:26 of the world is of higher quality than, say, YouTube comments. Plus, it helps to have content that can go back decades or even centuries, thus the deal's OpenAI struck with the AP and Axel Springer. One wonders if this is a negotiating tactic on the Times part or an actual lawsuit that will see through to seeking redress. Though the law firm, the Times retained for this suit, is the same one that represented Dominion voting systems in that defamation case against Fox News that resulted in a $787 million settlement, so that suggests they're serious. Amazon Prime Video plans to start showing ads, starting in two days on January 29th, though they do give you an option for customers to pay an additional $2.99 per month to avoid the ads,

Starting point is 00:04:17 quoting the verge. This will allow us to continue investing in compelling content and keep increasing that investment over a long period of time. The company said in an email to customers about the pending shift to what they call limited advertisements. We aim to have meaningfully fewer ads than linear TV and other streaming TV providers. No action is required from you, and there is no change to the current price of your prime membership, the company wrote. Customers have the option of paying an additional $2.99 per month to keep avoiding advertisements. The rest of the email summarizes the many benefits of a prime

Starting point is 00:04:47 subscription, no doubt an attempt to keep customers from canceling over this decision, end quote. Yeah, they want you to feel like nothing is changing. But Amazon Prime is 1499. a month, or you can get it for $139 if you pay for the whole year. But if you just want Prime Video, it's $899 a month. With this extra $2.99 per month to avoid ads, Prime would cost you just shy of $18 monthly. And if you're only into Prime Video, it's going to be almost $12. Amazon does have another streamer called FreeV. It's totally free, but it's got ads. So you're almost out of options to avoid ads on Amazon at this point unless you do pay up. Example number two in as many days that Sam Altman's ambitions have not taken a hit,

Starting point is 00:05:36 even after his recent issues with the Open AI board. Sources are telling Mark Gurman that outgoing Apple executive Tang Tan plans to join Johnny Ives' love from to work on some mystery hardware product that would incorporate AI heavily. Sam Altman is apparently planning to provide the software for this product. Quoting Bloomberg, the work marks one of the most ambitious efforts undertaken by Ives since he left Apple in 2019 to create Love From. The iconic designer is famous for the products he helped devise under Apple co-founder Steve Jobs, including the iMac, iPhone, and iPad. His hope is to turn the AI device work into a new company, but development of the products remains at an early stage,

Starting point is 00:06:16 according to the people. The efforts so far are focused on hiring talent and creating concepts. Tan will lead hardware engineering at the project while working at Love From. The people said, Bloomberg News previously reported that the executive was stepping down as Apple's vice president of iPhone and watch product design. He isn't slated to depart until February, though his responsibilities were already divided up this month. Altman, who was fired as OpenAI, Chief Executive Officer in November and quickly reinstated, irked some board members by raising funds for other endeavors. That included the effort with I've, Bloomberg News reported in November. Altman headquartered Softbank Chairman Masayoshi Son about making a multi-billion dollar investment in the new company.

Starting point is 00:06:56 Projects in the concept stage include devices for the home. OpenAI referred questions to love From, which declined to comment. Apple also declined to comment. For Apple, TAN's departure underscores an exodus of design talent since 2019, about 14 members of Ives' former team at Apple have left. Only roughly half a dozen of the designers who once reported to Ives still remain at Apple, end quote. Meanwhile, Love From's roster of clients include Airbnb and Ferrari, that three-year consulting gig it had with Apple ended last year, and roughly 20 former Apple employees are employed by Ive at love from. According to Pitchbook, Microsoft, Google, and Amazon have made investments amounting to two-thirds of the $27 billion raised by AI startups this year in 2023, vastly outspending VC firms,

Starting point is 00:07:50 quoting the Financial Times. The huge outlay, which exploded after the launch of OpenAIs chat GPT in November, highlights how the biggest Silicon Valley groups are crowding out traditional tech investors for the biggest deals in the industry. The rise of generative AI, systems capable of producing human-like video, text, image, and audio in seconds, have also attracted top Silicon Valley investors, but VCs have been outmatched, having been forced to slow down their spending as they adjust to higher interest rates and following valuations for their portfolio companies. Over the past year, we've seen the market quickly consolidate around a handful of foundation models with large tech players coming in and pouring billions of dollars into companies like OpenAI,

Starting point is 00:08:27 cohere, Anthropic, and Mistral, said Nina Akajian, a partner at U.S. Venture, firm Index Ventures, referring to some of the top AI startups. For traditional VCs, you had to be in early and you had to have conviction, which meant being in the no on the latest AI research and knowing which teams were spinning out of Google, DeepMind, meta, and others, she added, end quote. Yes, but a lot of this is strategic. After Microsoft swallowed up OpenAI, at least de facto, every other major tech platform needed its own big AI partner, its own LLM factory, if you will.

Starting point is 00:09:00 At the same time, every one of those players needed a. a big-pocketed partner with tons of compute capacity, so it's not crowding out so much as, you know, market fit in a sense. And speaking of one of those LLM factories, one of the questions investors have been asking of this AI moment is, can real money be made in AI? Well, quoting from the information, at least some of the players, some types of players seem to be quite capable. Quote, Anthropic and Open AI rival backed by Amazon and Google has projected it will generate more than $850 million in annualized revenue by the end of 2024, according to two people with knowledge of its financial picture. Just three months ago, Anthropic told some investors

Starting point is 00:09:47 it was generating revenue at $100 million annualized rate and expected that figure would reach $500 million by the end of 2024. It isn't clear why the latest projections are materially higher. The projection underscores the three-year-old startup's remarkable growth expectations and may offer more evidence that generative artificial intelligence is gaining steam among enterprises. annualized revenue typically reflects the prior month's revenue multiplied by 12, so the projection implies the startup will generate revenue of at least $70 million per month by the end of next year, an increase of more than eight times from its monthly revenue around September of this year. Anthropic generates revenue from the sale of Claude, AI, that can generate and analyze text. An Anthropic spokesperson did not have a comment for this article. The new projection comes as Anthropic is in the process of raising around $750 million from Menlo Ventures and other investors.

Starting point is 00:10:37 at a $15 billion pre-investment valuation at least three times its prior valuation from an earlier funding round in 2023. The new capital will be added to the billions of dollars in funding Amazon and Google have already committed to the startup. Anthropic will train its AI using Amazon's and Google's cloud servers. Anthropic is also selling Claude to the two cloud providers' customers. Some people close to the company believe Anthropic could reach $1 billion in annualized revenue next year or $83 million in revenue per month. A Mark Open AI passed during the summer. The creator of ChatGPT has likely passed $1.5 billion in annualized revenue by now, given its recent revenue growth trajectory and thanks in part to Microsoft selling the startup's technology to cloud customers, end quote.

Starting point is 00:11:25 Microsoft is testing a feature that would let Windows 11 users reinstall the operating system using Windows Update, a method that would preserve files, settings, and installed apps from Windows latest. A new Windows update feature could be a game changer for those scared of losing files or pictures when attempting to reinstall or recover their Windows 11 installations. The new feature, Fixed Problems Using Windows Update, lets you reinstall Windows 11 using Windows Update. Currently, if you want to repair or reinstall the installed version of Windows, you need to perform an in-place upgrade using the Media Creation Tool. The process lets you keep all your files, settings, and installed apps, but reinstall the installed version of Windows or sometimes

Starting point is 00:12:04 newer versions without losing files. Windows-in-place upgrades are generally considered a somewhat lengthy and time-consuming process, and some users fear they might lose their files or pictures. Thankfully, Microsoft is aware of the concerns and is adding a new feature called Fix Problems Using Windows Update. Microsoft is testing the feature with users in the beta channel, and you will find the fixed problems using Windows Update toggle under Settings, System Recovery, but it doesn't seem to work at the moment. With this new recovery tool, you can, quote, reinstall your current version of Windows, your apps, files, and settings will be preserved. that's according to the description of the toggle in the Settings app.

Starting point is 00:12:40 The idea is to repair the existing Windows installation by downloading a fresh copy of the OS from Windows Update. And the best part, it won't remove any file, settings, or apps, according to a support document from July of 2023. So how is this different from the existing reset your PC or cloud reset features? The key difference is that Windows will download the files from Windows Update without OEM bloatware, software, or other junk. Another notable difference is the new recovery feature won't delete your data or uninstall apps, unlike the reset options, end quote.

Starting point is 00:13:14 GM has had to pause Chevy Blazer EV sales as it works to fix software problems, including the infotainment screen going blank, which apparently is affecting a limited number of vehicles, quoting TechCrunch. The automaker told TechCrunch that owners are experiencing problems with the SUV screens and with charging at DC fast stations, confirming earlier reports from journalists who had access to review vehicles. GM said its engineering teams are, quote, working around the clock toward a solution and that when it has one ready, Blazor EV owners will have to bring their vehicles to a dealership for a software update.

Starting point is 00:13:47 GM also claimed a, quote, limited number of vehicles are affected but didn't offer a figure, and that the problems are, quote, not safety-related nor related to Ultimum or Google built-in. The Blazor EV only just had its official launch a few weeks ago, but pretty quickly two media outlets had big problems with their longer-term test cars. Kevin Williams at InsideEV saw his Blazers infotainment screen go completely blank and unusable. Then he had problems charging the car, including an alert to have the vehicle immediately serviced and was ultimately stranded mid-road trip, end quote. We've spoken about this dilemma before.

Starting point is 00:14:21 Car buyers are increasingly buying cars based on how easy the user experience is, especially digitally. Carmakers want to own that experience, thus the move away from Apple's CarPlay and Google's various implementations. And yet, UI and UX, it's not something the car companies have a long history of experience with. So, delicate balancing act. Finally today, just some interesting in the weeds technical stuff. As Moore high-performance computing chips have pushed past the one kilowatt mark, at least this year, I'm thinking of chips like Nvidia's GH-200 superchips, many manufacturers have begun embracing direct liquid cooling.

Starting point is 00:15:04 Quoting the register, as Moore's law slowed to a crawl, chips, particularly those used in AI and high-performance computing, have steadily gotten hotter. In 2023, we saw accelerators enter the... kilowatt range with the arrival of Nvidia's GH-200 superchips. We've known these chips would be hot for a while. Now, Nvidia has been teasing this CPU-GPU-Franc chip for the better part of two years. What we didn't know until recently is how OEMs and systems builders would respond to such a power-dense part. Would most of the systems be liquid-cooled or would most stick to air cooling? How many of these accelerators would they try to cram into a single box and how big would the box need to be? Now that the first systems based on the GH-200 make their way to market, it's become clear that

Starting point is 00:15:45 the form factor is very much being dictated by power density than anything else. It boils down to how much surface area you have to dissipate the heat. There are a couple of advantages to liquid cooling beyond a mere efficient transfer of heat from these densely packed accelerators. The higher the system power is, the more static pressure and airflow you need to remove the heat from the system. This means using hotter, faster fans that use more power, potentially as much as 20% of system power in some cases. Beyond about 500 watts per rack unit, most of the OEMs and ODMs seem to be opting for liquid-cooled chassis as fewer slower fans are required to cool lower power components like NICs, storage, and other peripherals, end quote. Nothing for you today. Talk to you tomorrow.

Tech Brew Ride Home - Wed. 12/27 – NYT Sues OpenAI

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.