Should Americans Get Shares in AI Companies?

**Nathaniel Whittemore** (0:01)
Today on the AI Daily Brief, as OpenAI and Anthropic race towards IPO, should AI be a public good?
Before that in the headlines, a new Nvidia chip means the Mac M series has some competition. The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI.
All right, friends, quick announcements before we dive in. First of all, thank you to today's sponsors, KPMG, Scrunch, Section, and OutSystems. To get an ad free version of the show, go to patreon.com/aidailybrief, or you can subscribe at Apple Podcasts. To learn more about sponsoring the show, send us a note at sponsors at aidailybrief.ai.
Honestly, if you want to find out anything else about the show, you can probably find it at aidailybrief.ai as well.
Today, we kick off some news from Nvidia. Honestly, it's another reminder of just how much the company is doing, even though the context in which we talk about them is mostly just their core product of chips. On Monday, the company held their GTC Taipei event, with the headline reveal being a new chip called the RTX Spark. While Nvidia billed it as a super chip, the RTX Spark is the first of their standalone prosumer-grade CPUs. The chip will feature 20 CPU cores married to over 6,000 integrated GPU cores supporting up to 128 gigabytes of unified memory. The chip is capable of delivering one petaflop of AI compute, which by comparison, an H100 outputs around four petaflops. Nvidia said that the chip will be available in Windows PCs and laptops by the fall, with models available from Asus, Dell, HP, Lenovo and Microsoft at launch. Pricing was not announced, but these will be premium devices designed to compete with high-end Mac products and gaming computers that pull double duty as inference workhorses. I saw a lot of people basically saying that this is their competition to the M5 series of Mac computers. In some ways, this is part of a trend and a shift in the compute workload as we prioritize inference. GPUs, of course, have been the core piece of AI hardware for almost a decade, but the CPU is now having a resurgence. GPUs are increasingly seen as hardware for AI training, while powerful CPUs are better for executing agentic tool calls. Kara Brischke, Nvidia's VP for GenAI Software, said of GPU-powered chatbots, that era is ending. Agents are the new workload. They will run everywhere from the data center to the edge.
In addition to those new machines, Nvidia CEO Jensen Huang also announced that Vera Rubin had entered full production. OpenAI and Anthropic have already taken delivery of their first units with plans to scale up into full data center buildouts this year. The Vera-Rubin nomenclature refers to the CPU-GPU pairing in the chip. Vera is the CPU while Rubin is the GPU architecture. For the first time for an Nvidia data center chip, the focus is on the CPU and its ability to supercharge agentic AI.
Said Huang, AI agents will be the largest users of computing. Vera is the first CPU designed for that future, built to run agentic AI at hyperscale with extraordinary performance, efficiency and programmability. Making that comparison that I just mentioned, The Verge argued that the RTX Spark could be the M1 moment for Windows. I.e. until now, Apple's M-series chips have been the go-to for running AI models locally, and now in its fifth generation, the M-series architecture has been utterly dominant. Apple has been the only choice for local inference, and many of us have the Mac minis to prove it. Nvidia is looking to replicate that breakout moment and capture a new slice of the market, going head-to-head with Apple for what Huang is calling the personal AI computer. Now, this idea of personal AI computing feels to me like it's very likely to be on display throughout the week as Microsoft Build kicks off.
Microsoft CEO Satya Nadella used the RTX Spark announcement as a kickoff, tweeting that their goal at Microsoft was to deliver unmetered intelligence to every home and every desk with Windows.
Now, staying on hardware for a minute, Meta is apparently joining the AI hardware competition with a new pendant. The information reports that Meta plans to begin testing the pendant over the next year as part of a broader AI hardware strategy, a leaked memo described to push into business-focused devices, which they are calling wearables for work. The core strategy, however, is to use wearables as the hook to increase use of Meta's AI models and drive consumer agent subscriptions. Now, it is worth noting that Meta is coming at this competition from a slightly different place, considering that the Meta Ray-Bans are the most popular AI device on the market. At this point, obviously, AI pendants are not a going concern for most people, but it appears that Meta wants to ensure that they have a product available just in case, preempting new devices from Google and OpenAI, and forming something of a natural conclusion of Meta acquiring AI pendants start up limitless at the end of last year. This also could just mark a shift in revenue strategy for Meta's beleaguered Reality Labs division. The division behind their Metaverse efforts continues to bleed money, even with the success of the Meta Ray-Bans. Last quarter, the division produced 4 billion in operating losses on revenue of 402 million, but Meta hopes the combination of useful consumer agents and new devices will drive new AI subscription revenue. In the memo, VP of Wearables, Alex Himmel wrote, To build a sustainable business beyond hardware margins, we need to monetize the software experiences that differentiate our devices. It has actually been some time since we talked about AI wearables, and one of the things that I'm watching most closely for on that front, is whether all those efforts fall for OpenAI, specifically in the category of sidequest that they are now trying to avoid, or whether that's still an area that they really plan on competing. Now, one more story on the Meta front, the company also just suffered a massive exploit, which many are blaming on AI. On Monday, numerous Instagram accounts were hijacked, including the Obama White House and Sephora. Users who had their account stolen said that they were unable to reach a human tech support worker. Back in March, Meta announced that they would be rolling out AI support to all accounts across Facebook and Instagram to carry out routine tasks like password requests. However, the system appears to have some pretty significant flaws, and over recent days, videos have been circulating on Telegram disploiting the exploit. An attacker can simply ask Meta's AI to link any arbitrary account to a new email address when requesting a password reset. For verified accounts, the AI bot will ask for a video of the person to prove liveness, but hackers found that an AI-generated video of the account owner would pass the checks. Hackers also needed to use a VPN to spoof the correct location, which became trivial when Meta added location data to Instagram profiles. Two-factor authentication was completely bypassed and many didn't notice anything was amiss until their account was gone. Commentators are fairly baffled by the misstep. In a series of tweets, Griglio-Rose writes, It's wild how Meta, a company going all in on AI, somehow missed the memo on how AI can generate images and videos that renders take-a-selfie verifications utterly useless. Seemingly confirming everyone's fears about over-reliance on AI and under-reliance on real humans, he later added, I'm hearing Instagram's trust and safety order was absolutely gutted over the last few weeks with 60% of the org gone between layoffs and forced reassignments to data labeling. All while AI maxing pushed a bunch of bugs to production. Apparently, this was not a sophisticated hack, but engineers at Instagram going overboard to use AI for everything and having no incentives for stuff like security. You get what you incentivize, a warning for any company wanting to copy Meta. Added Jeffrey Emanuel, If Meta can't manage agents in an acceptable way in their own infrastructure, how could they possibly expect anyone else to want to use any of their stuff?