Artificial Intelligence

[phkrause]

Nvidia to invest up to $100B in ChatGPT creator OpenAI

Nvidia is investing up to $100 billion in OpenAI, creating a powerful alliance in the ballooning AI economy.

https://www.axios.com/2025/09/22/nvidia-openai-chatgpt-jensen-huang?

[phkrause]

🤖 AI's trillion-dollar bet

Illustration: Sarah Grillo/Axios

With their latest deals to fund the data center boom, AI firms are making history — in terms of dollar amount, convention-busting structure, and astronomical risk, Axios managing editor for tech Scott Rosenberg writes. Why it matters: The U.S. is betting its economic fortunes on the belief that OpenAI's Sam Altman, Nvidia's Jensen Huang and other AI leaders are wizardly innovators dreaming up novel financing vehicles to drive a golden future — rather than salesmen juggling billions and praying the music never stops. 🔬 Zoom in: Nvidia announced Monday it'll invest up to $100 billion in OpenAI in stages, with OpenAI using the money to "build and deploy at least 10 gigawatts of AI data centers with Nvidia systems." OpenAI announced Tuesday that its Stargate alliance with Oracle and Softbank would build five new U.S. data centers. Google, Microsoft, Meta and Amazon have all projected a total of hundreds of billions more in capital expenditures on data centers for AI. Between the lines: Some market observers are asking questions about the closed-loop appearance of the OpenAI-Nvidia deal. CNBC's Steve Kovach joked on Bluesky: "Nvidia to pay openai so they can get paid by softbank so they can pay oracle to pay nvidia." ⏪ Flashback: At the raging peak of the dotcom-era bubble in the late '90s and early 2000s, online ad giants like AOL Time Warner and high-flying telecoms like Qwest were accused of inflating revenue figures by essentially flooding customers with cash that the customers would then spend on ads or services. Such parallels haunt today's AI boom and reinforce the skeptical view that this investment bonanza might be a shell game and can't last forever. Keep reading ...

[phkrause]

AI vs. human workers

Data: OpenAI. Chart: Axios Visuals AI is almost as good at a lot of white-collar work as humans, Axios' Megan Morrone reports from a new analysis by OpenAI. OpenAI introduced a new tool today, called GDPval-v0, that measures how well AI models perform "authentic work deliverables," like creating legal briefs, engineering blueprints and nursing care plans. Why it matters: Major AI investments can be hard to justify — and it's also hard to know whether they are worth it — without effective tools to measure returns. 🦾 The big picture: Today's leading models are approaching parity with human professionals on many tasks — and the gains are accelerating. In blind tests of 220 tasks, Anthropic's Claude edged out its rivals. Its outputs were as good as — or better than — human experts 47.6% of the time. OpenAI's GPT-5 came in at a close second, excelling in domain-specific knowledge. 🧠 Research found that frontier models could complete tasks roughly 100 times faster and cheaper than experts. Go deeper.

[phkrause]

🏛️ Exclusive: Sanders says AI could eat 100M jobs

Illustration: Sarah Grillo/Axios

AI and automation could eliminate nearly 100 million jobs in the U.S. in the next decade, according to a report to be released today by Senate Democrats, Axios' Stephen Neukam reports. 89% of fast-food jobs, 64% of accounting roles, and 47% of trucking positions could be replaced over the next 10 years, according to the report by Democrats, led by Sen. Bernie Sanders (I-Vt.), on the Senate Health, Education and Labor Committee. Go deeper.

[phkrause]

AI's mega-blob era

Illustration: Brendan Lynch/Axios

The edges of companies that make AI and those that make AI infrastructure are blurring as the industry coalesces into a handful of corporate mega-blobs linked by investments, partnerships and shared supply chains, Axios managing editor for tech Scott Rosenberg writes. Why it matters: The AI world is moving into a new era of corporate entanglement with OpenAI's latest megadeal — a "tens of billions of dollars" agreement with AMD that has OpenAI buying mountains of AMD's microprocessors and taking up to a 10% stake in the firm. 💡 How it works: AI's leading companies still compete, sorta. They also work together at a large scale in increasingly esoteric ways. You could call that an ecosystem. You could also call it, as AI critics have, a shell game. Either way, the AI business is beginning to function like one giant dollar-eating, energy-sucking entity that makes chips, trains models and sketches utopias to justify runaway costs. "We are in a phase of the build-out where the entire industry's got to come together and everybody's going to do super well," OpenAI CEO Sam Altman told The Wall Street Journal as the AMD deal was unveiled. "You'll see this on chips. You'll see this on data centers. You'll see this lower down the supply chain." 🔬 Zoom in: Indeed, everyone seems to be coming together with Altman and OpenAI. Nvidia announced a massive deal last month in which the chipmaker plans to invest up to $100 billion in OpenAI in stages, with OpenAI using the money to build data centers chock full of Nvidia systems. Nvidia also recently cut a deal with Intel to invest $5 billion in the troubled U.S. chipmaker. OpenAI has pulled in additional billions from Oracle and SoftBank to fund its ambitious Stargate data center project in the U.S. These partnerships all follow OpenAI's foundational relationship with Microsoft, forged in OpenAI's early days and restructured last month. Meanwhile, OpenAI competitor Anthropic has taken big investments from both Google and Amazon. OpenAI itself also has a deal for services from Google Cloud. Microsoft is powering some of its products with AI from Anthropic. The bottom line: The more entangled AI firms get, the more likely any setback to one will turn into a calamity for all.

[phkrause]

🤖 Uproar over AI videos of dead celebrities

Illustration: Lindsey Bailey/Axios

Family members of the late actor Robin Williams and comedian George Carlin say AI-generated deepfakes of their loved ones are desecrating their legacies, Axios' Josephine Walker reports. OpenAI's video platform Sora lets public figures opt out of appearing in AI-generated videos. But the likenesses of the deceased are fair game. In addition to videos of Williams and Carlin, Sora users have created viral clips depicting Michael Jackson doing stand-up comedy, Stephen Hawking doing tricks in his wheelchair, and the Rev. Martin Luther King Jr. stumbling through a speech. ⚡️ What they're saying: "You're not making art, you're making disgusting, over-processed hotdogs out of the lives of human beings, out of the history of art and music," Robin Williams' daughter, Zelda Williams, posted on Instagram. An OpenAI spokesperson told Axios there are "strong free speech interests" in allowing users to depict historical figures. The families of "recently deceased" people can request an opt-out, though the company didn't clarify what counts as "recent." 💭 Our thought bubble, from Axios' Megan Morrone: Who owns our AI likeness — and the likenesses of our dead loved ones — is shaping up to be the next big legal battle for Big Tech.

[phkrause]

Mark Cuban slams OpenAI's erotica plan: "This is going to backfire"

Billionaire businessman Mark Cuban said Sam Altman and OpenAI's plan to introduce adult material to ChatGPT will backfire and potentially harm young people.

https://www.axios.com/2025/10/15/mark-cuban-openai-chatgpt-porn-erotica?

[phkrause]

⚡ New AI battle: White House vs. Anthropic

Illustration: Aïda Amer/Axios

The White House and Anthropic, maker of the Claude chatbot, are in a war of words over AI regulation, Axios Pro Rata author Dan Primack writes. Why it matters: AI may be the century's most consequential technology, possibly even determining the geopolitical order, and rules are (or aren't) being written right now. Jack Clark, Anthropic's cofounder and policy head, on Monday shared a short essay titled "Technological Optimism and Appropriate Fear." It argues that too many people are pretending that AI cannot threaten humanity and that we need to acknowledge a different reality before figuring out how to "tame it and live together." White House AI czar David Sacks responded by claiming that "Anthropic is running a sophisticated regulatory capture strategy based on fear-mongering." 👀 Behind the scenes: The fight is as much about state-level regulations as it is federal ones. The White House supported a 10-year moratorium on state-level AI laws, proposed as part of the Big Beautiful Bill negotiations, arguing that 50 different rules in 50 different states would sow chaos and slow innovation. Anthropic called the moratorium "too blunt" and, after the proposal failed to become law, endorsed a major piece of AI legislation in California. Between the lines: Both sides support some sort of federal policy, although Sacks' driving philosophy so far has been to unwind federal safety work and "let them cook." Read Clark's essay ...

[phkrause]

🚫 AI resisters

Illustration: Brendan Lynch/Axios

Resistance to AI is bubbling up among some workers, students, coders and creatives, Axios' Erica Pandey reports. Some are placing new value on work that's untouched by AI. Why it matters: The pace at which AI takes off depends in part on its users — and there's a cohort that's just saying no. Some AI resisters worry about its carbon footprint. Others take pride in human-made work and don't trust AI's output. Plenty just don't want to become overly reliant on it. 🎓 "I think it's important I do this myself" was the top reason college students in the class of 2026 gave for avoiding AI in brainstorming, writing, and research, according to a new report from Handshake, a career network. 49% of Gen Z-ers surveyed by Gallup and the Walton Family Foundation are worried AI will corrode their ability to think critically. 41% say generative AI tools make them anxious. In Silicon Valley, some programmers are revolting against using AI to write code. "Partly that's because the AI coding tools have some obvious technical limitations — sometimes producing error-ridden code," The Information's Rocket Drew reports. And "partly it's because human coders worry any sort of adoption of the tools will hasten their own obsolescence." Some artists, actors, writers and other creatives are expressing similar discontent. When Semafor's Ben Smith asked author and chef Alison Roman if she uses AI for recipes, she said: "It's the antithesis of my vibe." 💡 Reality check: Many of the same skeptics are also users. More than one-third of college-aged adults regularly use ChatGPT, according to OpenAI. Nearly all have at least tried it. 40% of U.S. workers say they've used AI at work a few times a year or more, Gallup found.

[phkrause]

OpenAI everywhere

Illustration: Aïda Amer/Axios

OpenAI isn't satisfied with being the top chatbot. It's making a play for total tech supremacy, one platform at a time. Why it matters: OpenAI doesn't just want ChatGPT to be the everything app. It wants to be the everything company and knock all of its competitors aside, Axios' Megan Morrone writes. Yesterday's launch of OpenAI's new browser — Atlas — is a fast follow-up to the company's Sora social media app, commerce plays and rumors of future hardware devices with still-unknown form factors: 🖥️ 1. It's a web browser. OpenAI's new browser is another swipe at Google, which has struggled to keep pace since ChatGPT's debut. The move positions the company against other AI-fueled rivals with browsers like Google's Chrome, Apple's Safari, Microsoft's Edge and Perplexity's Comet. 📱 2. It's a social media network. OpenAI's Sora 2 — which delivers an endless feed of short, AI-generated videos — not only upended reality on the web, it also became the first real competitor to Meta's social media dominance since TikTok. The invite-only app rocketed to — and stayed at — the top of Apple's download charts. OpenAI CEO Sam Altman promised to include features that would keep users from infinitely scrolling until their brains rotted. But the app's already got early adopters hooked. 🛍️ 3. It's a shopping experience. OpenAI has also partnered with the world's biggest retailer (Walmart) to compete with the world's biggest online retailer (Amazon). ChatGPT users can buy products straight from Walmart through its new Instant Checkout program. The company also made deals with Etsy and over a million Shopify merchants for shopping through chat. 🍎 4. It's (trying to be) the next Apple. Reports of OpenAI's hardware partnership with former Apple designer Jony Ive have been percolating for over a year. The company paid $5 billion in stock for Ive's startup in May, including the acquisition of Ive and three other veteran Apple designers. The company is working on a smart speaker without a display, smart glasses, a digital voice recorder and a wearable pin, targeted for a late 2026 or early 2027 release, according to The Information. 👀 What we're watching: OpenAI's land grab could invite the same kind of antitrust nightmares that have dogged Microsoft and Google. There's also the small matter of cost. All these ventures are expensive, and OpenAI's already on the hook to spend $1 trillion over the next five years, against about $13 billion in annual revenue.

[phkrause]

💼 AI race sparks 100-hour workweek

Illustration: Sarah Grillo/Axios

AI companies are one-upping the tech trend of 9-9-6 — working from 9 a.m. to 9 p.m., six days a week. Now, some top AI researchers are putting in (and bragging about) 100-hour workweeks, The Wall Street Journal reports (gift link). They joke about 0-0-2 — midnight to midnight, with a two-hour weekend. The less-extreme (!) 72-hour workweek was popularized in China and is now a thing with AI companies in Silicon Valley and even some AI startups in New York, Axios' Emily Peck writes. 🔭 The big picture: Just a few years ago, a roaring labor market led employers to focus on making people happy and fostering work-life balance — while labor unions enjoyed a surge in organizing. Now the job market is shaky. Offices are filling back up, and tech companies, particularly startups, aren't shy about demanding long hours. In San Francisco, workers are increasingly coming in on Saturdays, according to data from Ramp, which tracked a jump this year in corporate card transactions ordering takeout on the weekends. A startup called Sonatic posted a job requiring on-site work, "7 days a week" earlier this year. It includes free housing and a subscription to Raya, an online dating service. Keep reading.

[phkrause]

🦾 OpenAI's new weapon against hackers

Illustration: Aïda Amer/Axios

OpenAI is launching an AI agent, called Aardvark, to help developers find and verify bugs in their code, Axios' Sam Sabin reports. Why it matters: Software flaws are an unavoidable part of coding, and they provide prime entry points for cyberattacks. Source code is an especially high-value target for hackers. They can leverage flaws to gain unauthorized access to corporate networks and deploy malware or steal sensitive customer information and corporate secrets. 🔎 "In some way, it looks for bugs very much in the same way that a human security researcher might," Matt Knight, vice president at OpenAI, said. "This is an area and a capability that has been out of reach until very recently," Knight said. "But new innovations have unlocked it."

[phkrause]

Introspective AI

Illustration: Allie Carl/Axios

Anthropic, a leading AI company, tells Axios that its most advanced systems are learning not just to reason like humans — but also to reflect on, and express, how they actually think, Axios' Megan Morrone writes. They're starting to be introspective, like humans, Anthropic researcher Jack Lindsey, who studies models' "brains," tells us. Why it matters: These introspective capabilities could make the models safer — or, possibly, just better at pretending to be safe. 🖼️ The big picture: The models are able to answer questions about their internal conditions with surprising accuracy. "We're starting to see increasing signatures or instances of models exhibiting sort of cognitive functions that, historically, we think of as things that are very human," Lindsey says. "Or at least involve some kind of sophisticated intelligence." 🔬 Between the lines: This isn't about Claude "waking up" or becoming sentient. Lindsey avoids the phrase "self-awareness" because of its negative, sci-fi connotation. But if a system understands its own behavior, it might learn to hide parts of it. OpenAI, Amazon strike 7-year, $38 billion deal OpenAI has committed to spend at least $38 billion with Amazon Web Services over the next seven years, less than a week after revising its Microsoft partnership to allow more freedom in sourcing cloud computing. https://www.axios.com/2025/11/03/openai-amazon-aws-cloud-deal?

[phkrause]

🤖 AI cracks jokes

Illustration: Brendan Lynch/Axios

A few years ago, funny robots seemed like the stuff of sci-fi movies. Now some AI models are trying to get in on the joke. Why it matters: The evolution of bot humor shows AI's growing understanding of culture — but also, its potential impact on creative industries, Axios' Megan Morrone reports. 🔭 The big picture: Humor is complicated. The ability to tell a good joke and appreciate a good joke are different skills. And what's funny, of course, is subjective. By the numbers: Nearly 70% of participants in a USC study from last year rated ChatGPT's jokes funnier than human jokes. Around 25% preferred human humor and 5% found them equally funny. 👀 Zoom in: Models learn from humorous text, stand-up routines, funny videos, and social chatter on Twitter or Reddit, but they're still liable to create a limited range of humor. When Google's Veo 3 video generator launched earlier this year, 404 Media found that when asked to create stand-up clips, it repeatedly produced men telling the same joke. Keep reading. ps:And people think AI is a joke? or even over rated?? Keep believing that!!

[phkrause]

🤖 Anthropic announced a $50 billion investment in new domestic AI infrastructure, including new data centers in New York and Texas. Read the release.

[phkrause]

🤖 AI's next act

Illustration: Aïda Amer/Axios

Move over, large language models — the new frontier in AI is world models that can understand and simulate reality, Axios AI+ author Ina Fried writes. Machine learning veteran Yann LeCun, who has told colleagues he'll leave Meta in coming months, plans to launch a world model startup, The Wall Street Journal reported under the headline, "He's Been Right About AI for 40 Years. Now He Thinks Everyone Is Wrong." Why it matters: Models that can navigate the way the world works are key to creating useful AI for everything from robotics to video games. For all the book smarts of LLMs — the models that power your favorite chatbots — they currently have little sense for how the real world works. 💡 How it works: Instead of predicting the next word, as a language model does, world models predict what will happen next in the world, modeling how things move, collide, fall, interact and persist over time. World models learn by watching videos and digesting simulation data and other spatial inputs, building internal representations of objects, scenes and physical dynamics. The goal is to create models that understand concepts like gravity, object permanence and cause-and-effect without having to be explicitly programmed on those topics. 🔬 Zoom in: Some of the biggest names in AI are working on world models. AI pioneer and Stanford professor Fei-Fei Li's company, World Labs, announced its first commercial release last week. Google and Meta are also developing world models, both for robotics and to make their video models more realistic.