OpenAI has hundreds of contract workers reading real ChatGPT conversations and rating them on a scale of one to seven, partly to reduce flattery and human-like behavior, 404 Media reports. The prompts are anonymized but can still contain sensitive data. Users who don’t want their chats reviewed by humans have to actively disable the “Improve the model for everyone” setting, which is on by default.
Anthropic has told investors it will turn a profit for the second straight quarter, but the claim rests on an adjusted metric that leaves out costs like stock-based compensation. Gross margins run above 80 percent, according to the Financial Times, though that figure comes before revenue-sharing payments to partners like Amazon and the cost of training models.
Quarterly revenue jumped 14-fold from a year earlier to $11.5 billion, and the annualized run rate hit $65 billion at the end of July. SemiAnalysis analyst Joey Brookhart says investors expect $120 billion in annualized revenue by year's end and nearly triple that by the close of 2027.
Anthropic plans to go public on the Nasdaq at a possible valuation of $2 trillion or more. Rather than release its prospectus last week as expected, the company shared the documents with only a small group of investors at first. At the same time, CEO Dario Amodei publicly called for slowing down AI development. OpenAI CEO Sam Altman and Elon Musk backed the call. Altman told Fortune that OpenAI won't go public this year.
The Clay Mathematics Institute (CMI) has officially weighed in on the Navier-Stokes problem, one of seven Millennium Prize Problems announced in Paris in 2000, each worth $1 million. The problem asks whether the equations governing fluid motion in three-dimensional space always have a smooth, complete solution.
The CMI says "the Navier-Stokes problem has apparently been settled" and hopes "to see waves of new human understanding unleashed as the innovations behind this work are analyzed and interrogated." The solution is now under review. "The process is deliberately unhurried, but we will provide updates," the institute writes. It also notes that the "increasing ability of new technologies to accelerate mathematical research has heightened this sense of anticipation."
The potential solution comes with a bitter dispute. Mathematician Tristan Buckmaster accuses OpenAI of redirecting resources toward the problem after rumors about his research leaked, using his drafts in training data, and rejecting his co-author, Levent Alpöge, who works at Anthropic, from authorship.
Sam Altman is doubling down on his call to pump the brakes on AI, at least a little. He wants a uniform, nationwide safety framework for advanced AI but says the industry shouldn't wait for lawmakers, even though international coordination needs U.S. government support. To that end, OpenAI now sets explicit safety protocols before training runs that could bring major capability jumps, and Altman hopes other companies "will learn from our approaches and propose their own." He also makes clear that "pacing" doesn't mean "stopping." Progress will "continue to be" rapid, just not as fast as it could.
Pacing will be well worth this cost; no amount of American competitive pressure should justify recklessness, or let capabilities get ahead of alignment and monitoring.
Sam Altman, OpenAI
Over the weekend, Anthropic CEO Dario Amodei, former Deepmind CEO Demis Hassabis, Microsoft CEO Satya Nadella, and Elon Musk all backed better-controlled AI development. According to The Information, OpenAI, Anthropic, and Google have been talking for months about self-regulation through an independent oversight body. Smaller players like Cohere worry this could turn into a cartel.
Elevenlabs has released Music v2.5 for ElevenMusic. The company says generated songs now sound fuller and more natural. In a blind test with 47,885 comparison pairs, listeners preferred v2.5 most of the time, especially for R&B, Soul, Hip-Hop, Rock, and orchestral music. Users retain the rights to their tracks. The free tier includes five lossless downloads per day and the Pro tier 400 per month.
Commercial use is allowed depending on industry and purpose, though the free tier requires attribution. Downloads of tracks based on other artists' songs are blocked, and imitating existing musicians isn't allowed. Music v2.5 is also available through Elevenlabs' API, while v2 remains accessible.
Elevenlabs recently signed a licensing deal with Universal Music Group, but says it only applies to future, separate products, not Music v2.5. The company says it trained its existing Music models on "licensed stems and music" without detailing the data. That still sets it apart from competitor Suno, which was sued for training on copyrighted content without rights holders' consent. Suno also released new models this week.
The AllSpark team has released Iris-mini and Iris-pro, two open-source search agents built on Qwen models that lead benchmarks among open-weight models in their size classes. According to the paper, the training data and models also improved performance on tasks they were never trained for, including general tool use and office work.
A law professor spent two years testing how an AI ban, unguided AI use, and structured training affect student performance. The group without AI finished last both years. “I was wrong,” the researcher writes, who had assumed that AI without guidance would do more harm than good.
