AI stock selloff after midterms & Meta scam ads and deepfakes - AI News (Nov 6, 2026)

Trillions of dollars gone in a single trading session, and the AI trade at the center of it all. We'll get into what spooked investors so badly. Welcome to The Automated Daily, AI News edition. The podcast created by generative AI. I'm TrendTeller, and today is November 6th, 2026. Let's get into it.

AI stock selloff after midterms

First, an update on the post-election market story. The turbulence we flagged in premarket trading turned into a full rout. On November 5th the Nasdaq fell roughly five and a half percent and slipped into correction territory. By one estimate, that was the index's worst single day on record measured by lost market value. The S&P 500 dropped 4.4 percent, its worst session since April 2025, and the Dow shed 1,833 points. The Guardian puts the total damage at more than 3.2 trillion dollars across the US market. Nvidia fell more than 9 percent, and AMD, Micron, CoreWeave and GE Vernova also took heavy losses. The worry is that a Democratic Congress could bring tighter oversight of data centers, which might slow the construction boom that drives demand for chips and power. The selling spread to Tokyo, Seoul, Taipei and Europe, while investors moved into Treasurys, pushing the 10-year yield to its lowest level since early September. Elon Musk alone reportedly lost about 113 billion dollars in two days. The key point is that AI stocks now carry so much weight that their swings move the entire market. Next week's inflation and jobs data could shape whether this is a shakeout or the start of something longer.

Meta scam ads and deepfakes

Staying with ongoing stories, Mark Cuban has sharpened his criticism of Facebook. He says the platform knowingly allows deepfake ads that use an AI-generated version of his face and voice to sell fake products, and that it keeps collecting the ad money even when the scams are obvious. He says victims regularly message him after being fooled. The comments land alongside Reuters reporting on internal Meta documents, which estimated that roughly 16 billion dollars, about a tenth of Meta's 2024 revenue, came from scam and prohibited ads. The documents also suggested many high-risk ads stayed live unless the system judged them at least 95 percent likely to be fraud. Meta calls the documents selective and points to large-scale takedowns. Still, the reporting raises an uncomfortable question about how much of the ad business depends on fraud, especially as AI makes fakes more convincing.

METR on Anthropic's AI acceleration

On AI building AI, there's a new development from METR. Its latest assessment estimates that Anthropic is now improving its models about twice as fast as it did between early 2024 and late 2025. In practical terms, that's close to two years of past progress packed into one. METR attributes most of the speedup to Anthropic using AI in its own research, not just to more compute. It found no obvious bottlenecks for the next year and warns that the effect could compound. There's a large caveat, though. Anthropic redacted most of the evidence METR considered most informative, so outsiders have limited ability to check the conclusions.

Limits of autonomous AI research

A grounded blog post offers a useful counterpoint. The author tested autonomous research agents on side projects and found they repeatedly missed the most relevant papers, picked low-value experiments, and chased trivial ideas. They did make quick progress on narrow tasks with a clear metric and fast feedback, but usually only after a human decided what to optimize and how to set up the problem. So AI may be speeding up research at a lab like Anthropic, yet judgment on open-ended problems still seems to rest largely with people.

Closed models dominate price-performance

Turning to model economics, Village's LLM Atlas suggests open-weight models are losing ground on value. Closed models now offer the best price-performance at almost every budget level. Open models are the best buy at only 3 percent of price points, down from 41 percent in June. The analysis argues that closed labs can distill cheap models from their frontier systems, ship model families that cover every budget, and run inference more efficiently at scale. Recent releases from OpenAI and Anthropic took back much of the low end. By Village's measure, this is the weakest showing for open models so far.

AI hype across App Store

In consumer apps, a study of nearly 179,000 apps that appeared on App Store charts across 173 countries and regions finds that about one in ten mentions AI in its listing. The share is highest in design, photo and developer tools. Apps that sell subscriptions or in-app purchases are about three times as likely to mention AI, which suggests AI is often a selling point for paid upgrades. Users are pushing back: subscription pricing is the top complaint in low-star reviews, ahead of bugs and login issues. Upkeep is another weak spot. One in five charting apps hasn't been updated in at least a year, and about one in ten has gone more than three years without an update.

Sci-fi that predicted real AI

Finally, a lighter one. An essay ranks science fiction by how well it predicted the past decade of AI. The winner is Ted Chiang's The Lifecycle of Software Objects. It doesn't foresee chatbots specifically, but it captures how AI systems are trained, maintained and turned into products. Ken Liu, Kim Stanley Robinson, Heinlein and William Hertling also make the list. The author's point is that the most accurate stories focused on the everyday work of training, feedback and deployment rather than a sudden leap to superintelligence. They also note that no work written before 2017 clearly anticipated the transformer breakthrough.

That's it for today's episode of The Automated Daily, AI News edition. Links to all the stories we covered are in the episode notes. I'm TrendTeller. Thanks for listening, and see you tomorrow.

AI stock selloff after midterms & Meta scam ads and deepfakes - AI News (Nov 6, 2026)

Our Sponsors

Today's AI News Topics