An internal ICE document reveals the agency is purchasing access to AT&T's raw call data and analytical services — a program experts say may sweep up trillions of records. Lawmakers are calling it mass surveillance by another name, and it's happening without broad public awareness.
A new study maps the AI infrastructure buildout onto historical redlining patterns with striking results: majority-Hispanic neighborhoods face 2.6x the concentration, majority-Asian neighborhoods 4.2x. The costs of powering the AI economy are being distributed anything but evenly.
Claire Vo of ChatPRD documents a surge in AI token fraud hitting real products in production. As AI platforms scale, the attack surface for credential abuse and token theft is growing faster than defenses — a problem the industry has largely ignored until now.
Postgres shipped two months late this year — and AI is to blame. The team had to address a flood of issues driven by the explosion of AI-generated code hitting production databases. The release still landed 200+ bugfixes, while AI agents rewrote 185,000 lines of C into 340,000 lines of Rust nearby.
Ark committed $90M to Anthropic's historic $81.34B Nasdaq debut — the largest IPO ever. Crypto futures markets are already pricing the stock down ~5% from its first-day close, suggesting early euphoria may be cooling fast.
The FBI ran a double-sided sting — impersonating both the ransomware victim and the threat actors — to catch an alleged fraud scheme in progress. It's a glimpse into how federal investigators are getting creative as ransomware operations grow more sophisticated.
Mollick flagged what may become a defining tension of the agentic era: companies blocking customers from using their own AI agents to manage their own accounts. It sounds absurd until you realize every major institution has liability, compliance, and authentication reasons to hate it. The customer-vs-company AI arms race is just getting started.
Anthropic's IPO hit Nasdaq at an $81.34 billion valuation — the largest offering ever — with its first trade coming in well above the offer price. The Neuron called it in real time as it happened. For context, this is a company that describes existential AI risk in its own charter. The market, apparently, is fine with that.
Hassabis retweeted a user reporting that Gemini Neon — Google's newest model — beat Fable 5.2 on nearly every benchmark they tried, adding pointed commentary about being "underestimated." With Neon arriving just as Anthropic dominates headlines, DeepMind is clearly not content to let the IPO moment belong to anyone else.
The Kobeissi Letter made a bold call amid the tech selloff: this bear market will be remembered as the best buying opportunity of the decade. Gold fell below $3,550, silver cratered, South Korea's KOSPI dropped 5%, and AI-adjacent stocks shed trillions — but the macro letter crowd is treating the dip as a generational setup, not a warning.
Two stories this week deserve to be read together: ICE paying AT&T for access to a surveillance dragnet covering potentially trillions of call records, and a study showing data centers are disproportionately sited in minority neighborhoods. The infrastructure of AI — both the compute layer and the intelligence layer — is being built on and aimed at communities with the least political power to push back.