FSD=full self deception: Tesla faces NHTSA engineering analysis of FSD, issues in low visibility conditions, possible recall - CNBC

"Available incident data raise concerns that Tesla’s degradation detection system, both as originally deployed and later updated, fails to detect and/or warn the driver appropriately under degraded visibility conditions such
as glare and airborne obscurants." - National Highway Traffic Safety Administration

Your car, Elon’s cloud compute: Tesla with GPUs and mega RAM will be a node in Musk's compute farm. - Not a Tesla App

Supermicro co-founder charged with conspiring to smuggle Nvidia AI servers to China, stock drops. - Financial Times

Nvidia to sell 1 million chips to Amazon by end of 2027 in cloud deal - Reuters

Jeff Bezos seeks up to $100 billion for industrial AI investment fund Project Prometheus - The New York Times

Blue Origin Seeks FCC Approval to Deploy 52,000-Satellite Orbital AI Data Center Network - The Wall Street Journal

OpenAI develops desktop 'superapp' to integrate ChatGPT and engineering tools - The Wall Street Journal

Anthropic Dispute Fractures Trump’s Silicon Valley Alliance - Financial Times

OpenAI is trying to do everything. Anthropic is perfecting one thing. - Sherwood News

Anthropic demanded changes to OpenCode, seemingly around plugins for using Claude Code Oauth. - GitHub

One could see a legit antitrust case around bundling Claude Code + lots of free API tokens to get training data and ice out OpenCode and Cursor, but what are laws anyway when the bosses are Trump and Bondi.

Cursor released Composer 2, a coding model scoring 61.3 on CursorBench (vs 44.2/38.0) and 61.7 on Terminal-Bench 2.0, priced $0.50/1M input and $2.50/1M output; faster default variant costs $1.50/1M input and $7.50/1M output. - Cursor

User Fynn on X discovered an exposed OpenAI/Cursor model path revealing Composer 2 maps to the Kimi K2.5 RL model (kimi-k2p5-rl-0317-s515-fast), exposing internal model provenance. - X (formerly Twitter)

why would you not use Kimi or another open-source frontier model

Big Tech pours money into midterms, targets Alex Bores, whose New York congressional race has become a bitter contest centered on artificial intelligence. - The Wall Street Journal

FBI is doing mass surveillance using the commercial tech built to serve ads. - Politico

Higher Oil Prices From Trump’s Iran War Risk Killing the AI Boom, WTO Says - Gizmodo

Dear AI, WTH is Ras Laffan and South Pars and what does it mean? - Bloomberg

Uber struck a $1.25 billion deal with Rivian to buy up to 50,000 autonomous electric vehicles from Rivian and will make an initial $300 million investment in the company. - Financial Times

Waymo says it's safer than human drivers, drove 170.7 million rider-only miles through December 2025 and reported 92% fewer serious-injury-or-worse crashes (35 fewer), 83% fewer airbag-deployment crashes (230 fewer) and 82% fewer any-injury crashes (544 fewer) versus human benchmarks. - Waymo

Alibaba and Tencent stocks get crushed as AI fails to boost profits - Bloomberg

Box CEO: Enterprise AI adoption shifts from hype to production-ready agentic workflows - Fast Company

Equinix achieves 50% ticket automation via AI-powered IT service management - TechRadar

AI Agents Boost Security Operations but Require Human Oversight - TechRadar

In testing, Rust-based Sashiko AI detects Linux kernel bugs with 20% false-positive rate, catching issues missed by humans - The Register

Google adds new vibe coding stack in AI Studio - Google

Google adds vibe design in Stitch. - Google

Cognizant study: 93% of US jobs face impact as AI shifts $4.5 trillion in labor tasks - Business Insider

Columbia Law School launched a course titled "Law of Artificial Intelligence" taught by Steptoe partner Michel Paradis, covering technical foundations, regulatory responses and intellectual-property implications to prepare lawyers for AI-related risks such as malpractice, sanctions and litigation. - Above the Law

Young Silicon Valley coders are developing AI chat agents that can impersonate relatives and causing some parents to mistake bots for their children . - The New York Times

A traveler stranded abroad said AI travel agents proved useful, though some users fear increased dependence on Big Tech companies. - Financial Times

Tinder to Scan User Camera Rolls with AI to Automate Profile Creation - 404 Media

Publishers are unprepared for AI-generated fiction. - The New York Times

AI replica of influencer Gracie Nielson draws 2.4 million views as virtual creator market heads toward $49B - Business Insider

Horny MAGA men rack up views of viral AI dream girl, soldier, and OnlyFans model - The Washington Post

Music Hacker Pleads Guilty to $8M Streaming Fraud Using AI-Generated Songs and Bots - Music Business Worldwide

AI-generated images and videos are inflaming animosity between Ethiopia and Eritrea. - Punch Newspapers

How Reddit CEO Steve Huffman used artificial intelligence to gain the upper hand. - Fast Company

We shape our tools, and then our tools shape us. - Big Think

Hot-pot robot can't stop, won't stop dancing - Reddit

Hot take: AI > crypto > prediction markets

AI can be very useful if people don't abuse it.

Crypto is almost always useless except for grifting, anything you can do with blockchain you could do it better in cloud with 'distributed-enough' systems like DNS and BlueSky. The grifting is the point.

Prediction markets corrupt everything they touch.

Follow the latest AI headlines via druce.ai on Bluesky