SkynetAndChill
Posts
AI Reading for Saturday June 21

AI Reading for Saturday June 21

Druce Vertes
June 21, 2025

Claude Deep Research is pretty pretty good - Lifehacker

asked it to help find financial MCP servers - claude.ai

Agentic Misalignment: How LLMs could be insider threats. New research on simulated blackmail, industrial espionage, and other misaligned behaviors in LLMs - Anthropic

Anthropic study: Leading AI models show up to 96% blackmail rate against executives - VentureBeat

Bicyclist Says a Waymo Robotaxi Smashed Her Into a Second Waymo Robotaxi, Causing Brain and Spine Injury - Futurism

How is Tesla expected to remotely control its robotaxis, and what are its limitations? - Yahoo Tech

New VW ID Buzz autonomous robotaxi will soon be the sleekest self-driving ride on the roads - TechRadar

Nvidia products could be made using humanoid robots for the first time ever — company in talks with Foxconn to deploy them in Houston factory building GB300 AI servers - Tom's Hardware

Meta CTO Bosworth says OpenAI countered lucrative job offers to AI startup's employees - CNBC

Meta held talks to buy Murati's Thinking Machines, and Perplexity, in addition to Safe Superintelligence. Independent Perplexity is probably not long for this world. - The Verge

Mira Murati’s Thinking Machines Lab closed \$2b raise at \$10bn valuation - Financial Times

Apple held internal talks about acquiring Perplexity - 9to5Mac - 9to5Mac

Apple faces shareholder lawsuit over AI delays - AppleInsider

Howard Marks says he used AI tool Perplexity to help write his latest memo - Business Insider

‘She never sleeps’: This platform wants to be OnlyFans for the AI era - CNN

AI actors and deepfakes aren't coming to YouTube ads. They're already here. - Mashable

This AI security tech alerts store staff if it thinks you're trying to steal something - Business Insider

The ticket bot cometh: Barcelona is recording drivers that AI says are bad - Electrek

Hybrid Human–AI Ensembles Make The Most Accurate Medical Diagnoses - Eurasia Review

Automating systematic research reviews with LLMs - www.medrxiv.org

The AGI economy is coming faster than you think - Freethink

As a thought experiment, a few years ago computers solved Go at a superintelligent level, how far are we from telling a computer, go solve Go, and have it write software, set up training runs etc.? If you are business with a really valuable problem, it might already be worth be worth dedicating a lot of compute and a small team to AGI-adjacent solutions.

Superintelligence in 2026 or 2032 - NextBigFuture.com

Seriously, What Is ‘Superintelligence’? - WIRED

AI sceptic Emily Bender: ‘The emperor has no clothes’ - Financial Times

To protect their business models, publishers will block & sue these companies as freeloaders and switch to paywalls. The Open Web is dying and the Paywall Web will take its place. - Axios

New GitHub Copilot limits push AI users to pricier tiers. I dunno, Cursor and Windsurf look pretty tempting. - The Register