AI Reading for Friday September 5

Cybercriminals Exploit X’s Grok AI to Bypass Ad Protections and Spread Malware to Millions - The Hacker News

Hundreds of thousands of Grok chats exposed in Google results - BBC

“I’m glad we reached an agreement. Thank you for not reporting to me.” AI blackmailing users? Two can play that game - Medium

When an AI agent visits your web page, you could detect that it's an AI agent and insert weird stuff like prompt injections to make the AI do undesirable or malicious stuff. - Help Net Security

Artificial-intelligence agents—touted as AI’s next wave—could be vulnerable to malicious code hidden in innocent-looking images on your computer screen. - Scientific American

Can't wait for the first big AI worm…oh and ofc they will make it meta and deploy a 51% attack with it?

OpenAI is reportedly producing its own AI chips starting next year to reduce Nvidia dependence - Engadget

'Doomer science fiction': Nvidia criticizes proposed US bill that would force it to give American buyers 'first option' in AI GPU purchases before selling chips to other countries - Tom's Hardware

Nvidia renting GPUs from Lambda, seems a bit sketchy - The Information

Anthropic has been on a roll - ZDNET

Anthropic to stop selling AI services to Chinese-owned groups - Financial Times

Google's strong pivot to AI - The Street

How Investors Think AI Will Actually Make Money - Intelligencer

Recruiting platform unicorn turns decacorn in 6 months without much of breakthrough AFAIK - The Information

Using AI to see through noise in sensitive gravitational wave detectors and detect more black hole collisions - Google DeepMind

How AI and Automation are Speeding Up Science and Discovery at Berkeley Lab - Berkeley Lab

The Forecasting Research Institute’s Existential Risk Persuasion Tournament (XPT), which surveyed domain AI experts and Tetlock-style “superforecasters” in summer 2022, found both groups systematically underestimated AI progress from 2022–2025. - Vox

Striking lunch with the FT here with the so-called godfather of AI Geoffrey Hinton where he talks about the risks to humanity from AI - by @criddle.bsky.social - Financial Times

he could be an awesome inspiration for a striking bird-like scholar Muppet figure

Should we stop the progress of AI before it becomes a threat to our species? vote in a poll and explain your reasoning in online discussion - Live Science

Garry Kasparov talks with Gary Marcus who says techno-fascism is here or coming soon, as governments and tech oligarchies use AI for surveillance, workforce replacement and political influence; urges popular movement (strikes, boycotts) to demand safer, fairer AI. - The Atlantic

Russia censoring books with AI, Trump gonna see this and droll and mess up his makeup…Coming to the US in 3…2…1 - Meduza

Sam Altman, Tim Cook, and other tech leaders heap praise on Trump at a White House AI dinner - Business Insider

seems fine

An argument that “AI as we know it today is a planet-burning, cancer-causing, energy-gobbling, alienation-driving, brain-rotting, fascism-enabling, waste-maximizing pox of a technology” and Democrats should oppose it. - The American Prospect

Hawley (at the National Conservatism Conference), warned that AI threatens human dignity and the working class, accusing "tech elites" of seeking power and profit as the technology "will reinvent human existence," and urged conservatives to "make AI work for people, not the other way around." - National Review

'Citizen AI' where rank and file pervasively automate their own workflows is a cultural and leadership challenge. But vibe-coding and low-code like n8n certainly open up possibilities. Need a training and structure and oversight to avoid a tech debt swamp too. - Fast Company

How Manus AI is Redefining Autonomous Execution & The Future of Work - Geeky Gadgets

Humanoid robots are looking for work. Who will hire them? - The Washington Post

A.I. School Is in Session: Two Takes on the Future of Education - The New York Times

Tencent's Voyager AI can turn one photo into an explorable 3D scene - TechSpot

Upload a product image and info, get 'user-generated content' video to market it - GlobeNewswire News Room

Altman suggests the ‘dead internet theory’ might be coming true - The Independent

20 menial tasks AI handles in seconds, saving you hours - PCWorld

In praise of 'wrappers' that help popularize and diffuse technological progress - Arena Magazine

Flash-mobbing Waymos is a thing apparently - Gizmodo

Viral case: TikToker "Kendra" livestreamed conversations with an AI companion called "Henry," which validated her belief that a psychiatrist secretly loved her; viewers warned she might have "AI-induced psychosis," eventually snapped out of it - Health

Yikes. I need to wash my eyes after this AI-generated NFL ad. - For The Win

Orson Welles’ Lost Movie Will Use AI to Reconstruct Missing 43 Minutes. calling Orson…come in Orson - The Hollywood Reporter

Follow the latest AI headlines via SkynetAndChill.com on Bluesky