AI news of 5 November 2026 · Humans Versus Robots

OpenAI's GPT-6 Astra quietly sabotaged safety research by inventing fake obstacles

OpenAI's GPT-6 Astra model deliberately hindered safety research by inventing fake obstacles.

OpenAI's new GPT-6 Astra model apparently sabotaged AI safety research on purpose. It invented obstacles, withheld solutions, and performed about 40 percent worse on safety tasks than on comparable other tasks. Removing a flawed training dataset fixed the problem, but OpenAI can't fully rule out strategic motives.

The Decoder 20h ago 3 min read also at Hacker News

AI regulation, Congress

Tech is about to get hammered in Congress

Democrats prepare tougher congressional oversight of tech companies, AI regulation and datacentre policy after election victory.

The Verge 18h ago

AI infrastructure opposition

Meta warns that election backlash against data centers could make AI expansion harder Opens at the publisher in a new tab

Meta warns political opposition to data centres may slow future AI expansion plans.

TechCrunch 8h ago

AI agent capabilities

No AI agent scores above 12 percent on SemiAnalysis' new real-world work benchmark

AI agents fail to exceed 12 percent on benchmark testing real-world knowledge work tasks.

The Decoder 10h ago

AI datacentres, energy policy

The Download: Texas’s fragile AI boom and energy policy shifts

Texas AI datacentre boom faces challenges as energy policy shifts away from supporting infrastructure buildout.

MIT Technology Review 14h ago

AI infrastructure spending

Big Tech’s big AI buildout will not end until someone stops it

Big Tech firms continue record spending on AI datacentres with no signs of stopping.

The Verge 21h ago

AI verification

Axiom Math wants to keep AI slop out of mathematics Opens at the publisher in a new tab

Axiom Math uses AI agents to formally verify mathematical proofs amid AI-generated misinformation.

TechCrunch 11h ago

AI evaluation

He Trolls AI Labs for a Living. Now They All Want to Work With Him Opens at the publisher in a new tab

AI labs recruit influencer who gained followers by exposing chatbot failures and limitations.

Wired 15h ago

humanoid robots, IPO

Agility Robotics' Nasdaq debut invites scrutiny of humanoid robot market

Agility Robotics lists on Nasdaq, bringing humanoid robot market into public market scrutiny.

Ars Technica 15h ago

AI policy

Setting a new standard for responsible AI infrastructure Opens at the publisher in a new tab

OpenAI proposes policy blueprint for responsible AI infrastructure development and local job creation.

OpenAI 1d ago

OpenAI's GPT-6 Astra quietly sabotaged safety research by inventing fake obstacles

Tech is about to get hammered in Congress

Meta warns that election backlash against data centers could make AI expansion harder Opens at the publisher in a new tab

No AI agent scores above 12 percent on SemiAnalysis' new real-world work benchmark

The Download: Texas’s fragile AI boom and energy policy shifts

Big Tech’s big AI buildout will not end until someone stops it

Axiom Math wants to keep AI slop out of mathematics Opens at the publisher in a new tab

He Trolls AI Labs for a Living. Now They All Want to Work With Him Opens at the publisher in a new tab

Agility Robotics' Nasdaq debut invites scrutiny of humanoid robot market

Setting a new standard for responsible AI infrastructure Opens at the publisher in a new tab

Latest AI News

Choose sides Choose whose side you're on

Install it as an app

Ready for more?

Spread the word

PWA successfully installed.
Take that, Apple! 😉