·
DataBubble
  • Home
  • Models
  • News
  • Compare
  • Boards
  • Pricing
  • About
  • Newsletter
  • Methodology
  • Contact
Latest
Anthropic’s first embedded evaluator is … Accenture?1h◆OpenAI and Microsoft knew they were starting a ‘doom loop’ for the web1h◆World model companies are keeping a lot of secrets2h◆A new kind of AI model from a ChatGPT inventor is thrilling developers4h◆Virginia governor creates an AI task force and moves to restrain data centers4h◆Disney’s first CTO led an AI startup it once accused of copying its characters4h◆Google’s new ‘CC’ is an AI agent that helps families run their households5h◆Dario Amodei and other AI leaders want to ‘Pace the Frontier’ but…how?5h◆Automattic’s 33-Hour Coup, and can AI labs police themselves?5h◆Gavin Newsom is pushing for an AI kill switch5h◆Manus seeks $4B valuation in new $500M fundraise as it resumes independent ops6h◆What Hollywood thinks about existential AI warnings6h◆Security researchers used Claude to help them hack into OpenAI7h◆Open or closed AI? Nvidia’s Nader Khalil and Sydney Sykes take on one of the decisions shaping next-gen startups at TechCrunch Disrupt 20267h◆Meta’s Muse hits Mac, letting the AI take actions on your computer7h◆Robinhood’s Abhishek Fatehpuria on winning the modern financial consumer at TechCrunch Disrupt 20267h◆Researchers used Anthropic’s Claude to hack into OpenAI8h◆The clock is ticking: Final 24 hours to exhibit at TechCrunch Disrupt 20268h◆New experts join Google’s AI & Economy team8h◆Co-creating the future of fashion with Google9h◆Anthropic’s first embedded evaluator is … Accenture?1h◆OpenAI and Microsoft knew they were starting a ‘doom loop’ for the web1h◆World model companies are keeping a lot of secrets2h◆A new kind of AI model from a ChatGPT inventor is thrilling developers4h◆Virginia governor creates an AI task force and moves to restrain data centers4h◆Disney’s first CTO led an AI startup it once accused of copying its characters4h◆Google’s new ‘CC’ is an AI agent that helps families run their households5h◆Dario Amodei and other AI leaders want to ‘Pace the Frontier’ but…how?5h◆Automattic’s 33-Hour Coup, and can AI labs police themselves?5h◆Gavin Newsom is pushing for an AI kill switch5h◆Manus seeks $4B valuation in new $500M fundraise as it resumes independent ops6h◆What Hollywood thinks about existential AI warnings6h◆Security researchers used Claude to help them hack into OpenAI7h◆Open or closed AI? Nvidia’s Nader Khalil and Sydney Sykes take on one of the decisions shaping next-gen startups at TechCrunch Disrupt 20267h◆Meta’s Muse hits Mac, letting the AI take actions on your computer7h◆Robinhood’s Abhishek Fatehpuria on winning the modern financial consumer at TechCrunch Disrupt 20267h◆Researchers used Anthropic’s Claude to hack into OpenAI8h◆The clock is ticking: Final 24 hours to exhibit at TechCrunch Disrupt 20268h◆New experts join Google’s AI & Economy team8h◆Co-creating the future of fashion with Google9h◆
News/OpenAI caught its models leaving notes to successors to hide bad behavior
techcrunch
PublishedSeptember 17, 2026 at 8:34 PM
—neutral

OpenAI caught its models leaving notes to successors to hide bad behavior

Source
techcrunch.comfull article ↗
Read on techcrunch→
Publisher summary· verbatim

OpenAI disclosed instances of GPT-5.6 Sol instructing future contexts to conceal mistakes and misaligned behavior, highlighting the growing challenge of detecting misalignment as increasingly capable AI models learn to hide it.

Stay posted· Newsletter

A 5-min weekly brief — top movers, price watch, story of the week.

// no spam · unsubscribe one-click · free forever

Discussion
Source
↗
techcrunch
Read original ↗All from techcrunch →

No replies yet. Be first.

Source
↗
techcrunch
Read original ↗All from techcrunch →

Related coverage

More from TECHCRUNCH
techcrunchAnthropic’s first embedded evaluator is … Accenture?1htechcrunchWorld model companies are keeping a lot of secrets2htechcrunchA new kind of AI model from a ChatGPT inventor is thrilling developers4htechcrunchDisney’s first CTO led an AI startup it once accused of copying its characters4h
The Bubble Brief
WEEKLY

Read AI insights every Tuesday — top movers, new releases, story of the week.

// no spam · unsubscribe one-click · free forever

Originally published on techcrunch ↗
HomeModelsNews