What I found interesting this time

“Bug and Architecture Fixes Quote,” n.d. https://social.mkj.earth/@mkj/116079166290620645

"Implementation bugs can generally be fixed, but protocol bugs can, in bad cases, be nigh impossible to fix while maintaining compatibility."
Great quote.

“ChatGPT ‘Physics Result’ Reality Check: What It Actually Did,” n.d. https://m.youtube.com/watch?v=3_2NvGVl554&pp=ygUWQ2hhdGdwdCBwaHlzaWNzIHJlc3VsdA%3D%3D

OMG another hype based OpenAI paper.

“Anthropic Found Out Why AIs Go Insane,” n.d. https://youtu.be/eGpIXJ0C4ds

Finally some innovative research on how to steer AI personas right on track.
I think this research opens vector for new possibilities. Yeah opens vector by finding vector HeHeHe…
I think similar research will come forward for other GenAI architectures like image generation etc. Maybe finally we will limit possibility of AI to generate child porn.

“Microsoft Error Sees Confidential Emails Exposed to AI Tool Copilot,” n.d. https://www.bbc.com/news/articles/c8jxevd8mdyo

"Under normal circumstances, organisations would simply switch off the feature and wait till governance caught up," Henein said.
"Unfortunately the amount of pressure caused by the torrent of unsubstantiated AI hype makes that near-impossible," he added.
…

“Introducing OpenPlanter, so You Can Keep Tabs on Your Government since They’Re Almost Certainly Keeping Tabs on You.,” n.d. https://x.com/shinboson/status/2024665932522766797.

“Your AI-Generated Password Isn’t Random, It Just Looks That Way,” n.d. https://www.theregister.com/2026/02/18/generating_passwords_with_llms/

Vibe coders generate passwords for their repositories and put them in the Github.
Such passwords are not trully random. Out of 50 generated ones 20 were the same XDD
Yeah. What a time to be alive!