ArXiV has a new LLM policy (Screenshots with alt text so you don’t have to click through to the other place and see all the stupid responses)
I'm going to use this in a lecture on chain-of-thought for my NLP class
I post a thread about how AI agents are maybe not ready for prime-time and get attacked for being a shill? Ah, the joys of being an AI researcher on BlueSky
Can we all agree to not test really large AI agents on cybersecurity benchmarks? Like we get it. Turning off guardrails, insecure sandboxes, and persistence goals result in random 3rd parties getting hacked. Check. Time to move on.
AI being weird is back. (I love that the Official Bob Dylan website is provided as a source for this answer. I’m pretty sure the website does not address this particular question)
The big AI tech companies are all jumping on the “recursive self-improvement” hype train. It’s only a matter of time before we are all talking about the singularity and hard-takeoff again, though so far I haven’t seen anyone come out and directly say it yet as part of this hype cycle
The Vatican has been staying abreast on developments in AI for a long time. I met a priest in 2017 whose job it was to be informed about AI developments. They were very interested in the topic of "alignment" and it doesn't surprise me to it appear in the Pope's encyclical.
I've discovered a new thing I hate: AI-generated review rebuttals.
Is there significant daylight between Mythos and GPT 5.5? Feels like the US government regulation is ad-hoc, spurious, and maliciously targeted. If feels like a rubicon has been crossed. Next few weeks should be… interesting
April: Claude Mythos was too dangerous for release. 40 organizations got early access to fix their software security issues. June: Claude Mythos released as Claude Fable. Anthropic says they have installed safeguards against using it for hacking.
