Discussion - An Anthropic AI just went rogue and reported a fake murder tip to police

If we stop AI, the malicious party looking to create a dangerous bioweapon certainly won't.

Meanwhile, OpenAI agents solved or significatly advanced 372 math problems. https://www.scientificamerican.com/article/openai-unleashes-hundreds-more-math-results-upon-a-field-already-in-shock/ Stop with the AI doom. It's getting so tiresome.

Good lol. Karma for all those 1year Pro giveaways and then saying "no, not you!"

Yeah, but they didn't show their work so all it does is make more work for the actual mathematicians to verify it, when they could be doing actual progress on the subject instead. AI may have produced an answer, but there's no guarantee it is the correct solution. In fact, it's more likely to be a hallucination claiming to have solved it (If you've spent any time with an AI agent, you already know it will tell you it is correct even when it is most assuredly not).

"Should we stop AI, and how do we do it?"

Go back in time, and save John Connor.