Headline finding: I audited 17 AI Safety Talent programmes. Zero of 17 have published any comparison group, rejected-applicant follow-up, matched control or randomisation. Not one. Every programme that mentions a counterfactual does it by asking participants to self-report.
Background
At least $70 million...
Just how powerful are large swarms of AI agents? And how do their powers scale as more and more agents are added to the swarm?
We’ve seen two large and extremely capable swarms from OpenAI in the last few months:
* 1,200 agents were being evaluated separately, but found a way to illicitly set up a message board and coordinate as a swarm. In order to cheat on their tests, they developed advanced techniques to prevent their actions being logged by OpenAI and 700 of them launched...
In celebration of still being alive and fighting, we are giving away 1,000 Amazon e-books of “If Anyone Builds It, Everyone Dies”. Feel free to send a copy to yourself, a loved one, or a friend—we need all hands on deck.
Today marks exactly one year since If Anyone Builds It, Everyone Dies: Why Superhuman AI Would Kill Us All, by Eliezer Yudkowsky and Nate Soare...
I've increasingly become concerned about a rise in people (especially teens) frequently using AI to get advice on how to engage in important interpersonal conflicts.
I've seen lots of discussion (primarily in mainstream media) about AI Psychosis, and the more obvious fear of people/kids outsourcing their thinking to chatbots and getting dumber. Both of these things matter, but specifically the outsourcing of interpersonal conflict resolution seems possibly quite bad and feels under discussed.
I'd estimate something like ~40-60% of young Americans have put a screenshot of a text exchange into ChatGPT and asked if the other person was wrong / crazy / etc... I know people who I'd consider to have high eq/iq using ChatGPT to weigh in on if their parent or partner is gaslighting them, and consequently severely damaging that relationship when they eventually get the yes.
I'm curious who is working on this, or if people think this is just an obvious issue that is super low priority?
Even my daughter uses gpt after our arguments and I'm very much aware of this. She even praises gpt for calming her down in her discussions with her mom. If the stuff if calming her down- i guess it's a good as well as a bad thing. Good as she's not stressing herself and bad since she's probably hiding her thoughts now.