LESSWRONG
LW

All of mjt's Comments + Replies

There have been 3 planes (billionaire donors) and 2 have crashed
mjt10

I’m on East Coast US somewhat near VT, and would also love to hear more if you can disclose.

Reply
Apply to the Redwood Research Mechanistic Interpretability Experiment (REMIX), a research program in Berkeley
mjt31

One thing I never get about these types of research programs is how anyone with a “normal” job can reasonably participate. They say they encourage people with software backgrounds who want to get into AI safety but maybe don’t have any other AI experience. Okay, that’s me. I have a software job on the East Coast of the US, not related to AI, but I spend a great deal of my free time reading AI research and doing my own projects. My employer isn’t gonna allow me to take a couple months off to go do this thing I personally am very interested in and think I ... (read more)

Reply
1Dan Valentine
  Have you considered asking them about it? I've worked at several software jobs where this would have been no problem. I've also seen a few people take sabbaticals and there was no issue with it, their teammates generally thought it was really cool. One guy I know took a 1-year sabbatical to live in a van and drive around Europe.    This is all anecdotal and your situation may be different of course. I just wanted to add this data point as it seemed like you may be prematurely dismissing sabbaticals as some crazy thing that never happens in real life.
Why I think strong general AI is coming soon
mjt10

Is this research into 'actual reasoning' that you're deliberately being light on details about something that is out in the public (e.g. on arxiv), or is this something you've witnessed privately and anticipate will become public in the near future?

Reply
2awg
Here is a paper from January 2022 on arXiv that details the sort of generalization-hop we're seeing models doing.
1porby
Most of it is the latter, but to be clear, I do not have inside information about what any large organization is doing privately, nor have I seen an "oh no we're doomed" proof of concept. Just some very obvious "yup that'll work" stuff. I expect adjacent things to be published at some point soonishly just because the ideas are so simple and easily found/implemented independently. Someone might have already and I'm just not aware of it. I just don't want to be the one to oops and push on the wrong side of the capability-safety balance.