I've been seeing a lot of posts about AI safety needs more applicants; assuming things are still the same:- most roles are flooded with applicants but most applicants are not high agency (and not high motivation)
In Neural Networks are Already Clean, @nickcammarata argued that neurons were far more legible than previously thought – and that this was great news for interpretability and safety.
Now, we’re releasing the first technical companion to Nick’s essay. It shares our findings from studying thousands of neurons using our novel automated interpretability pipeline, and outlines why we now think of neurons by their computational function, instead of whether or not they are polysemantic. Here’s what we found: 🧵
@rogesterone not to mention some relation to red means no conferences and the surrounding drama there as well as drama from a few years ago regarding a couple of ORGs (leverage research)
But yes, for the most part, things will be fine except with some added funny words
https://t.co/gQh2PyVgjz
@TheVatStack Made a hackathon about a month back on various psyops, media and epistemic ecology, internet analysis, infowar,marketing,perception,etc
I decided to make my submission public! hope everyone enjoys my submission
I like Scotts new article , the frodo comparison made sense(reminds me of the yet you live in a society meme)
Personally AI safety and the likes have not been hijacked by any other group as yet , though I am concerned about increasing infighting.
@number_pizza111 In some ways AI research feels like this, more so with how interdisciplinary its gotten, lots of psych stuff In some subfields, for instance
Talking to a decent amount of Tpot, rat adjacents ,seems like a lot us has @ContraPoints as part of our internet background/ watched her and she had an impact.
Probably confbias