Aidan Clark
@_aidan_clark_
Followers
10K
Following
1K
Media
68
Statuses
2K
Qualitative Mathematics @OpenAI Ex: @DeepMind, @BerkeleyDAGRS Hae sententiae verbaque mihi soli sunt
Joined November 2020
Man this made me wanna pick up WC3.
6
0
32
I think it’s easy to be mad at the uninformed, but we should probably save that anger for the obviously so. This sort of thing is a bit subtle.
2
0
10
I’m not upset the ultra wealthy exist for many reasons, but the simplest is that I can’t think of any system which disallows this but which also seems to work! If such a system existed, it seems better than what we have!.
2
0
14
It’s actually quite nuanced why it seems so hard to build a system encouraging what US capitalism does without creating the same level of inequity. Not getting why those left behind by our current system might be mad that someone has amassed so much wealth is similarly dumb AF.
This is the ultimate litmus test for retardation. If you think “Wow look at all the wealth he stole. This is bad” You are retarded. If you think “Wow, he created something so valuable that millions were willing to pay for it, and entire industries were revolutionized in a way.
4
3
45
My bar for AGI is when all human labor has been replaced by nanobot swarms, until then let’s just keep the hype in check.
5
3
93
Two days a month we turn off all the silent tests just to see who on here is imagining new models.
4
1
24
RT @nervouscomputer: it's hard to appreciate just how deep the talent pool at openai is until each wave of departures.
0
1
0
Said as a sad Californian resident who would love to be able to add a working Sten to his collection.
0
0
9
This is a good thread. I think if you’re the kind of person that would be cool with every American being mailed a machine gun, no questions asked, then probably it’s reasonable to say no AI safety is A-OK.
I didn't want to post on Grok safety since I work at a competitor, but it's not about competition. I appreciate the scientists and engineers at @xai but the way safety was handled is completely irresponsible. Thread below.
3
0
26
I can't recommend enough how refreshing it is to go on a spree of muting every anon account you see on your TL. It'll become so much less toxic.
12
2
138
I made the mistake of emailing a car dealership and I’ve been reached out to *literally* a dozen times in the last day. Lesson learned.
1
0
19
Presumably folks are already taking this as an opportunity to jeer, say “its so over”, and make references my ancient millennial brain will only half-get, but the reality is this model is incredible, we were so excited to share it, and we remain so— we just want to do it right.
11
2
148
Hi, .We’re delaying the open weights model. Capability wise, we think the model is phenomenal — but our bar for an open source model is high and we think we need some more time to make sure we’re releasing a model we’re proud of along every axis. This one can’t be deprecated!.
we planned to launch our open-weight model next week. we are delaying it; we need time to run additional safety tests and review high-risk areas. we are not yet sure how long it will take us. while we trust the community will build great things with this model, once weights are.
41
18
634
If you pay close attention you can tell one out of every three or four animals David Attenborough just makes up.
1
0
13
Buuuuuut I think that's because my coding work has a relatively high ratio of complexity vs lines-of-code needed. That's not always true, and if you're on the other end of the spectrum I think it probably even now can be a major accelerant.
0
0
6
This aligns with my own experience: I don't feel an AI tool would help accelerate the core of my workload, but there exist a distribution of tasks which it's a massive accelerator for (almost all of which look like "do something very different than what you normally do").
We ran a randomized controlled trial to see how much AI coding tools speed up experienced open-source developers. The results surprised us: Developers thought they were 20% faster with AI tools, but they were actually 19% slower when they had access to AI than when they didn't.
3
0
33
Credit to the guy trying to shoot his shot with a hilariously targeted-to-OAI cold call ending "You won't be waiting on alignment meetings, you'll be shipping", lotta layers here.
2
0
27
I agree with the conclusion of this thread, and some of the arguments but disagree that the argument is the cause of the conclusion. AI 2027 is wrong, as many likeminded manifestos are similarly wrong, because the authors don’t understand how people work.
I'll go ahead and put a flag in the ground here. This AI 2027 thing is wrong, the intuitions behind it are bad, and in 2027 these guys will publish a post-mortem that mostly amounts to "we were right, just got the dates wrong".
3
1
39