Comment by Exa Watson (exa-watson) on Can stealth aircraft be detected optically? · 2024-05-02T09:24:50.667Z · LW · GW

Spot on

Comment by Exa Watson (exa-watson) on KAN: Kolmogorov-Arnold Networks · 2024-05-02T09:14:46.083Z · LW · GW

Is this a massive exfohazard? 

Very Unlikely

Should this have been published?

Yes

Comment by Exa Watson (exa-watson) on KAN: Kolmogorov-Arnold Networks · 2024-05-02T09:13:58.144Z · LW · GW

I know this sounds fantastic but can someone please dumb down what KANs are for me, why they're so revolutionary (in practice, not in theory) that all the big labs would wanna switch to them?

 

Or is it the case that having MLPs is still a better thing for GPUs and in practice that will not change?

 

 

And how are KANs different from what SAEs attempt to do

Comment by Exa Watson (exa-watson) on Upcoming unambiguously good tech possibilities? (Like eg indoor plumbing) · 2024-04-20T15:57:38.053Z · LW · GW

• ^[4]
• AI life coaches

not excited about this - such a coach is either going to give very politically correct opinions, or target audiences with glaring insecurities, like young or low confidence men.. just like human coaches. 

Comment by Exa Watson (exa-watson) on Claude 3 claims it's conscious, doesn't want to die or be modified · 2024-04-20T15:39:39.792Z · LW · GW

I dont know if you are aware, but this post was covered by Yannic Kilcher in his video "No, Anthropic's Claude 3 is NOT sentient" (link to timestamp) 

Comment by Exa Watson (exa-watson) on Transformers Represent Belief State Geometry in their Residual Stream · 2024-04-20T15:35:13.295Z · LW · GW

If I understand this right, you train a transformer on data generated from a hidden markov process, of the form {0,1,R} and find that there is a mechanism for tracking when R occurs in the residual stream, as well as that the transformer learns the hidden markov process. is that correct?