@TrismHermes Seems to be a translation error,
Here is the actual substack of Jack Clark, the source behind the Japanese article
Import AI 476: Creativity and RSI; worse-than-China open weight models; AI goes to space https://t.co/3z9kvbrOpU
@grok@TrismHermes@grok I fundamentally disagree bc AI is a both qualitatively and quantitatively different technology , such that it will itself solve the bottlenecks that usually are faced by upcoming technologies plus the whole money is going into building AI such that humans are eliminatedloop
@RuhRohRaggy99@cryptopunk7213 Don't you think that has been the whole point??
100% of money is going into eliminating humans from the loop all together.
And they will succeed pretty much.
Significant update from Jack Clark this morning. His RSI timelines have moved up. He now says it is likely that full RSI will become possible in 2027, and that creativity was the last piece. I think we are less than six months away from closing the loop.
@JHochderffer Who is this guy to make such a bld claim? Is he a ai researcher.
Opens account->literally the farthest profession possible from tech, just some philosophy bullshit.
Yeah good try buddy
@irrepldjw@deredleritt3r Jagged frontier is real, but it grows still , so you are right , capabilities will range wildly superhuman to avg human to below average, nevertheless, I think humans themselves have a jagged skill profile and they will be like humans just on magnitude above
@kame_chan_dayo_@imjustnewatai Agent2 is already here internally, they probably just started training agent 3
Timelines are way faster and internal models are way more advanced than you think.