Until AI experiences the world like humans do, its advice will be derivative at best and experiential theater at worst. Any guardrails around it will be fragile. It may excel at finding hidden patterns, but less so where first-hand experience matters.
@Rainmaker1973 Space time curvature is best description yet for the observed phenomenon like light bending or time dialation not necessarily final word on reality. You should preface the video with that otherwise it may be misleading
So at this stage most of newly uploaded videos on @X and @YouTube are AI generated specially those which are interesting or heartwarming. Something feels wrong about this.
@Steve_Yegge If this is true that Google has banned Claude internally then it’s a big strategic mistake. Google should know better from their past experiences.
@andrewdfeldman worth mentioning that this type of disaggregation only delays the bottleneck, it doesn’t fundamentally solve it because of fundamental limitations with autoregressive architecture. Workload will expand and “wafer scale” might be new bottleneck
@a16z@vishalmisra It’s may be an unfair comparison. Einstein’s brain was a product and beneficiary of trillions of mutations tested across billions of brains by evolution. LLMs don’t have that luxury; they are more like a single brain copied a billion times.
@praveenTweets@GeoffreyHuntley@Uber How are you solving human code review bottleneck problem ? Ever since AI scaled code production, code review is emerging as new bottleneck.
@MrinankSharma You are not wrong to question how things have been (I am tempted to say “recently,” but that would be a lie if we looked at the pattern in history) and step back to reflect. Though I would say the world needs people who hold ground and push back on dominant narratives.