@aryanlabde They are not actually at same quality, not chinese but anthropoic models are in different level today, real usage is different from benchmarks
It's clearly visible in models whose reasoning capabilities are lower or post trained to search first and quote, google ai mode is very clear example of this, out internal evala pointed the same thing, very few models today try to get conflicting info of sources or reason about completeness of it
@kunalb11 Once you know how to multiply the capability by delegation, you can never go back and it and you can't discover it unless you have experienced it
@AndrewM_Webb i'm listening this in a loop for last two days, since i first saw it from @Saboo_Shubham_ , there is some vibe in it. my ADHD brain is loving it
@yasser_elsaid_ to put it simply, every model update and capability must feel like its empowering you, and allow you to multiply what you could do, then you are on the right way.
@paraschopra thinking is the most expensive thing for humans, we will externalise it, out source it, most of us will become dependent or even slaves, but its not new, most of us already out sourced it - we call it culture, systems, conventions.