Very considerate. But also, can we trust them that they're not downgrading without saying anything? The incentives are there, to bill at Mythos prices for Opus costs.
Would be nice to have a distributed, independent AIs, each being trained their own way. Maybe it would have to be a really slow training process to keep costs low (years even?).
We absolutely can, we absolutely should, and we absolutely have been doing that.
When I get on a bus with other people, I can, with a fairly high degree of certainty, rest assured that the other passengers will not just randomly kill me to get something from me. That's because the people, just like me, have some level of comfort and their basic needs being met.
The further we slide away from that, the higher the risk for everyone. And to maintain that needs constant work, from everyone.
Amazing. I feel like you could write entire books about this new way. Think of the fame and fortune that awaits such visionaries! Surely this is the way of the future. The next "Andrej Karpathy", "Boris Cherny", "Peter Steinberger" are just waiting to be discovered.
As is often repeated here, "technology is a force multiplier".
AI has been multiplying the wrong things (for the common people) at an amazing rate, while the good things it multiplies have been few, arguable, and lagging behind.
Usually to be enthusiastic about new tech, it needs to be the reverse. New tech that multiplies the "good things" first while the wrong things catch up later (see Internet, cars, commercial aviation, smartphones).
Sure, fellow software engineers who spend 8+ hours a day writing code see it as the opposite, since AI _is_ seen as doing the "good things" (i.e. writing code). But eventually the bad things will catch up with us too.
Nah. At least with Uber the driver has self-preservation as an incentive to not just fuck around. What incentive would a freelance nobody have to not do the funniest shit possible inside a stranger's home at least once.