noting that input:output:cached is 10:50:1 for astra and luna but 10:50:0.5 for sol. this doesn't mean a lot without "tokens per task" information but it's still interesting for there to be a "dip" like this instead of a monotonic change in one direction or the other
what is the…epistemic status, for lack of a better way to put it, of the probabilities? what do they mean? what (probabilistic) guarantees do we have about, say, the responses to
- is the capital of france paris?
- it is august. is it raining in paris?
(forgive the examples; they're probably not semantically the sort of thing jev is trained to work on. but i figure the point translates to various kinds of questions that come up in "inner loop of agentic pid controller" contexts)
a normal text-generating model if asked to produce a number will also do that just fine. i assume in jev's case it was actually rled to essentially learn to express priors over things using its implicit world model, which definitely ought to help, but can we say more?
In keeping with our commitment to user responsibility, following the official launch of V4.1 Flash and prior to the release of V4.1 Pro, all requests to the Pro model will be routed to V4.1 Flash and billed at Flash's price.
just in terms of user perception when selling this sort of service, this is what they call a "good look"
claude -p "${curl $url}\nreply only with a substantive comment in 'statement. question?' form" | dd conv=lcase
as a service
i wonder how hard it would be to write a decent classifier for individual comments like this (in practice you can just look at the account history which would make it an easy problem)
reply