Andrew Lee
speaker
2,833 appearances
3 recordings
1 series
first heard Mar 2025
last heard 15 May
Andrew Lee’s voice in public audio — every appearance, attributed to the second.
Trend
recordings per month · last 12 monthsRecordings per month over the last 12 months — 2 in all, peaking in May 2026 with 1.
Appearances
I think so far we've been able to, uh, and our approach has been like, you know, maybe we'll make some prompting tweaks that'll try to like address issues in one model.
Well like not breaking it in the other model.
And I think I think so far that's mostly that's mostly worked.
I think over time the
APIs of these things have converged and the basic capabilities of these things have converged.
So my my hope is that will get easier over time, not harder.
But I could see us having some some like very model specific harness things potentially and then thinking about ways to do that in like a really modular way.
So it's like not a huge uh amount of overhead.
But uh yeah, definitely something on my mind.
Yep.
Yeah, like in i in the example of uh anthropic and open AI, for open AI, they have a very simple caching API, which is basically like they'll just cache any prefix for twenty-four hours and they do it automatically.
Anthropic has a much more explicit caching API, and you can only cache four points in your uh in your call, and there's a lot more code kind of making it happen.
So in this case we're kind of lucky in that like once you've done the work to make anthropic work.
Making open AI work is pretty easy, but yeah, in that case we do have different code to sort of translate our context to like a cacheable context in each case.
Uh
It is so hard to stay up to date on this stuff.
We we have the ability internally to like test models pretty quickly.
Um it it's harder to actually ship things in production because for example, like the way thinking blocks work is different across different providers and and like if you have bugs, you might have to tune the prompts and things.
So we we haven't shipped that many, but like we've tested GLM internally.
Um we've tested the Google models, KMA, Deep Seek, um probably some others I'm not thinking.
Showing 361–380 of 2,833 · page 19 of 142
← Previous
Next →