Aaron Levie
speaker
876 appearances
1 recordings
1 series
first heard Apr 2025
last heard Apr 2025
Aaron Levie’s voice in public audio — every appearance, attributed to the second.
Trend
recordings per month · last 12 monthsNo recordings in the last 12 months.Older appearances are listed below; set an alert to hear about the next one.
Appearances
You know, there there's a lot of things that you can do where where um and actually this was this started to become an interview question that I had uh for people coming into the AI team.
I was like, Do you want to fine tune a model?
And and if they said yes, I had to beat that out of them and
Make sure that they weren't like deeply committed to it.
Um, because it was just like like AI is happening so fast.
I can't have somebody who's like so committed to one particular architecture right now.
We need range of motion, we need to be able to like like flip out something, you know, instantaneously.
Like internally, we kind of have a a a mental model of if anybody has a breakthrough in the model space that we have.
Built scaffolding or plumbing around to mitigate the lack of that breakthrough previously.
We should just like kill our stuff and just adopt whatever that breakthrough is.
And like, you know, maybe not like within 10 minutes, but but certainly the next couple sprints that we're planning for, we should be thinking about like, oh, we were doing some crazy optimization on images, the, you know, with some technology.
Like if the model can now do that, just get rid of your stuff and just keep moving up the stack.
And
And and because what you don't want to do is ever have this culture set in of like our thing is better because just you just happen to have the people that built that thing be so committed to it that you're you're kind of missing the new breakthroughs that that are actually happening in the space.
like there's like lots of stuff.
You have to do.
Like we have uh we we have uh one of our top AI guys.
Um we think we invented a uh a thing we call e-rag um that does this sort of rag enhancement that that is able to um uh do better a better job at like entity extraction from uh from a large uh you know set of chunks of data and embeddings.
Um so we'll we'll do a we do a lot of stuff ourselves, but it's only in
service of of again kind of like optimizing around the models right now uh in in areas that the models are not not particularly good at but but nothing that we think is is gonna lock us in place, which is why fine tuning I've just been against, um, or built you know, training our own model.
Showing 581–600 of 876 · page 30 of 44
← Previous
Next →