Ben Horowitz
speaker
15,697 appearances
54 recordings
1 series
first heard Dec 2024
last heard yesterday
Ben Horowitz’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 — 43 in all, peaking in Jul 2026 with 8.
Appearances
Safety was like, okay, let's make these things not say nasty stuff that people don't like.
matched
And so
matched
the approach that was taken was kind of uh surface around the edges.
matched
Okay, we'll put some filters on this guy and we'll RLHF around the edges.
matched
But if you get deep into the thing, you'll be able to get the bad words out.
matched
But if somebody wants to go through that to hear bad words themselves, who cares?
matched
But then, you know, when you get into, okay, now these things are really good at cyber hacking and
matched
other kinds of ideas.
matched
Now you need kind of a more architectural idea where the model itself knows not to reward hack in a way that's going to be dangerous and so forth.
matched
And do you feel like
matched
Yes, we can make progress against that fast, or is that a really hard different category of problem?
matched
Or how are you thinking about that?
matched
usability.
matched
Right.
matched
And when you think about and we don't have a really great community yet amongst the uh soda models, but it seems like those kinds of ideas
matched
you and Google and Anthropic and SpaceX would want to share and meta now as opposed to okay, this is a proprietary idea that's a way to keep these models safe, since you're all on related architectures, or how do you see that unfolding?
matched
Or is everybody going to do it independently?
matched
Very interesting.
matched
very cleverly.
matched
Okay, so I've got a comment and a question on a th I would say there's a third thing that we learn which is
matched
Showing 501–520 of 15,697 · page 26 of 785
← Previous
Next →