Nathan Lambert
speaker
1,814 appearances
3 recordings
2 series
first heard Feb 2025
last heard 1 Feb
Nathan Lambert’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 Feb 2026 with 1.
Appearances
But it's a set of model weights and a configuration that is deployed to the users.
We see this with TikTok.
You open it.
I don't use TikTok.
Supposedly in like five minutes, the algorithm gets you.
It's locked in.
And those are language models doing recommendations.
I think there are ways that you can do this with a language model.
Within like five minutes of chatting with it, the model just gets you.
And that is something that...
people aren't really ready for.
Don't give that to kids, at least until we know what's happening.
Like a lot of researchers at these companies are so well-motivated and they definitely...
The likes of Anthropic and OpenAI are culturally so want to do good through this for the world.
And it's such a, I'm like, I don't want to work on this.
Because on the one hand, a lot of people see AI as a health ally, as somebody they can talk to about their health confidentially.
But then it bleeds all the way into this.
I'm talking about mental health and things where it's heartbreaking that this will be the thing where somebody goes over the edge, but other people might be saved.
There's things that as a researcher training models, it's like, I don't want to train image generation models and release them openly because I don't want to enable somebody to have a tool on their laptop that can harm them.
other people, like I don't have the infrastructure in my company to do that safely.
Showing 421–440 of 1,814 · page 22 of 91
← Previous
Next →