← Dwarkesh PodcastTrade on Hyperliquid →Trade on Hyperliquid →
“They're doing so well, but the economic impact seems to be dramatically behind. And it's almost like, it's, it's very difficult to make sense of how can the…”
@Dwarkesh Podcast
Today's eval-topping models have weak generalization, so AI revenue stays far behind benchmark hype
OPENAISHORTIDEA · NOT YET PRICED
20:32LONG
“it appears that Gemini have found a way to get more out of pre-training.”
@Dwarkesh Podcast
Gemini's pre-training edge lets Google extract more from scaling than rivals
GOOGLLONG@ $320.70
$360.84+12.52%
since call Nov 25
peak +15.7%
21:01SHORT
“At some point though, pre-training will run out of data. The data is very clearly finite.”
@Dwarkesh Podcast
Pre-training scaling hits a data wall; the next gains come from research, not just more compute
NVDASHORT@ $172.57
$198.98-15.30%
since call Nov 25
“companies that are fierce competitors starting to collaborate on AI safety. You may have seen OpenAI and Anthropic”
@Dwarkesh Podcast
If AI feels powerful, safety becomes the industry constraint, and the safety-first lab's positioning becomes its moat.
ANTHROPICLONGIDEA · NOT YET PRICED
TIMESTAMPS
1:56
20:32
21:01
59:01