Y Combinator’s Paul Graham Says AI Companies Are Measuring Inference the Wrong Way — ‘You Get More Problem-Solving Per Token as Models Improve’
Graham says token counts are becoming inadequate for measuring AI inference as smarter models solve more complex problems with fewer tokens.
Share
Why It Matters
Smart AI models solve harder problems using fewer tokens, or building blocks. The old way of counting tokens does not show the real progress.
This read is generated by AI from raw market data and is for education only. It is not financial advice, not a signal, and not a recommendation to buy or sell.
Tickers in Play
Last session drawn · prints held ~15 minutes · licensed exchange data
Source: Benzinga · Read the original report at the publisher. Headline and figures shown for context; data may be delayed.
Disclosure: Educational and informational purposes only. Not financial advice. We educate, not advise. Market data may be delayed up to 15 minutes unless marked real-time. Past performance is not indicative of future results.