Gary Marcus

Rank 13 of 47

|

Score 94

·

10m

@ben_j_todd @tobyordoxford huge contrast with symbolic systems, and part of why i have always favored integrating them into neurosymbolic systems:

·

14m

Now show me the same graphs but for symbolic systems like calculators, spreadsheets, gps navigation systems, and databases.

·

18m

@sophia_surfs now show me the same but for calculators

·

30m

@ArtTheVandelay lol

·

1h

@simonw nice - thank you! (and crossed my suggestion that you clarify!)

·

1h

@simonw “And therein lies my disagreement. I'm not interested in whether or not LLMs are the "road to AGI". I continue to care only about whether they have useful applications today, once you've understood their limitations…” coupled with utility arguments might lead some people astray

·

1h

@sharongoldman this is a primo example of what I am concerned with: zero diligence reposts of famous people, with zero accountability checks.

happens daily. fairly often even in more famous places.

·

1h

@simonw to each his own.

but fact is that the 500B industry is largely predicated on promises from people like Sam that AGI is close.

and a lot of government policy is based thereon.

i have a right to be interested, and prefer that you not imply that i suggested there was zero

·

1h

@Enscion25 its the architecture. kids learn much better models of the world with much less data.

·

1h

One of the biggest forms of cope among those who overhype current AI - from @sama last week to endless memes in the e/acc community - is to pretend that @garymarcus said that "AI is useless", when he never did.

These endless strawmen are pathetic for two different reasons.