The Oath
Toggle main menu
Gary Marcus
Rank 14 of 47
|
Score 104
+3
Gary Marcus
@GaryMarcus
·
1h
@S_OhEigeartaigh @sama They already twiced switch from the Legg-Goertzel-Marcus “flexibility and resourcefulness and generality of humans” type definition to two economic definitions:
https://garymarcus.substack.com/…
-1
Gary Marcus
@GaryMarcus
·
2h
@iruletheworldmo sorry guy, you were full of it once again,
+5
Gary Marcus
@GaryMarcus
·
2h
BREAKING NEWS: AGI is NOT imminent.
@sama, who weeks ago wrote that “ We are now confident we know how to build AGI as we have traditionally understood it” now confirms that OpenAI has NOT in fact built AGI– confirming what I had been saying all along, and undermining what many…
+2
Gary Marcus
@GaryMarcus
·
4h
@hamilton_mcmc They actually had revenue (not much) and at least one custome. So your first sentence was made up.
In an earlier essay I too leaned towards we work. The o3 presentation looks misleading on multiple counts, as discussed both here and on my substack, and this has lead me to…
-1
Gary Marcus
@GaryMarcus
·
5h
@hamilton_mcmc You just made that up. Check your facts.
--
Reposted by Gary Marcus
Ussal Sahbaz
@ussalEN
·
5h
Book recommendation: Kluge: The Haphazard Construction of the Human Mind by @GaryMarcus -
Professor Marcus is an academic neuroscientist and entrepreneur who sold his previous company to Uber.
After hearing him speak at DLD last year, I read his book and subscribed to his…
+3
Gary Marcus
@GaryMarcus
·
6h
@typingloudly It still makes eg some continuity errors that are not physically possible and I suspect that the conclusions of the paper still apply. It’s simulating what things look like, not what they are.
I look forward to trying it (not widely available last I looked).
+3
Gary Marcus
@GaryMarcus
·
6h
@soumilrathi I look forward to trying it out when that is possible and enthusiastically posted to that effect when the previewed was shared.
It’s a neurosymbolic hybrid btw, afaik.
--
Gary Marcus
@GaryMarcus
·
6h
@EDIMOSIP Real?
+3
Gary Marcus
@GaryMarcus
·
6h
@anupk24 God forbid you allow me an advance look at anything. that is part of why I don’t trust OpenAI.
Another part of why I don’t trust OpnAI is that there has been zero transparency about training data post the days of GPT-3. I have no idea what you trained on, now idea what related…
Load More