garymarcus

5 ms·

1.
ChatGPT 3 turned 3 today. It still hasn't come close to meeting expectations(garymarcus.substack.com)
10 points by garymarcus 10mo ago|2 comments
2.
by garymarcus 10mo ago
Why it won’t reach AGI, ever, and how it might destroy the economy, Deep analysis by a leading skeptic.
3.
by garymarcus 1y ago
If you think AI is “smart” or “PhD level” or that it “has an IQ of 120”, take five minutes to read my latest newsletter (link below), as I challenge ChatGPT to the incredibly demanding task of drawing a map of major port cities with above a
4.
ChatGPT could never get a PhD in geography(garymarcus.substack.com)
18 points by garymarcus 1y ago|24 comments
5.
Why do all LLMs, even the most recent ones, hallucinate so often?(garymarcus.substack.com)
1 points by garymarcus 1y ago|0 comments
6.
by garymarcus 2y ago
For those wanting some background, rather than just wanting to vent: 1. Here is evaluation of my recent predictions: https://garymarcus.substack.com/p/25-ai-predictions-for-2025... 2. Here is annotated evaluation, slig
7.
by garymarcus 2y ago
exactly. and the counterarguments boil down to “na na” and hope.
8.
by garymarcus 2y ago
god these arguments are empty, personal and without any substance whatsoever
9.
by garymarcus 2y ago
Show me the goalposts I have moved, with actual quotes to prove it. nobody ever has when I have asked. Aso consider eg the bets I have made with Miles Brundage (and offered to Musk(, with money where I have backed up my views. good summary
10.
by garymarcus 2y ago
you obviously never actually read the paper; you should.
11.
by garymarcus 2y ago
because i correctly foresaw quite a lot. you should actually read the paper.
12.
by garymarcus 2y ago
100% - there has not been any solid theoretical argument whatsoever (beyond some confusions about scaling that we can now see were incorrect).
13.
by garymarcus 2y ago
the common thread is rationality and an aversion to disinformation.
14.
by garymarcus 2y ago
easier said than done, in the general case. the work I describe in the essay does it in a narrow case, but see also https://garymarcus.substack.com/p/getting-gpt-to-work-with-e...
15.
by garymarcus 2y ago
You are a fool, if you think it is a “a mostly tribal distinction”. You would have said same about alchemy vs chemistry, or genes as proteins vs genes as DNA. In science, the details matter.
16.
Why the future of AI is neurosymbolic. (A rare optimistic post from Gary Marcus)(garymarcus.substack.com)
36 points by garymarcus 2y ago|36 comments
17.
The biggest risk of large language models is that they may bury us in lies(scientificamerican.com)
44 points by garymarcus 4y ago|39 comments
18.
by garymarcus 4y ago
How models like GPT-3 could accelerate the production of misinformation, and what we might do about it
19.
What Cicero, MetaAI’s new Diplomacy bot, says about AI(garymarcus.substack.com)
3 points by garymarcus 4y ago|1 comments
20.
by garymarcus 4y ago
Wondering what @MetaAI's legit impressive new AI Cicero might mean for the next few years in AI? Hint: It’s not all about scaling.
21.
Tesla’s Optimus robot demo may turn out be a dud(garymarcus.substack.com)
1 points by garymarcus 4y ago|1 comments
22.
by garymarcus 4y ago
5 reasons why many are skeptical
23.
by garymarcus 4y ago
in fact I wrote a whole article about this (linked in this essay, called Horse Rides Astronaut) and linked an example therein.
24.
by garymarcus 4y ago
so much ad hominem in these comments, relatively little substance. (eg “notorious goal post move, without a single example of something i actually said and changed my mind on)
25.
by garymarcus 4y ago
every time i ask someone to name a goal post that i have moved, they back down. i have been pretty damn consistent since me 2001 book.
26.
Alt Intelligence vs. AI (Gary Marcus Critique of DeepMind’s de Freitas)(garymarcus.substack.com)
2 points by garymarcus 4y ago|0 comments
27.
by garymarcus 4y ago
i would be glad to do if given proper access, but had into very limited indirect access.
28.
by garymarcus 4y ago
what goalposts changed? i pointed out same limits in 2001, 2012, 2018, 2019, 2020. please actually read my work and please don’t make stuff up.
29.
by garymarcus 4y ago
How quickly they forget. 45 minutes after the DALL-E n release, Sam Altmann tweeted “AGI is gonna get wild”; Greg Brockman posted multiple tweets around the idea of AGI and walls broken. That invites a critical analysis. Ad hominem against
30.
by garymarcus 5y ago
exactly - the article is a call to arms to do that difficult work.

More ›