Some thoughts: The author claims that 'TESCREALism' is like a religion, but also acknowledges that AI is drastically changing the world, and seems to want to slow it down. This is a common...
Some thoughts:
The author claims that 'TESCREALism' is like a religion, but also acknowledges that AI is drastically changing the world, and seems to want to slow it down. This is a common propaganda technique, making your enemy seem both weak and strong at the same time. In this case dismissing LLMs as 'stochastic parrots' but also accepting that OpenAI will centralize power by replacing doctors.
It seems they aren't making an important distinction between the TESCREALists who want to slow down AI development, and those who want to accelerate it, lumping them together into one group. This is very strange since LWers seem predominantly against the development of AI.
I think you're missing some nuance here – Gebru claims that LLMs are "stochastic parrots" (which she was fired from Google for). She does not claim that they will replace doctors because they are...
The author claims that 'TESCREALism' is like a religion, but also acknowledges that AI is drastically changing the world, and seems to want to slow it down. This is a common propaganda technique, making your enemy seem both weak and strong at the same time. In this case dismissing LLMs as 'stochastic parrots' but also accepting that OpenAI will centralize power by replacing doctors.
I think you're missing some nuance here – Gebru claims that LLMs are "stochastic parrots" (which she was fired from Google for). She does not claim that they will replace doctors because they are better but rather because OpenAI will claim that their model is "clever" (which is impossible) and therefore should replace them.
It seems they aren't making an important distinction between the TESCREALists who want to slow down AI development, and those who want to accelerate it, lumping them together into one group. This is very strange since LWers seem predominantly against the development of AI.
I think you're confusing two notions of AI
(a) AI in the here and now (e.g. neural networks) which have real-world immediate harms
(b) the possible invention of clever AI which might want to kill us and might be able to do so at some distant and undetermined point in the future
The TESCREAL people care a lot about (b) which they think is the really worrying thing. Gebru cares a lot more about (a); this XKCD comic explains the distinction well.
I think this is a somewhat silly take. If LLMs aren't really intelligent enough to replace any intellectual labor, then the amount of power they can centralize is very limited. While there's a lot...
I think you're missing some nuance here – Gebru claims that LLMs are "stochastic parrots" (which she was fired from Google for). She does not claim that they will replace doctors because they are better but rather because OpenAI will claim that their model is "clever" (which is impossible) and therefore should replace them.
I think this is a somewhat silly take. If LLMs aren't really intelligent enough to replace any intellectual labor, then the amount of power they can centralize is very limited. While there's a lot of hype---just like there was for expert systems/GOFAI in the past---if you truly believe that it's just hype, and the AI is a stochastic parrot, you shouldn't have anything to worry about.
The TESCREAL people care a lot about (b) which they think is the really worrying thing. Gebru cares a lot more about (a); this XKCD comic explains the distinction well.
True, but I still think that it's doing her argument a disservice to lump the groups that believe (b) and those that don't, because they have different beliefs and different actions. It's missing a lot of nuance, especially when those who believe in (b) could be allies to her cause.
OpenAI will claim that their model is "clever" (which is impossible)
Impossible? We don't actually really know how intelligence or cleverness works, so it would be hard to disprove any claims of such.
At risk of poisoning the thread with irony, here's an AI-generated summary expandable by timestamp: https://www.summarize.tech/www.youtube.com/watch?v=P7XT4TWLzJw
At risk of poisoning the thread with irony, here's an AI-generated summary expandable by timestamp:
Some thoughts:
I think you're missing some nuance here – Gebru claims that LLMs are "stochastic parrots" (which she was fired from Google for). She does not claim that they will replace doctors because they are better but rather because OpenAI will claim that their model is "clever" (which is impossible) and therefore should replace them.
I think you're confusing two notions of AI
The TESCREAL people care a lot about (b) which they think is the really worrying thing. Gebru cares a lot more about (a); this XKCD comic explains the distinction well.
I think this is a somewhat silly take. If LLMs aren't really intelligent enough to replace any intellectual labor, then the amount of power they can centralize is very limited. While there's a lot of hype---just like there was for expert systems/GOFAI in the past---if you truly believe that it's just hype, and the AI is a stochastic parrot, you shouldn't have anything to worry about.
True, but I still think that it's doing her argument a disservice to lump the groups that believe (b) and those that don't, because they have different beliefs and different actions. It's missing a lot of nuance, especially when those who believe in (b) could be allies to her cause.
Impossible? We don't actually really know how intelligence or cleverness works, so it would be hard to disprove any claims of such.
Does anyone know of a transcript, or a way to get one? I'm interested to hear what she says, but I don't have a 48 minute attention span.
At risk of poisoning the thread with irony, here's an AI-generated summary expandable by timestamp:
https://www.summarize.tech/www.youtube.com/watch?v=P7XT4TWLzJw