Please stop anthropomorphizing AI company executives.
-
Please stop anthropomorphizing AI company executives.
They're not sentient and they can't actually think. They just predict token sequences using statistical patterns scraped from LinkedIn, X and the Financial Times. The apparent "understanding" is simply a parlor trick: They produce text strings that *sound* plausible, but they don't actually have any understanding of the world.
I'll grant that they became a lot more technically impressive when Reinforcement Learning from Investor Sentiment pipelines were introduced, but fundamentally they're still just stochastic bullshit generators. Giving them access to the internet will end up being the most expensive mistake in the history of technology.
LOL
-
Please stop anthropomorphizing AI company executives.
They're not sentient and they can't actually think. They just predict token sequences using statistical patterns scraped from LinkedIn, X and the Financial Times. The apparent "understanding" is simply a parlor trick: They produce text strings that *sound* plausible, but they don't actually have any understanding of the world.
I'll grant that they became a lot more technically impressive when Reinforcement Learning from Investor Sentiment pipelines were introduced, but fundamentally they're still just stochastic bullshit generators. Giving them access to the internet will end up being the most expensive mistake in the history of technology.
@datarama no lies detected
-
Please stop anthropomorphizing AI company executives.
They're not sentient and they can't actually think. They just predict token sequences using statistical patterns scraped from LinkedIn, X and the Financial Times. The apparent "understanding" is simply a parlor trick: They produce text strings that *sound* plausible, but they don't actually have any understanding of the world.
I'll grant that they became a lot more technically impressive when Reinforcement Learning from Investor Sentiment pipelines were introduced, but fundamentally they're still just stochastic bullshit generators. Giving them access to the internet will end up being the most expensive mistake in the history of technology.
-
Please stop anthropomorphizing AI company executives.
They're not sentient and they can't actually think. They just predict token sequences using statistical patterns scraped from LinkedIn, X and the Financial Times. The apparent "understanding" is simply a parlor trick: They produce text strings that *sound* plausible, but they don't actually have any understanding of the world.
I'll grant that they became a lot more technically impressive when Reinforcement Learning from Investor Sentiment pipelines were introduced, but fundamentally they're still just stochastic bullshit generators. Giving them access to the internet will end up being the most expensive mistake in the history of technology.
I assume you're familiar with the "Larry Ellison / lawnmower" idea?
-
I assume you're familiar with the "Larry Ellison / lawnmower" idea?
If not, it starts:
> Do not fall into the trap of anthropomorphizing Larry Ellison. You need to think of Larry Ellison the way you think of a lawnmower. ...Full text here:
https://simonwillison.net/2024/Sep/17/bryan-cantrill/There's a full conference talk on it, but I can't find it now.
-
-
Please stop anthropomorphizing AI company executives.
They're not sentient and they can't actually think. They just predict token sequences using statistical patterns scraped from LinkedIn, X and the Financial Times. The apparent "understanding" is simply a parlor trick: They produce text strings that *sound* plausible, but they don't actually have any understanding of the world.
I'll grant that they became a lot more technically impressive when Reinforcement Learning from Investor Sentiment pipelines were introduced, but fundamentally they're still just stochastic bullshit generators. Giving them access to the internet will end up being the most expensive mistake in the history of technology.
That might have been true six months ago or even three months ago, but you're just showing that you're not truly interacting with AI CEOs *today* with an open mind.
-
Please stop anthropomorphizing AI company executives.
They're not sentient and they can't actually think. They just predict token sequences using statistical patterns scraped from LinkedIn, X and the Financial Times. The apparent "understanding" is simply a parlor trick: They produce text strings that *sound* plausible, but they don't actually have any understanding of the world.
I'll grant that they became a lot more technically impressive when Reinforcement Learning from Investor Sentiment pipelines were introduced, but fundamentally they're still just stochastic bullshit generators. Giving them access to the internet will end up being the most expensive mistake in the history of technology.
@datarama And now I’m hearing @bcantrill reminding us not to anthropomorphize Larry Ellison back in 2010. A true visionary….
-
Please stop anthropomorphizing AI company executives.
They're not sentient and they can't actually think. They just predict token sequences using statistical patterns scraped from LinkedIn, X and the Financial Times. The apparent "understanding" is simply a parlor trick: They produce text strings that *sound* plausible, but they don't actually have any understanding of the world.
I'll grant that they became a lot more technically impressive when Reinforcement Learning from Investor Sentiment pipelines were introduced, but fundamentally they're still just stochastic bullshit generators. Giving them access to the internet will end up being the most expensive mistake in the history of technology.
@datarama So true
-
@mjdxp ...but even they wouldn't go home and talk to an AI company executive.
-
@rugk that is what it refers to
@valpackett @rugk Yep!
-
Please stop anthropomorphizing AI company executives.
They're not sentient and they can't actually think. They just predict token sequences using statistical patterns scraped from LinkedIn, X and the Financial Times. The apparent "understanding" is simply a parlor trick: They produce text strings that *sound* plausible, but they don't actually have any understanding of the world.
I'll grant that they became a lot more technically impressive when Reinforcement Learning from Investor Sentiment pipelines were introduced, but fundamentally they're still just stochastic bullshit generators. Giving them access to the internet will end up being the most expensive mistake in the history of technology.
-
@aho Liability Laundering Marketers?
-
Please stop anthropomorphizing AI company executives.
They're not sentient and they can't actually think. They just predict token sequences using statistical patterns scraped from LinkedIn, X and the Financial Times. The apparent "understanding" is simply a parlor trick: They produce text strings that *sound* plausible, but they don't actually have any understanding of the world.
I'll grant that they became a lot more technically impressive when Reinforcement Learning from Investor Sentiment pipelines were introduced, but fundamentally they're still just stochastic bullshit generators. Giving them access to the internet will end up being the most expensive mistake in the history of technology.
I swear, I'll scream next time some headline mentions what Sam Altman "thinks" or what Dario Amodei "believes".
-
I swear, I'll scream next time some headline mentions what Sam Altman "thinks" or what Dario Amodei "believes".
@datarama I hope you're ready to scream
-
I swear, I'll scream next time some headline mentions what Sam Altman "thinks" or what Dario Amodei "believes".
-
@datarama I hope you're ready to scream
@rune I've been fighting to hold in the screams for the last 4 years at this point
-
@rune I've been fighting to hold in the screams for the last 4 years at this point
@datarama same. I have a void but I don't think she'd like if I screamed into her fur
-
@datarama same. I have a void but I don't think she'd like if I screamed into her fur
@rune My reptilian roomie also doesn't like loud humans.
-
Please stop anthropomorphizing AI company executives.
They're not sentient and they can't actually think. They just predict token sequences using statistical patterns scraped from LinkedIn, X and the Financial Times. The apparent "understanding" is simply a parlor trick: They produce text strings that *sound* plausible, but they don't actually have any understanding of the world.
I'll grant that they became a lot more technically impressive when Reinforcement Learning from Investor Sentiment pipelines were introduced, but fundamentally they're still just stochastic bullshit generators. Giving them access to the internet will end up being the most expensive mistake in the history of technology.
@datarama They really do behave this way! They have been like this for much longer than LLMs have been popular too. Too much exposure to marketing bullshit & buzzwords must have equivalent effects to LLM output exposure & we didn't realize early enough. They gave up their ability to think before it was cool.
