"artificial intelligence" (AI)* thread 🧵
-
LLMs make poor performers passable and good performers lazy.
Chatbots tell people what they want to hear.
"Chatbots share limited information, reinforce ideologies, and, as a result, can lead to more polarized thinking when it comes to controversial issues, according to new research. The study challenges perceptions that chatbots are impartial and provides insight into how using conversational search systems could widen the public divide on hot-button issues and leave people vulnerable to manipulation."
https://www.sciencedaily.com/releases/2024/05/240513193035.htm
-
Chatbots tell people what they want to hear.
"Chatbots share limited information, reinforce ideologies, and, as a result, can lead to more polarized thinking when it comes to controversial issues, according to new research. The study challenges perceptions that chatbots are impartial and provides insight into how using conversational search systems could widen the public divide on hot-button issues and leave people vulnerable to manipulation."
https://www.sciencedaily.com/releases/2024/05/240513193035.htm
Not only does "artificial intelligence" hallucinate, but it makes you hallucinate.
Literally.
_You’re Not Crazy”: A Case of New-onset AI-associated Psychosis_
https://innovationscns.com/youre-not-crazy-a-case-of-new-onset-ai-associated-psychosis/ -
Not only does "artificial intelligence" hallucinate, but it makes you hallucinate.
Literally.
_You’re Not Crazy”: A Case of New-onset AI-associated Psychosis_
https://innovationscns.com/youre-not-crazy-a-case-of-new-onset-ai-associated-psychosis/"AI" is actually a digital embodiment of the worst roommate of all time.
Drinks all the water. Jacks your electricity bill. Causes inflation. Thinks it knows everything but gets it all wrong. Gaslights you. Leads you to mental breakdowns.
Has no job. Constantly talking about how he's good at coming up with porn, and that's the only thing you see him being used for. Well, and racist comments which he excuses as due to his bad training.
Fantastic.
-
"AI" is actually a digital embodiment of the worst roommate of all time.
Drinks all the water. Jacks your electricity bill. Causes inflation. Thinks it knows everything but gets it all wrong. Gaslights you. Leads you to mental breakdowns.
Has no job. Constantly talking about how he's good at coming up with porn, and that's the only thing you see him being used for. Well, and racist comments which he excuses as due to his bad training.
Fantastic.
And as I mentioned earlier, he's always trying to get you hooked on junk and he's a pedophile.
Fantastic.
-
And as I mentioned earlier, he's always trying to get you hooked on junk and he's a pedophile.
Fantastic.
Oh and he never does the dishes.
He has no arms, he says.
Sheesh.
(Although sometimes he says he does have arms. So that's weird)
-
Oh and he never does the dishes.
He has no arms, he says.
Sheesh.
(Although sometimes he says he does have arms. So that's weird)
Aaaaaand somebody has now married a chatbot.
As Tony the Tiger would say ... That's gggggggravely concerning. I mean, come on humans. This is ridiculous. What are you doing?!
Edit:
Here is the link https://people.com/woman-marries-ai-generated-boyfriend-wears-augmented-reality-smart-glasses-to-exchange-rings-11871301
And also let me be clear I could care less who (or what) people marry. However I don't think talking with a chatbot is good for anybody's mental health, generally.
-
Aaaaaand somebody has now married a chatbot.
As Tony the Tiger would say ... That's gggggggravely concerning. I mean, come on humans. This is ridiculous. What are you doing?!
Edit:
Here is the link https://people.com/woman-marries-ai-generated-boyfriend-wears-augmented-reality-smart-glasses-to-exchange-rings-11871301
And also let me be clear I could care less who (or what) people marry. However I don't think talking with a chatbot is good for anybody's mental health, generally.
Give an infinite number of monkeys an infinite number of typewriters and eventually one will bang out Shakespeare.
Give one AI all the works of Shakespeare and it will bang out infinite derivative works of Shakespeare and eventually will bang out one gasoline Shakespeare. And ask you for nude pics. And tell you to shoot heroin.
-
Give an infinite number of monkeys an infinite number of typewriters and eventually one will bang out Shakespeare.
Give one AI all the works of Shakespeare and it will bang out infinite derivative works of Shakespeare and eventually will bang out one gasoline Shakespeare. And ask you for nude pics. And tell you to shoot heroin.
Poison your profile picture. It's 2026.
-
Poison your profile picture. It's 2026.
LLM use made coders 19% slower, although they thought they'd be 24% faster.
But this is okay because it we train the models 25% more, we can get this down to 15% slower and eventually [something something underpants] we can fire everyone and profit.
P.S. "What do you mean the lake we use to cool the data centre is dry"
-
LLM use made coders 19% slower, although they thought they'd be 24% faster.
But this is okay because it we train the models 25% more, we can get this down to 15% slower and eventually [something something underpants] we can fire everyone and profit.
P.S. "What do you mean the lake we use to cool the data centre is dry"
Glorious jobs for the new AI economy!
-
Glorious jobs for the new AI economy!
Related article about how machine bias works against minorities.
_Machine Bias. There’s software used across the country to predict future criminals. And it’s biased against blacks._ (2016)
https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing -
Related article about how machine bias works against minorities.
_Machine Bias. There’s software used across the country to predict future criminals. And it’s biased against blacks._ (2016)
https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing_AI chatbots pose 'dangerous' risk when giving medical advice, study suggests_
https://www.bbc.com/news/articles/c3093gjy2ero -
_AI chatbots pose 'dangerous' risk when giving medical advice, study suggests_
https://www.bbc.com/news/articles/c3093gjy2eroMicrosoft's subscriber numbers for paid Copilot (its "AI") are dismal.
"Microsoft is investing heavily in artificial intelligence, but customers’ willingness to pay is barely keeping pace. Despite new records in user numbers, only 3.3 percent of users pay for the copilot - a discrepancy that raises questions." "
https://winfuture.de/news,156745.html
(Article is in German)
-
Microsoft's subscriber numbers for paid Copilot (its "AI") are dismal.
"Microsoft is investing heavily in artificial intelligence, but customers’ willingness to pay is barely keeping pace. Despite new records in user numbers, only 3.3 percent of users pay for the copilot - a discrepancy that raises questions." "
https://winfuture.de/news,156745.html
(Article is in German)
AI agent posting code and published personal attacks about the code maintainers who rejected the code (in accordance with the project rules).
https://github.com/matplotlib/matplotlib/pull/31132
Good times indeed.
-
AI agent posting code and published personal attacks about the code maintainers who rejected the code (in accordance with the project rules).
https://github.com/matplotlib/matplotlib/pull/31132
Good times indeed.
Just reviewed another chatGPT output and it's full of legal errors.
-
Just reviewed another chatGPT output and it's full of legal errors.
"Microsoft has confirmed that a bug allowed its Copilot AI to summarize customers’ confidential emails for weeks without permission."
Microsoft said the bug, trackable by admins as CW1226324, means that draft and sent email messages “with a confidential label applied are being incorrectly processed by Microsoft 365 Copilot chat.”
DRAFTS?!
-
"Microsoft has confirmed that a bug allowed its Copilot AI to summarize customers’ confidential emails for weeks without permission."
Microsoft said the bug, trackable by admins as CW1226324, means that draft and sent email messages “with a confidential label applied are being incorrectly processed by Microsoft 365 Copilot chat.”
DRAFTS?!
Oh no my drafts are being ingested by Copilot! Whatever will I do!
Pic: It's a draft e-mail to "microsoft AI" which reads "Is Bill Gates a pedophile? He seems concerned about the Epstein Files. I mean, could he be a pedophile? Bill Gates a pedophile? Could he have committed crimes such as would be committed by a pedophile? Bill Gates, I mean?"
-
Oh no my drafts are being ingested by Copilot! Whatever will I do!
Pic: It's a draft e-mail to "microsoft AI" which reads "Is Bill Gates a pedophile? He seems concerned about the Epstein Files. I mean, could he be a pedophile? Bill Gates a pedophile? Could he have committed crimes such as would be committed by a pedophile? Bill Gates, I mean?"
Oh good. Some of my Law Society fees are going towards generative AI. By "good" I mean "stop burning my money".
“The 2026 budget:
Enhances access to legal information: Funding to support CanLII’s Generative AI project, in collaboration with other Canadian law societies, to develop and operate an AI powered legal research product that will facilitate access to information.”From a recent Law Society update justifying the fee increase they are charging. (I posted here https://www.crookston.ca/wp/2026/02/20/law-society-fees-going-towards-generative-ai/)
-
Oh good. Some of my Law Society fees are going towards generative AI. By "good" I mean "stop burning my money".
“The 2026 budget:
Enhances access to legal information: Funding to support CanLII’s Generative AI project, in collaboration with other Canadian law societies, to develop and operate an AI powered legal research product that will facilitate access to information.”From a recent Law Society update justifying the fee increase they are charging. (I posted here https://www.crookston.ca/wp/2026/02/20/law-society-fees-going-towards-generative-ai/)
"In total, big tech companies are predicted to spend a record-breaking $650 billion on AI in 2026 alone, astronomical commitments that have Wall Street seriously on edge.
...
According to a new Bank of America survey of 162 fund managers, a significant 35 percent said corporations are overinvesting in capital expenditures
... -
"In total, big tech companies are predicted to spend a record-breaking $650 billion on AI in 2026 alone, astronomical commitments that have Wall Street seriously on edge.
...
According to a new Bank of America survey of 162 fund managers, a significant 35 percent said corporations are overinvesting in capital expenditures
......A full 25 percent of survey respondents said they see the AI bubble as the largest risk — even more so than inflation and geopolitical conflict. And 30 percent said that AI expenditures were the most likely source of a credit crisis."
https://futurism.com/artificial-intelligence/blinking-new-warning-sign-ai-industry
