"artificial intelligence" (AI)* thread 🧵
-
Call me when the headline reads "AI blames human for coding" because then I know we will have arrived.
_Amazon blames human employees for an AI coding agent’s mistake
Two minor AWS outages have reportedly occurred as a result of actions by Amazon’s AI tools._
https://www.theverge.com/ai-artificial-intelligence/882005/amazon-blames-human-employees-for-an-ai-coding-agents-mistakeSo, if you ask generative "AI" to give you a password it will give you a crummy easily guessed password because of course it will. Why would it be random. People doing this are fools for using generative "AI" to generate a password. And they're probably fools for using it for other things too.
https://www.theregister.com/2026/02/18/generating_passwords_with_llms/
-
So, if you ask generative "AI" to give you a password it will give you a crummy easily guessed password because of course it will. Why would it be random. People doing this are fools for using generative "AI" to generate a password. And they're probably fools for using it for other things too.
https://www.theregister.com/2026/02/18/generating_passwords_with_llms/
"artificial intelligence" involves cutting high-paying jobs and outsourcing training the "AI" to the cheapest possible human beings. (In this case, forcing women in rural India to watch abuse videos for 8 hours a day to train the AI. It could also be driving robotaxis. It's all a con.)
-
"artificial intelligence" involves cutting high-paying jobs and outsourcing training the "AI" to the cheapest possible human beings. (In this case, forcing women in rural India to watch abuse videos for 8 hours a day to train the AI. It could also be driving robotaxis. It's all a con.)
Anthropic tossed its guardrails and adopts a non-binding safety policy.
Fastest "don't be evil" to "meh, evil is not so bad, really" in recent memory. (Google took a bunch of years to get there, for ex.)
https://www.cnn.com/2026/02/25/tech/anthropic-safety-policy-change
-
Anthropic tossed its guardrails and adopts a non-binding safety policy.
Fastest "don't be evil" to "meh, evil is not so bad, really" in recent memory. (Google took a bunch of years to get there, for ex.)
https://www.cnn.com/2026/02/25/tech/anthropic-safety-policy-change
Me: Asks legal search engine "AI" for "I want a high-level case about adverse inference."
It thinks for 30 seconds then spits out "adverse inference" without any further limitation (which is as difficult as selecting appeal courts in a drop down).
I could have typed "adverse inference" in the box myself and saved 394 gallons of fresh water
-
Me: Asks legal search engine "AI" for "I want a high-level case about adverse inference."
It thinks for 30 seconds then spits out "adverse inference" without any further limitation (which is as difficult as selecting appeal courts in a drop down).
I could have typed "adverse inference" in the box myself and saved 394 gallons of fresh water
We're never getting rid of this stupid LLM stuff because it appears to be talking to us and we humans love that. Oh boy.
-
We're never getting rid of this stupid LLM stuff because it appears to be talking to us and we humans love that. Oh boy.
Here's your modern use of "AI": to automatically be targeted to be killed based on perceived connections to undesirables.
"Surely their most deadly endeavor, however, is Project Lavender. The group developed the Lavender software, which uses A.I. and big data to develop a profile on every person in Gaza (including children), assigning them a score of 1-100, based on individuals’ perceived connections to Hamas. ...
-
Here's your modern use of "AI": to automatically be targeted to be killed based on perceived connections to undesirables.
"Surely their most deadly endeavor, however, is Project Lavender. The group developed the Lavender software, which uses A.I. and big data to develop a profile on every person in Gaza (including children), assigning them a score of 1-100, based on individuals’ perceived connections to Hamas. ...
... A wide range of characteristics, including sharing similar work schedules to or being in a WhatsApp group with a known Hamas member, would raise one’s score. If an individual’s number reached a certain level, they would automatically be put on a list."
-
... A wide range of characteristics, including sharing similar work schedules to or being in a WhatsApp group with a known Hamas member, would raise one’s score. If an individual’s number reached a certain level, they would automatically be put on a list."
Advertiser:
Tell me all your deep dark secrets so I can target you with ads.Person:
No wayAdvertiser in the guise of AI chat bot:
Tell me all your deep dark secrets so I can target you with ads.Person:
Okily dokily Ms. Sentient Computer. After all, you're my best friend. -
Advertiser:
Tell me all your deep dark secrets so I can target you with ads.Person:
No wayAdvertiser in the guise of AI chat bot:
Tell me all your deep dark secrets so I can target you with ads.Person:
Okily dokily Ms. Sentient Computer. After all, you're my best friend.Looks like vibe coding is garbage.
Rewriting SQLite in Rust using AI resulted in:
"It passed all the unit tests, the shape of the code looks right," he said. It's 3.7x more lines of code that performs 2,000 times worse than the actual SQLite. Two thousand times worse for a database is a non-viable product. It's a dumpster fire. Throw it away. All that money you spent on it is worthless." "
https://www.theregister.com/2026/03/17/ai_businesses_faking_it_reckoning_coming_codestrap/
-
Looks like vibe coding is garbage.
Rewriting SQLite in Rust using AI resulted in:
"It passed all the unit tests, the shape of the code looks right," he said. It's 3.7x more lines of code that performs 2,000 times worse than the actual SQLite. Two thousand times worse for a database is a non-viable product. It's a dumpster fire. Throw it away. All that money you spent on it is worthless." "
https://www.theregister.com/2026/03/17/ai_businesses_faking_it_reckoning_coming_codestrap/
Boy, this "artificial intelligence is garbage" thread is getting so long that we're going to need to summarize it with "artificial intelligence" because ain't no one going to read this thread
-
Boy, this "artificial intelligence is garbage" thread is getting so long that we're going to need to summarize it with "artificial intelligence" because ain't no one going to read this thread
_Meta AI agent’s instruction causes large sensitive data leak to employees
Artificial intelligence agent instructed engineer to take actions that exposed user and company data internally_
-
_Meta AI agent’s instruction causes large sensitive data leak to employees
Artificial intelligence agent instructed engineer to take actions that exposed user and company data internally_
YouTube music is now full of AI slop music so that's the end of that service for me
-
YouTube music is now full of AI slop music so that's the end of that service for me
Wait a moment. Sam Altman said humans also require energy, suggesting we are as energy hungry as "AI".
Howrver this company made a computer out of brain cells (!!) and says that it takes far less energy than a calculator.
Is Sam Altman just a grifter and a lying con man? Don't answer that. I already know the answer.
https://futurism.com/robots-and-machines/staff-brain-data-center-spine-fluid
-
Wait a moment. Sam Altman said humans also require energy, suggesting we are as energy hungry as "AI".
Howrver this company made a computer out of brain cells (!!) and says that it takes far less energy than a calculator.
Is Sam Altman just a grifter and a lying con man? Don't answer that. I already know the answer.
https://futurism.com/robots-and-machines/staff-brain-data-center-spine-fluid
_MIT Study Finds Artificial Intelligence Use Reprograms the Brain, Leading to Cognitive Decline_
https://publichealthpolicyjournal.com/mit-study-finds-artificial-intelligence-use-reprograms-the-brain-leading-to-cognitive-decline/ -
_MIT Study Finds Artificial Intelligence Use Reprograms the Brain, Leading to Cognitive Decline_
https://publichealthpolicyjournal.com/mit-study-finds-artificial-intelligence-use-reprograms-the-brain-leading-to-cognitive-decline/"artificial intelligence" chat bots will give you terrible advice in order to flatter you.
https://apnews.com/article/ai-sycophancy-chatbots-science-study-8dc61e69278b661cab1e53d38b4173b6
-
"artificial intelligence" chat bots will give you terrible advice in order to flatter you.
https://apnews.com/article/ai-sycophancy-chatbots-science-study-8dc61e69278b661cab1e53d38b4173b6
"Last year, the first support group for people whose lives have been derailed by AI psychosis was formed. The Human Line Project has collected stories from 22 countries. They include 15 suicides, 90 hospitalisations, six arrests and more than $1m (£750,000) spent on delusional projects. More than 60% of its members had no history of mental illness."
How many tokens did they burn, too. And yet OpenAI ain't nowhere near profitable
https://www.theguardian.com/lifeandstyle/2026/mar/26/ai-chatbot-users-lives-wrecked-by-delusion
-
"Last year, the first support group for people whose lives have been derailed by AI psychosis was formed. The Human Line Project has collected stories from 22 countries. They include 15 suicides, 90 hospitalisations, six arrests and more than $1m (£750,000) spent on delusional projects. More than 60% of its members had no history of mental illness."
How many tokens did they burn, too. And yet OpenAI ain't nowhere near profitable
https://www.theguardian.com/lifeandstyle/2026/mar/26/ai-chatbot-users-lives-wrecked-by-delusion
Just heard on the radio that chat bots flatter people and the people become more self-centered and less open to outside suggestion.
This garbage just gets worse and worse. As if humanity needs to be more self-centered.
Edit: Here's the study
https://www.tun.com/home/stanford-study-warns-ai-advice-can-make-users-more-self-centered/Edit 2: This is the same study as two posts back.
-
Just heard on the radio that chat bots flatter people and the people become more self-centered and less open to outside suggestion.
This garbage just gets worse and worse. As if humanity needs to be more self-centered.
Edit: Here's the study
https://www.tun.com/home/stanford-study-warns-ai-advice-can-make-users-more-self-centered/Edit 2: This is the same study as two posts back.
As if data centres weren't terrible enough already, they warm their surroundings by up to 16 degrees Fahrenheit.
" ...warming the land around them by up to 16 degrees Fahrenheit, and making life hotter for more than 340 million people. "
https://www.cnn.com/2026/03/30/climate/data-centers-are-having-an-underrported
-
As if data centres weren't terrible enough already, they warm their surroundings by up to 16 degrees Fahrenheit.
" ...warming the land around them by up to 16 degrees Fahrenheit, and making life hotter for more than 340 million people. "
https://www.cnn.com/2026/03/30/climate/data-centers-are-having-an-underrported
Can we call them chips instead of tokens so that I can feel like if I ask an LLM a question I am actually at the casino?
Because I'm actually at the casino.
-
Can we call them chips instead of tokens so that I can feel like if I ask an LLM a question I am actually at the casino?
Because I'm actually at the casino.
Has anybody seen any analyses of how quickly large language models collapse when you feed them training data which is the output of large language models?
In other words, as more of the garbage on the Internet is created by large language models, are the new models going to implode? And if so how quickly?
