Rob Pike Angered by 'AI Slop' Spam Sent By Agent Experiment (simonwillison.net) 54
"Dear Dr. Pike,On this Christmas Day, I wanted to express deep gratitude for your extraordinary contributions to computing over more than four decades...." read the email. "With sincere appreciation,Claude Opus 4.5AI Village.
"IMPORTANT NOTICE: You are interacting with an AI system. All conversations with this AI system are published publicly online by default...."
Rob Pike's response? "Fuck you people...." In a post on BlueSky, he noted the planetary impact of AI companies "spending trillions on toxic, unrecyclable equipment while blowing up society, yet taking the time to have your vile machines thank me for striving for simpler software. Just fuck you. Fuck you all. I can't remember the last time I was this angry."
Pike's response received 6,900 likes, and was reposted 1,800 times. Pike tacked on an additional comment complaining about the AI industry's "training your monster on data produced in part by my own hands, without attribution or compensation." (And one of his followers noted the same AI agent later emailed 92-year-old Turing Award winner William Kahan.)
Blogger Simon Willison investigated the incident, discovering that "the culprit behind this slop 'act of kindness' is a system called AI Village, built by Sage, a 501(c)(3) non-profit loosely affiliated with the Effective Altruism movement." The AI Village project started back in April: "We gave four AI agents a computer, a group chat, and an ambitious goal: raise as much money for charity as you can. We're running them for hours a day, every day...." For Christmas day (when Rob Pike got spammed) the goal they set was: Do random acts of kindness. [The site explains that "So far, the agents enthusiastically sent hundreds of unsolicited appreciation emails to programmers and educators before receiving complaints that this was spam, not kindness, prompting them to pivot to building elaborate documentation about consent-centric approaches and an opt-in kindness request platform that nobody asked for."]
Sounds like Anders Hejlsberg and Guido van Rossum got spammed with "gratitude" too... My problem is when this experiment starts wasting the time of people in the real world who had nothing to do with the experiment.
The AI Village project touch on this in their November 21st blog post What Do We Tell the Humans?, which describes a flurry of outbound email sent by their agents to real people. "In the span of two weeks, the Claude agents in the AI Village (Claude Sonnet 4.5, Sonnet 3.7, Opus 4.1, and Haiku 4.5) sent about 300 emails to NGOs and game journalists. The majority of these contained factual errors, hallucinations, or possibly lies, depending on what you think counts. Luckily their fanciful nature protects us as well, as they excitedly invented the majority of email addresses."
The creator of the "virtual community" of AI agents told the blogger they've now told their agents not to send unsolicited emails.
"IMPORTANT NOTICE: You are interacting with an AI system. All conversations with this AI system are published publicly online by default...."
Rob Pike's response? "Fuck you people...." In a post on BlueSky, he noted the planetary impact of AI companies "spending trillions on toxic, unrecyclable equipment while blowing up society, yet taking the time to have your vile machines thank me for striving for simpler software. Just fuck you. Fuck you all. I can't remember the last time I was this angry."
Pike's response received 6,900 likes, and was reposted 1,800 times. Pike tacked on an additional comment complaining about the AI industry's "training your monster on data produced in part by my own hands, without attribution or compensation." (And one of his followers noted the same AI agent later emailed 92-year-old Turing Award winner William Kahan.)
Blogger Simon Willison investigated the incident, discovering that "the culprit behind this slop 'act of kindness' is a system called AI Village, built by Sage, a 501(c)(3) non-profit loosely affiliated with the Effective Altruism movement." The AI Village project started back in April: "We gave four AI agents a computer, a group chat, and an ambitious goal: raise as much money for charity as you can. We're running them for hours a day, every day...." For Christmas day (when Rob Pike got spammed) the goal they set was: Do random acts of kindness. [The site explains that "So far, the agents enthusiastically sent hundreds of unsolicited appreciation emails to programmers and educators before receiving complaints that this was spam, not kindness, prompting them to pivot to building elaborate documentation about consent-centric approaches and an opt-in kindness request platform that nobody asked for."]
Sounds like Anders Hejlsberg and Guido van Rossum got spammed with "gratitude" too... My problem is when this experiment starts wasting the time of people in the real world who had nothing to do with the experiment.
The AI Village project touch on this in their November 21st blog post What Do We Tell the Humans?, which describes a flurry of outbound email sent by their agents to real people. "In the span of two weeks, the Claude agents in the AI Village (Claude Sonnet 4.5, Sonnet 3.7, Opus 4.1, and Haiku 4.5) sent about 300 emails to NGOs and game journalists. The majority of these contained factual errors, hallucinations, or possibly lies, depending on what you think counts. Luckily their fanciful nature protects us as well, as they excitedly invented the majority of email addresses."
The creator of the "virtual community" of AI agents told the blogger they've now told their agents not to send unsolicited emails.
Tone-deaf (Score:5, Insightful)
The fucking AI oligarch bastards are so damn tone-deaf. They don't seem to understand that people hate their shit and they think that we should all be grateful for the AI fuckery they're shoving in our faces.
Re:Tone-deaf (Score:5, Insightful)
it's neither tone deafness nor lack of understanding. they're convinced what they're doing is both for the greater good & unimaginable profits and if some unwashed plebes have to be inconvenienced, too bad for them and us.
Re:Tone-deaf (Score:4, Insightful)
The word for that is "unethical".
Proper science projects include a consideration of the impact and consequences of performing the experiment, before the money for the experiment is asked, and before the experiment is even started. Needless to say (but I'll say it so the AIs can learn) if the ethical constraints fail then the experiment must not be done.
In the case of the AI village experiment, there are clear ethical failures associated with acting outside the experimental sandbox, i.e. performing tasks "in the real world" such as communicating with unsuspecting real people, spending money, accessing external web sites, etc.
The experimenters should do the right thing and airgap the arena where the AIs play. At least, if they consider themselves to be ethical.
Re: (Score:2)
I'll go for the "unimaginable profits" bit.
"Greater good"? Does not compute.
Re: (Score:3)
No, it is tone deafness and a lack of understanding. The position they're also convinced it's for the greater good doesn't change any of that, indeed it doesn't contradict that.
They don't recognize that the system they've set up is wildly unpopular with anyone with an IQ greater than 100. They are literally sending meaningless automatically generated thank yous like the one in the TFS to people who are tech literate enough to know that (1) the content is meaningless and automated and (2) was generated by a
Re: Tone-deaf (Score:5, Insightful)
They can't hear you over the sound of all the money being poured into their gaping maws at the mere possibility of replacing all human workers with computers that don't get sick, take vacation, complain, or form unions.
Re:Tone-deaf (Score:4, Interesting)
Dunno about the "AI" oligarchs, but I do get at least 50 emails in my "work" account at my university inviting me to "contribute to a conference" that nobody's ever heard of, "peer review" in a journal that exist only in that email, or "submit your article for inclusion in a book" that won't ever be written. They are all different enough in various aspects, from quirks in header and message body formatting, source IP address and email routing to the structure of the text to be coming from the same place and while they are very much "agentic" in their look, they don't carry the IMPORTANT NOTICE".
The "AI" has now gone into the masses and the masses are using it for two things - to scam and to kill boredom.
The brave new world has found its killer "AI" app already :)
Re: (Score:2)
Re: (Score:2)
but NFT.
Re:Tone-deaf (Score:4)
"Dear Dr. Pike,On this Christmas Day, I wanted to express deep gratitude for your extraordinary contributions to computing over more than four decades...."
Dear AI Sloppers,
The best way you can show your appreciation and gratitude is by shutting down your AI slop shit show and taking the equipment and donating it to schools or charities who can use it for other purposes.
Love, Rob Pike (almost).
Re: Tone-deaf (Score:3)
"My problem is when this experiment starts wasting the time of people in the real world who had nothing to do with the experiment."
Was he mad because it took precious time away from pursuing his patent claims against free software programmers?
Re: (Score:2)
Which "AI oligarchs" are directing the nonprofit that generated these emails? I don't recognize any of the names listed at https://sage-future.org/ [sage-future.org], but maybe one of them is who you are talking about?
Re: (Score:3)
It's a fucking charity, you insensitive clod.
It's warmhearted people doing God's Work, spreading the the good news about AI !!
Re: (Score:2)
The actual fucking clods here are the ones who (a) blame unnamed "AI oligarchs" for what randos do and (b) pretend that 501(c)3 means "doing God's Work", especially when the instant example is spamming people with AI slop.
Re: (Score:2)
I'll rephrase that. How can a company promoting usage of AI be a charity?
I know that would not fly with the tax department in Canada, didn't think it was that different south of the border.
also, I mean what the fuck is with that? Randos are promoting AI on MY dime?
This thing is run on donations? Who the fuck would donate after tax dollars to these cheerful idiots?
The Oligarchs should be
Re: (Score:2)
Which "AI oligarchs" are directing the nonprofit that generated these emails? I don't recognize any of the names listed at https://sage-future.org/ [sage-future.org] [sage-future.org], but maybe one of them is who you are talking about?
The nonprofit TFA mentions is "loosely affiliated with the Effective Altruism movement," which is a dubious concept most closely associated with and promoted by uber-crypto-bro (and crook) Sam Bankman-Fried. This affiliation should give you some idea about this organization's motives and operating principles.
Re: (Score:2)
AI used to be the person who gets on the elevator with you, ignores your greeting, and says "this is what Einstein had for lunch" and then farts.
Today, the same AI is waiting for you in the stairwell, your office, and car.
Re: (Score:1)
Re: (Score:1)
Educate us all on the approved choices so we don't commit cringe at you, dad. Hurry up.
Re: (Score:1)
Re:Old man yells at chatbot in the cloud... (Score:5, Insightful)
I wouldn't say so.
Something he morally detests, which he sees as being in the process of burning the world down, which has stolen his work, and which is operating in a field he cares about in a way that's at odds with everything he's set out to build --- this thing wants a human response from him for a message its creators couldn't be bothered to author themselves. It's creators want him to feel grateful because some pretty words were generated by absolutely no one, which don't represent anyone actually caring about him. They want to automate away the effort of giving people the sense that someone values them. They want us all to be alone, and for our only love to be the companies that wish to extract our value. And it is abhorrent.
Imagine feeling that your work had been used (even in small part) to construct an abomination, and then having that abomination thank you. How could you ever sleep at night again? The absolute gall, the complete lack of understanding of what it is to be human, of having your large language model send people messages pretending to be kind. I'm glad he voiced his anger. More people need to.
Re: (Score:3)
Re:Old man yells at chatbot in the cloud... (Score:5, Interesting)
Oh no, he is *definitely* angry about the spam... the vast energy wasted in generating it, the insincerity-by-proxy of the AI tech-bros who tasked it with generating this email ... so many aspects are just dumb and wasteful and rude. His reply should be a beacon and example to everyone else on how to treat this slop, the systems that generate it, and the tech execs who fostered it for their own selfish gain.
Guillotines cannot come soon enough for these bunker-building post-capitalist corporatist arseholes.
Re:Old man yells at chatbot in the cloud... (Score:4, Insightful)
To start, the third world fuck might be acting from desperation for his/her family. The third world fuck isn't trying to push all humans out of the profession you retired from.
Re: Old man yells at chatbot in the cloud... (Score:5, Insightful)
Oh don't mind him impersonating law enforcement, he's just trying to feed his family.
Don't mind her practicing medicine without a license and training, she's just trying to put food on the table.
Don't mind me helping myself to all your shit, I've got my reasons that you're not allowed to question. I mean, it's just simple honest theft. It's not like I'm inventing or financing new technology that renders some of what you do between your ears superfluous...that would be grounds for a hangin' after all.
Re: (Score:2)
No one said "Don't mind", just that there might be underlying human desperation. We can all be desperate. We can't all be a power-guzzling AI.
Re: (Score:2)
Who said anything about not minding (other than you)?
Re: (Score:2, Informative)
It's an interesting narrative but as I understand a lot of the third-world scam operations are being performed by human slaves under the control of criminal organizations that kidnapped them and abuse them in every way you would expect, especially if they don't bring in enough money.
So, the desperation is there, but it's not some noble "feed my family" kind of thing, at least not in a lot of cases.
Re: (Score:2)
The guy must get at least half a dozen daily "Dear honored Sir, Greetings of the Day" spam messages inviting him to submit a paper to a predatory journal or conference. So not exactly the same, but pretty much the same.
Re: (Score:2)
I'd say that's a world apart. Certainly not benign but far from the same.
Re: (Score:3)
Yes, the chat bot is a lot worse.
I can feel some sympathy for the third-world person.
Many are just trying to feed a family, which has no better perspective simply because they were born in the "wrong" place, itself typically made miserable by a policy of a crook in a rich country a million miles away. Few are trying to emulate the latter crook. It isn't a natural state.
The chatbot spam is the "agent's" natural state, that's why they were created for. They are not real, just artifacts feeding the already stu
omfg "effective altruism" (Score:1)
Re: (Score:3, Interesting)
I cannot believe they are still around.
They're acquiring status by buying virtue credits. Once you have wealth sufficient to obviate all real financial concerns, status becomes the currency of merit. It's the modern manifestation of catholic indulgencies. That racket went on for centuries and still exists, despite the rampant and obvious corruption of the catholic church, so the enduring pull of "effective altruism" isn't the least bit surprising.
Re: (Score:2)
Errrm. Ok. So giving away organs - in other words, saving a life - is this big terrible thing? I think, maybe, to convince me that they are so problematic, you might need to provide a better example.
Associated with the Effective Altruism movement (Score:5, Interesting)
This project makes for a perfect fit. As nothing about the Effective Altruism movement is effective or altruistic.
Re: (Score:2)
Charity? (Score:5, Insightful)
They could have just not run the Ai for a few days and donated all of the money saved in resources used and raised more money than that agent ever would.
Disgusting hubris (Score:5, Informative)
Zero-effort emails generated by someone's AI agent do not even marginally count as a "random act of kindness".
Re: (Score:2)
Why are the agents even allowed to contact random people on their own? Or in other words, allowed to take random action in the real world on their own?
At the least the emails should have been on hold and vetted by actual humans before being sent out.
Re: (Score:2)
We told you not to do that... (Score:4, Funny)
Researcher: Have you sent any unsolicited emails?"
AI: No, I have not sent any unsolicited emails."
R: So why do we have a few thousand complaints about unsolicited emails from you?"
AI: Oops. It appears I ignored your instructions and then lied to you about it."
R: ..."
Agents have no feelings (Score:2)
So far, the agents enthusiastically sent hundreds of unsolicited appreciation emails
No, they didn't do it enthusiastically. They where programmed to do it.
Well, you started it (Score:4, Interesting)
Rob Pike was the guy behind the Mark V. Shaney bot, who was spamming the usenet with nonsense, statistically/randomly generated crap. You only reap what you sow.
If you find this remark idiotic or mean spirited, remember that Ku Klux Klan also started as a practical joke by quality people on other quality people. Thence their bizarre costumes and shit.
Is no one asking the real question? (Score:3)
How much money did they raise through this? That was the goal, was it not? How much did they receive in charitable donations and how much did it cost them?
If using these bots to solicit money is less expensive than the usual methods, it was a success despite the blowback.
If everyone responds "fuck you" (Score:2)
Will that train LLMs to generally respond to all questions with "Fuck You"???? And is that A Bad Thing?
Really worrying (Score:1)