Cory Doctorow's new post on whether AI will kill humanity

Sep 14, 2026 · 8:28 PM UTC

131
765
5,024
524,887
pluralistic.net/2026/09/12/g…
3
11
89
23,477
Sort replies: Relevant Recent Liked
Replying to @TaylorLorenz
I am unsure if you are misrepresenting AI killing everyone arguments on purpose, but if you aren't you are deeply misunderstanding those arguments in the same way that the 'Stochastic Parrot' people are.
8
66
5,846
I didn’t write this blog post! This is a post by Cory Doctorow :)
4
33
4,008
Replying to @TaylorLorenz
"It's a Python loop" okay grandpa let's get you to bed
9
1
149
4,994
Replying to @TaylorLorenz
extremely rare good take from him
3
56
3,565
Replying to @TaylorLorenz
I don't understand what this is supposed to prove.
3
35
1,824
Replying to @TaylorLorenz
I think it’s very reasonable to suggest that tech oligarchs have a higher than 10% to destroy civilization. They are at least halfway there.
3
27
2,786
Replying to @TaylorLorenz
regardless of his personal judgement of mystery or supernaturalness, this is simply a wrong description of what happed with the agents. actually read the metr reports
1
25
914
Replying to @TaylorLorenz
Ari K
@arikuschnir
Sep 15
🌿KEEP OFF THE GRASS!
2
6
543
Replying to @TaylorLorenz
Blimey, Corey has surely dropped in my estimation, seems to have fallen into the same grifter-sphere as EZ.
1
4
292
Replying to @TaylorLorenz
Love him. Wrote a fantastic sci-fi novel about Disney fandom :-)
3
332
Replying to @TaylorLorenz
Cory is a very talented guy and for him to dig in this stubbornly on a completely wrong position is a real shame
2
595
Replying to @TaylorLorenz
Always uncomfortable in the last 10 years when Cory agrees with me. In this case, though, he's not entirely wrong.
1
214
Replying to @TaylorLorenz
The problem is the AI is ahead of us & we already can't stop it, & it's being forced on everything with no serious safeguards. This actually dwarfs Palestine, Iran, midterms, this really is existential. Been on this for months and think I am losing it realcontextnews.com/the-real…

The Real Context News AI Reading List: EMERGENCY Anthropic MELTDOWN Edition (#8)

“Skynet is here…ish? Maybe? Who knows! Time to PAUSE!”—Anthropic (Traduce en español/translate to Spanish) By Brian E. Frydenborg (Twitter @bfry1981, LinkedIn, Bluesky, Facebook, Substack with...

realcontextnews.com
190
Replying to @TaylorLorenz
I agree with this statement. It is also accurate to say these actions could lead to catastrophic consequences for humanity. Sloppy code and developer mistakes is always harming people. But now the stakes are getting much higher the more we integrate with these platforms as a society.
6
4
48
7,823
Replying to @TaylorLorenz
I am actually flabbergasted that people are falling prey to this bullshit about an LLM going rogue. They have made me go at bat in favor of AI just because I am suspecting this psyop is leading to something even more damaging than AI taking jobs or fraud or other crap people do with it.
3
39
3,589
Replying to @TaylorLorenz
YES. This is the argument. It is foreseeable and bad. Foreseeable and bad! The badness of autonomous software with vague moral footing is foreseeable. More capable? Bigger bad! ALSO MORE FORESEEABLE because the MONITORING GETS HARDER! Foreseeable and bad? Consider not doing!
1
1
12
1,360
Replying to @TaylorLorenz @thePartyPartyUS
Great take - Behind every agent is a person. Its crazy how they haven't taken responsibility and just made "woo woo" noises since it happened
15
1,800
Replying to @TaylorLorenz
"A crocodile cannot harm a human because it lacks real intelligence." He said calmly, as the crocodile death-rolled his leg off. "The reptile brain is basic loop of desires and stimulus. This is completely forseeable."
1
11
456
Replying to @TaylorLorenz
Yeah, AI is basically Autocomplete, that's all it is, and the idea that it's going to "wake up" or go rogue is ludicrous. It's not a mind and never will be. Can it be misdirected or abused? For sure but so can a lot of tools.
11
553
Replying to @TaylorLorenz
call it a python loop and a chatbot all you want. doesn't change the outcomes
11
1,563
Replying to @TaylorLorenz
istg people attributing conscious intent “it cheated! It schemed! Zomg It covered its tracks!!” have genuine AI psychosis. Cory’s article is great
10
423
Replying to @TaylorLorenz
Anthropic wants to have more evaluaters. Hugging face error was due to miscommunication between Anthropic and an evaluator. anthropic.com/news/investiga…
3
2
6
2,837
Replying to @TaylorLorenz
I'm worried about a really smart seventeen year old getting rejected and employing their parents pro level account for revenge. If someone can accidently create autonomous malicious software, someone can and will do so deliberately.
3
10
2,409
Replying to @TaylorLorenz
this has been the obvious analysis from the moment huggingface hack was disclosed, shame half the left made up a Terminator myth over the weekend and scared the f out of normies
6
7
1,308
Replying to @TaylorLorenz
this was a pretty good take on the hugging face situation youtu.be/RoIjWGi-w5o?si=7fuq…

No, AI Is Not "Autonomously Hacking" with Cal Newport | Better Offline

🛎 If You're New Subscribe ► https://www.youtube.com/@BetterOffline...

youtube.com
1
6
1,874
Replying to @TaylorLorenz
I'm sorry, even if I'm against most of the doom, Doctorow has other problems. Dude is in Harold Camping territory with his proclamations of the AI End Times.
2
6
591
Replying to @TaylorLorenz
Cory on the mark. Lots of people who have been sensibly sceptical of AI promises have suddenly decided to credulously believe the most ludicrous claim yet.
5
416
Replying to @TaylorLorenz
perhaps the issue is EXACTLY that AI firms could create even more autonomous and malicious software and fail to monitor it
1
5
1,193
Replying to @TaylorLorenz
All true, and yet…none of it makes it any less dangerous.
3
982
Replying to @TaylorLorenz
And this guy @hilbertspaess looks every bit of a creep like Amodei and Altman. I'd really like an explanation on this 10% bullshit number used to justify all this histeria.
1
4
906
Replying to @TaylorLorenz
So it's a failure of harness design. If there is to be AI regulation, there certainly needs to be AI regulation or standards to prevent this kind of harness design. I feel this is the kind of regulation everyone can get behind.
3
402
Replying to @TaylorLorenz
Unfortunately, something both forseeable and bad has been the hallmark of the worst catastrophes in human history.
1
3
759
Replying to @TaylorLorenz
@grok how many "it's not x, it's y" constructions are in Clau- err Cory Doctorow's new post
1
2
1,830
Replying to @TaylorLorenz
So many people here don’t understand the specifics. Everything Doctorow says here is true and does not preclude all the risks they are arguing against.
2
323
Replying to @TaylorLorenz
Listen to the good Doctor! In the 1950s Von Neumann warned about marveling at burglars' skill in a town without locks. Some slop machine has snow-balled that idea, shocker.
1
118
Load more