Sign in

Matt Rosoff

@mattrosoff.bsky.social
4.5K followers 520 following 2K posts

Editor-in-Chief of The Register www.theregister.com/Author/Matt-Ros…

PostsRepliesMedia
Reposted by Matt Rosoff
Garthtoons by Garth German @garthtoons.bsky.social · 24/09/2026
Look up “regulatory capture.” Follow for more cartoons!
610449
Matt Rosoff @mattrosoff.bsky.social · 15/09/2026
'"AI" – chatbots that wake up, "set their own goals," and "spontaneously" start hacking servers – is fake. It doesn't have "a 10% chance of ending the human race." The Hugging Face hack isn't a mysterious, supernatural occurrence. It's a Python loop and a chatbot.' pluralistic.net/2026/09/12/g...
pluralistic.net
Pluralistic: LLMs are real, AI is fake (12 Sep 2026) – Pluralistic: Daily links from Cory Doctorow
040
Reposted by Matt Rosoff
The Register @theregister.com · 15/09/2026
Apple iPhone Duo makes developers think in folds
theregister.com
Apple iPhone Duo makes developers think in folds
Cupertino's $2,000 foldable gives devs more screen, and more UI state to manage
031
Reposted by Matt Rosoff
Garthtoons by Garth German @garthtoons.bsky.social · 14/09/2026
Great discussion and overview of this push for AI regulation, why the big companies want that, and how it could actually increase the risk of the US falling behind in the AI race. Also a measured talk on how much of a risk AI really is to human existence
032
Reposted by Matt Rosoff
Garthtoons by Garth German @garthtoons.bsky.social · 14/09/2026
Great discussion, thanks for sharing.
011
Matt Rosoff @mattrosoff.bsky.social · 14/09/2026
If you want to understand what all this killer AI talk is really about, @lot49.com @tobiasmann.bsky.social and @bvig.bsky.social had a helpful discussion on this week’s Kettle - listen here
131
Matt Rosoff @mattrosoff.bsky.social · 14/09/2026
Now that our horses are off to the races, it’s vital to bolt the barn door. www.theregister.com/ai-and-ml/20...
theregister.com
Big AI sets out its terms for regulatory capture and calls it ‘Pace the frontier’
Amodei, Altman, Nadella and Musk agree on how government can tame the monster they created
020
Reposted by Matt Rosoff
Daniel Swain @weatherwest.bsky.social · 13/09/2026
Fielding questions re: this article w/bizarre factual errors. First, there's no "Montecito County" in CA. Second, 2 fires are actually in *Monterey,* & 3rd in *Mendocino*. Third, 60k ac have *not* burned in either fictional Montecito nor real Mendocino.🤨
latimes.com
Three Montecito County fires near Big Sur have burned nearly 60,000 acres
In Mendocino County, the total area burned by three fires grew to nearly 60,000 acres on Saturday.
72845232
Matt Rosoff @mattrosoff.bsky.social · 12/09/2026
I have to wonder if any of these people had children and paid any attention to their brain development in the first five years. Maybe their nannies could teach them something about the actual miracle of human intelligence.
050
Matt Rosoff @mattrosoff.bsky.social · 12/09/2026
The problem came when we allowed these companies to call probabilistic computer programs based on massive amounts of training data “AI,” which allowed everybody to project their favorite sci-fi movie or book onto them.
282
Matt Rosoff @mattrosoff.bsky.social · 12/09/2026
I feel like critical thinking has completely disappeared from the press and the tech industry. Is this like some Covid brain fog thing?
050
Matt Rosoff @mattrosoff.bsky.social · 12/09/2026
Fast food company: this one new menu item we helped popularize keeps poisoning people and we have no idea why. Better ban all similar menu items from all competitors until we can figure it out.
030
Matt Rosoff @mattrosoff.bsky.social · 12/09/2026
Plane company: This new kind of plane we’re pioneering keeps blowing up and we have no idea why. Solution: ban all competition and stall all development in the field until we can figure out why.
050
Matt Rosoff @mattrosoff.bsky.social · 12/09/2026
Car company: this totally new type of car we’re pioneering keeps blowing up and we have no idea why. The solution is to give us complete control over this nascent industry until we can figure it out.
040
Reposted by Matt Rosoff
monkeyrotica @monkeyrotica.bsky.social · 12/09/2026
Translation: Give us another hundred buhzillion dollars or we will release Ultron.
static.klipy.com
Ultron: Everyone Creates The Thing They Dread
ALT: Ultron: Everyone Creates The Thing They Dread
2311
Reposted by Matt Rosoff
Feral T. Robot, Oil Swamp Utility Bot #7629 @feralrobots.bsky.social · 12/09/2026
Yes, it is, but much more so in people's heads than in the computers.
01131
Reposted by Matt Rosoff
Camerado @dogsinautumn.bsky.social · 12/09/2026
It’s really bizarre how much of these guys’ priors are dependent on science fiction they themselves invented within the last 15 years. They’re extremely comfortable with massive leaps in reasoning because they assume they’re talking about the behavior of a god
642714
Reposted by Matt Rosoff
David Burbach 🇺🇸 @dburbach.bsky.social · 12/09/2026
Given current prices, if AI ships me a kickass video card and new SSDs for free, I'll consider letting in nerd Nosferatu
566216
Matt Rosoff @mattrosoff.bsky.social · 12/09/2026
Most of the people commenting on AI now have next to no idea how AI works.
040
Matt Rosoff @mattrosoff.bsky.social · 12/09/2026
How it started: I think I’ll just wander down to the basement and see if I still have an international adapter… 90 minutes later:
030
Reposted by Matt Rosoff
Matt Gorecki @goreckis.co · 11/09/2026
Has this just been market manipulation all along?
1371
Reposted by Matt Rosoff
gchance60.bsky.social @gchance60.bsky.social · 11/09/2026
Sounds like an Enron level audit is in order.
1793
Reposted by Matt Rosoff
Ed Zitron @edzitron.com · 11/09/2026
NVIDIA has sold about $535bn of gpus and associated gear (mostly GPUs) in the data center segment since the beginning of fy2023 (February 2022). If Microsoft only has 2gw or so, I think it’s unlikely anyone else is even close. Where are the GPUs?
1760459
Reposted by Matt Rosoff
Ed Zitron @edzitron.com · 11/09/2026
Holy shit. Bloomberg just verified one of my biggest scoops - that Microsoft only has 2gw of operational capacity. I had it at 1.993GW in prep for the story that @aishadown.bsky.social and I did for the guardian. Vindication! www.theguardian.com/technology/2...
Brody Ford @BrodyFord
Lot of speculation out there about how many GPUs are actually deployed.
Today, Microsoft has about 2GW of Al-specific capacity live. around 17% of the total fleet.
Only about 2 gigawatts of the company's current 12-gigawatt capacity is centered on Al-specific chips, the people said. This is expected to grow to about a third of the 38 gigawatts Microsoft plans to have online in 2032, they said.
Brody Ford @BrodyFord_ • 23h
scoop: Microsoft is planning a massive expansion of its data center fleet to 38 gigawatts. $msft w/ @mattmday bloomberg.com/news/features/...
3:15 PM • Sep 11, 2026 • 12.5K Views



1,993,608.00
W


MW


GW





Are Microsoft's AI
plans being held back by a shortage of chips?
Guardian investigation finds
apparent discrepancy between what tech company has said about its AI capacity - and the
number of advanced chips it
has in operation
By Aisha Down and Ed Zitron
A Microsoft datacentre in the North Holland village of
Middenmeer in the Netherlands. Photograph: Ramon van
Flymen/EPA
251938362
Reposted by Matt Rosoff
Ketan Joshi @ketanjoshi.co · 11/09/2026
It takes 0.0001 seconds to realise the AI-bros are lying about "extinction risk" as a marketing trick, given how none of them seem bothered about the deadly fossil fuels they've become obsessed with using to power their data centres.
512453
Reposted by Matt Rosoff
Bertha Coombs @berthacoombs.cnbc.com · 11/09/2026
A friend called 9/11 the Pearl Harbor of our age. In this day and age, especially, let’s remember the bravery, the kindness and the grace so many showed that day. We saw the best in us, in the face of the worst of humanity.
0287
Reposted by Matt Rosoff
Tom Hearden @followtheh.bsky.social · 11/09/2026
Remember how the flights were forced to land in Canada and how wonderful the Canadian people treated us? I do. The fucking lunatic in the White House does not.
830344
Reposted by Matt Rosoff
Sunny @flynndreno.bsky.social · 10/09/2026
These guys want to pretend they’re working on the Atomic bomb so badly.
031
Reposted by Matt Rosoff
Teeg Dougland @arobertbitter.bsky.social · 10/09/2026
They must really be having trouble getting that IPO ready. I notice no S-1 yet. Quick! Start freaking out about Skynet!
122
Reposted by Matt Rosoff
emergentempire.bsky.social @emergentempire.bsky.social · 10/09/2026
Look, if you knew what me and my fellow Gaters knew about what the Nibiruians have planned once Hale-Bopp passes you wouldn't think it's so funny!
161
Reposted by Matt Rosoff
babsbeaty @babsbeaty.bsky.social · 10/09/2026
I swear all these people have a super villain fetish. Delusions of grandeur + narrow world view + self-hatred = I’m going down and taking you all with me. It’s inevitable.
0111
Reposted by Matt Rosoff
Fi 🏳️‍⚧️🇵🇸 @hansmoletran.bsky.social · 10/09/2026
Roko's basilisk is stupid for any number of reasons, but one of my favourite is the idea that there would be any continuity of consciousness between me and a digital recreation of me in the machine's torture simulator
1181
Reposted by Matt Rosoff
JoshiRibombee @joshihive.bsky.social · 10/09/2026
he's scared, but he believes that it will happen no matter what, and that if he tries to prevent it fruitlessly, the AI will take revenge on him when it gets sentience this is Roko's Basilisk and it is very stupid
2212
Reposted by Matt Rosoff
Lostmyshape @lostmyshape.bsky.social · 10/09/2026
AI Guy: I’m just scared I’m going to eat this handful of anthrax Me: Oh, well why don’t you just put it down? AG: I mean, I don’t WANT to eat it M: No, I get that. So just put it down and then maybe you won’t eat it AG: I’m so scared
0754
Reposted by Matt Rosoff
nibbish @nibbish.bsky.social · 10/09/2026
If they really believed any of this, they'd be, like, a million times more specific. HOW is ai is going to kill us all. Is it nukes? Is it poison? Is it some weird 25th century nanomachine nonsense? What is the mechanism?
516114
Reposted by Matt Rosoff
nibbish @nibbish.bsky.social · 10/09/2026
WHY is ai going to kill us all? Is it to make more paperclips more efficiently? Is it the old "kill them all to save the world" nonsense? Something more esoteric? Does it have a plan for how all the servers that house all of its data are going to be upkept once it deletes us?
2974
Reposted by Matt Rosoff
nibbish @nibbish.bsky.social · 10/09/2026
That they (and, tbh, all the people that are simply buying this at face value) aren't elaborating on either of these questions at all, and are simply responding to questions with "hurr, durr, try reading a BOOK sometime." leads me to believe that this is all bluster.
6825
Reposted by Matt Rosoff
Too Big to Fail @toobigtofail.bsky.social · 10/09/2026
This, from an Anthropic employee, either reflects a lack of imagination or is evidence that this is all just lies. If you were really that scared, you'd pull the plug on all the servers and wipe everything from ARIN, APNIC, LACNIC, AFRINIC, and RIPE NCC. No more internet; problem solved.
I would burn my equity to the ground in a heartbeat for a 1% higher chance we make it out of this situation alive. I expect a great many of my colleagues across the industry would as well. 

I promise you, we are actually just fucking scared, it's not galaxy brained marketing.
1221552201
Reposted by Matt Rosoff
MGoCoder @mgocoder.bsky.social · 10/09/2026
The #1 reason we got Trump is that the media gives Republicans a pass on everything.
021
Reposted by Matt Rosoff
Paul Waldman @paulwaldman.bsky.social · 10/09/2026
So maddening that Trump gets about 1/10 as much criticism from reporters for saying "This is the greatest economy in history" as Biden got for saying "I know times are tough, but it's getting better" which was supposedly such a terribly insensitive affront to Americans' lived experience
1037373
Reposted by Matt Rosoff
Mar Hicks @histoftech.bsky.social · 10/09/2026
The way AI is going to “kill everyone” is just by manufacturing consent for ever worse outcomes & conditions, like many governments do. It is a governing technology after all. And like bureaucracy it launders responsibility so the people & institutions who deploy it aren’t accountable for the harms
13680230
Reposted by Matt Rosoff
Ed Zitron @edzitron.com · 10/09/2026
“It’s not hyperbole that we could cure cancer” actually yes it is. That’s exactly what it is
Jacob, is there a reason that kept you showing up to work at OpenAl and Anthropic every day for the last few years? Do you feel like there is a path forward here where everything is fine?
I really want to see the benefits of this technology.
For example, it's also not hyperbole that we could cure cancer. Maybe you saw the achievement yesterday on Navier-Stokes, the novel mathematical problem. There really is no reason that we can't transfer that to scientific domains like biology and make breakthroughs on questions we've struggled with for a long time.
2163466
Reposted by Matt Rosoff
Dandy G🍌 @dangyd.bsky.social · 10/09/2026
He can’t identify what he’s concerned about because then it becomes a solvable problem. He’s angling for an impossible problem he can spend a few years “solving” for whomever pays best. The only insider info he’s revealing here is he doesn’t think equity will hold its value by the time it vests.
1303
Reposted by Matt Rosoff
Ed Zitron @edzitron.com · 10/09/2026
Aaaaand there you have it folks archive.md/2026.09.10-0...
I'm sure the next couple days will be crazy, but after things settle down, have you figured out what's next for you?
I mentioned AI 2027 and AI 2040 before. I think these are really interesting predictions of the way the future looks. They're not done by any big labs; these were done externally by researchers that just cared about predicting what the world looks like. These have been very good resources for me and a lot of other researchers for trying to think through the implications of what you're doing.
So I expect to do something like this-some sort of independent commentary on where things are going, at least for the short term before we get some sort of pacing agreement or slowdown. Ilf| it looks like we're not going to rush straight into recursive self-improvement, maybe I'd consider going to an auditing agency or some third-party regulation or transparency agency. But for now, I think I'm just gonna try and take stock of where the situation is.
Update 9/9/26 7:40pm ET: This story has been updated with a statement from Anthropic.
2048085
Reposted by Matt Rosoff
Ed Zitron @edzitron.com · 10/09/2026
This Jacob Coxon thing is weird. He worked at Anthropic for six weeks and OpenAI for nearly three years. He’s doing a whole media tour without any real statements about what it is he’s scared of other than vague references to recursive self improvement, a theoretical ai that teaches itself.
471406264
Reposted by Matt Rosoff
Emmanuel Valle @vallecito.bsky.social · 10/09/2026
How did the entire tech and journalism world turn into idiotic sociopaths
6743
Matt Rosoff @mattrosoff.bsky.social · 10/09/2026
Somebody on the Search Answers team at Google has a juvenile sense of humor: type in four hyphens and check out what the AI returns. (Found by a @theregister.com reader.)
000
Reposted by Matt Rosoff
Tom Hearden @followtheh.bsky.social · 10/09/2026
gm and Happy Uncle Larry Day to those who celebrate -- one year ago today Larry Ellison was briefly worth >$400 Bil and was the richest person on the planet $ORCL is -50% since (you hate to see it)
3455
Matt Rosoff @mattrosoff.bsky.social · 10/09/2026
I’m in this dumb motel next to the Taco Bell without a hope in hell, I can’t believe that I’m still around.
020
Reposted by Matt Rosoff
Kylie Eleison @gonebabygone.bsky.social · 07/09/2026
Tech companies don't even tell you what they do anymore, they just look at you like this
26824901