TechTakes

Promptfondlers are biologically incapable of telling good from bad, video game edition.

For something lighter, here's an AI bro getting wowed by the shittiest "video game" I've ever seen (trust me, the screenshot doesn't do it justice):

In lieu of sneering this shit, I'd like to argue that arts education should become mandatory for all students post-bubble, regardless of their profession. STEM, humanities, tech, doesn't matter - give them four years of art so they don't turn out like this guy.

JFC some CSAM collector stormed a Wikipedia conference with a gun https://slate.com/technology/2025/10/wikipedia-editors-conference-gunman-culture-war.html

A second teardown of Omarchy has hit the web

https://eepy.moe/notes/ae7rhgxvpw8d09kw

Moderator actions Moderators may hide, delete, or issue a warning if a post includes any of the following:

• Harassment or threats: Any comments, posts, or replies that attack, demean, or intimidate an individual or group (e.g., homophobic slurs, anti-trans sentiment, calling other users “Nazis” or “fascists”, etc.

today in I fucking called it fedora aka mostly red hat has decided to allow slop code in a way that violates even their utterly mid stated principles around the tech

if you’re downstream from any fedora packages (and I don’t know the scope of this policy so it might be safe to consider anything owned by red hat in general to be tainted — yes I realize most of us are downstream from a bunch of red hat shit) it might be time to evaluate an alternative if available

sigh

I am the scream

Charities Using AI-Generated Photos of Starving Children to Raise Money

Morally, attacking the poor and downtrodden through polluting the world with AI is abhorrent, and anyone doing this should be permanently barred from working at any charity.

Practically, the sight of blatant AI-generated poverty porn is going to drive people away from giving to these charities, damaging their ability to do good.

Trump freed Binance fraudster, SBF pardon futures mooning rn

Oh boy, another AI doom video popped up on my feed. Time for more morbid curiosity. The topic is about Big Yud and Nate Soares's new book ("If You Build It, Everyone Dies") about how AI is gonna kill us all. I have better things to waste 30 minutes on, so I'm not watching the full video, but the thumbnail ("The 7 Minute War") kinda suggests what the contents are gonna be.

Thankfully, the description of the video has a Google doc with their sources! I'm sure it's full of hard evidence from careful experiments that logically demonstrate why their doomsday scenario is something to worry about, not just a random assortment of Anthropic blog posts and completely unrelated events.

Somehow, there are a bunch of sources for the first 2 minutes of the video.

"In the New York Times' best-selling book, which was endorsed by Nobel laureates and the godfathers of AI" Geoffrey Hinton — Personal estimate >50% existential risk.

Geoffrey "All radiologists will be replaced in 5 years" Hinton, Nobel laureate in physics, famous for his work in ... physics.

"researchers from the Machine Intelligence Research Institute describe in detail one potential example future" Machine Intelligence Research Institute — The Sable scenario from If Anyone Builds It, Everyone Dies by Yudkowsky & Soares. Fictional narrative illustrating risks, not prediction.

This is not the first we've seen from MIRI, and I have a feeling it will not be the last. The monster under my bed is a fictional narrative illustrating risks, not prediction.

"AI researchers have known this has been potentially a very bad idea since at least 2024" Anthropic/Apollo Research — Multiple 2024 papers document deceptive/self-preserving behaviors in controlled evaluations.

They are still trying to flog the Anthropic/Apollo Research claims that chatbots will lie to you if you tell them to lie to you.

"They spin up 200,000 GPUs and let Sable think for 16 hours straight" xAI/NVIDIA — Colossus supercomputer in Memphis scaling toward ~200,000 GPUs for Grok training.

What does this even demonstrate? Some people can do some stuff with some GPUs? I ate some oatmeal today. Now everyone should be thoroughly convinced of my oatmeal-eating abilities.

I watched for a few seconds around the timestamp, and it seems to be the beginning of their scifi story, I mean, AGI scenario. Yes, if you want to convince people that your scenario is plausible, I'm sure this is the part that you need serious amounts of evidence for. Remember, almost half the sources have timestamps for the first two minutes of the video.

"a stunt to see if Sable can crack famous math problems like the Riemann hypothesis" Clay Mathematics Institute — Riemann Hypothesis remains unsolved after 160+ years, considered most famous unsolved problem in pure mathematics.

Again, what does this demonstrate? I tried solving P vs NP with a cheeseburger. That didn't work either. The only purpose of mentioning this is for narrative window dressing, because Math Is For Smart People.

These are the sources for just the first two minutes. After that, they get a bit sparse.

"Back in 2024, smaller models showed flashes of the same behavior" Multiple Papers — Documented deception/scheming findings in frontier models.

"Claude 3.7 was caught repeatedly cheating on coding tasks even when told to stop"

More Anthropic blog posts and system cards? Come on, I can't sneer the same thing twice in one post!

"Steal cryptocurrency from weak exchanges just like hackers did to Mt. Gox in 2011" U.S. Department of Justice — Russian nationals charged for 2011 Mt. Gox hack. 647,000-850,000 BTC stolen.

I don't know what this has to do with supporting the validity of their AI doomsday scenario, but kudos to them for showing why cryptocurrency is also stupid, I guess.

"or Bybit in 2025" Reuters/FBI — Largest cryptocurrency theft to date. FBI attributed to North Korean Lazarus Group.

More? I guess this is hard evidence for showing why cryptocurrency is stupid. I still don't understand how this demonstrates that AI is scary.

"Reminder, this scenario is based on years of technical research by the Machine Intelligence Research Institute, laid out in the book If Anyone Builds It Everyone Dies" MIRI — Meta-commentary explaining the scenario is illustrative, not predictive.

I knew MIRI would be back. It's illustrative, not predictive! Please don't blame us if none of this even remotely happens! But it's based on years of technical research. An entire graduate student's worth of output in a decade.

"In 2023, a human gave an LLM access to the internet and created an X account, Terminal of Truths, which gained hundreds of thousands of followers and launched its own crypto meme coin that reached a literal billion dollar market cap" Terminal of Truths — Real-world example of AI agent gaining social media following and wealth.

The link they give references ... another one of their own videos. You really are not beating the circular reference allegations here. Even if the purported story is somehow accurate, this again demonstrates how cryptocurrency is stupid. At least they use an LLM as a prop this time.

"Gain of function research. Any one of them could be hijacked to unleash catastrophe." Science/CIDRAP — Fouchier and Kawaoka created ferret-transmissible H5N1. Controversial GOF research began 2011.

I think Yud is obsessed with this topic in particular. Better than diamondoid bacteria, I guess. Again, the AI just magically comes in and uses this stuff somehow.

"The number one and number two most cited living scientists across all fields think scenarios like this are not only possible but likely to happen. And the average AI researcher thinks there is a 16% chance of AI causing human extinction."

Okay, let me be completely serious for this one. What would someone do if they truly believed that their work would lead to a horrible disaster, such as the extinction of humanity? Would they continue to work in the field, let alone make enough contributions to rise to the top? Alright I'm done.

openai released their spyware browser and it is... not good

https://www.anildash.com//2025/10/22/atlas-anti-web-browser/

look at the depth of this grifting

a whole One (1!) H100! in space!

note how it mentions nearly absolute fucking nothing about the supporting cast. about storage and networking, about interface capabilities, what kind of programmatic runtimes you could have! none of it. just gonna yeet a sat into space, problem solved! space DCs!

compute! in space! "what do you mean 'compute what'? compute!" I hear, as the jackass rapidly packs up their briefcase and starts edging towards the door. who needs to care about getting data to and from such a device? it'll run Gemma![0] magic!

SAR, in particular, generates lots of data — about 10 gigabytes per second, according to Johnston — so in-space inference would be especially beneficial when creating these maps.

scan-time "inference", like you'd definitely know every parameter you'd want to query and every result you'd want to have, first-time, at scan! there's a fucking reason this shit gets turned into datasets, and that the tooling around processing it is as extensive as it is.

and, again, this leaves aside all the other practical problems. of which there are many. even just the following ones should make you wince: launch, maintenance, power, heat dissipation (vacuum is an insulator!), repair, (usable) lifetime, radiation. and that's before even touching on the nuances in those, or going further on the list

good god.

I guess the one good bit here is that it isn't the "we're gonna micromachine them in orbit!" bullshit fantasy, but I bet that's not far behind

[0] - "multimodal and wide language support" so literally a Local LLM, but that means it needs... input... and... response... which again goes back to all those pesky "interaction" and "network" and "storage" questions.

So kurzgesagt put out an anti AI-slop piece 2 weeks ago. It's hits good angles too, going hard on the society-eroding effects. The conclusion is a little mushy, but hey, I wasn't expecting an instigation to machine-breaking from kurzgesagt.