Big brain tech dude got yet another clueless take over at HackerNews etc? Here's the place to vent. Orange site, VC foolishness, all welcome.
This is not debate club. Unless it’s amusing debate.
For actually-good tech, you want our NotAwfulTech community
Ah, I've hit an extremely satisfying professional milestone and thought I'd share:
"Being acquainted-enough with the codebase to be able to, with full conviction and a lot of glee, answer my LLM-head colleague simply: 'Claude is wrong'".
(Claude was extremely obviously wrong once you spent more than half a second thinking about its claim.)
Why! Why are media so uncritical in their reporting!
https://www.tagesschau.de/wirtschaft/unternehmen/openai-ki-hackerangriff-100.html
Sorry, this is in German. It's about the marketing stunt open ai pulled vis-à-vis "oh no, our new model breached containment and went ahead on its own, yep all by itself, and hacked huggingface! So scary! BTW we'd like to IPO soon and-"
The outlet above, if you're not aware, isn't a German news outlet, it's THE German news outlet. They're independent of financial interests, in that they're publicly funded. They have an incredibly solid history of factual reporting.
Which is why it's so aggregating to see them parrot OpenAI's claims 1:1.
Also, don't go in the comments. For every sane "this is obviously a marketing stunt" comment there's 40+ "we're cooked" comments sincerely believing this shit.
I'm getting more and more convinced that in the minds of most people, LLMs are alive, waking, thinking beings with intents and an inner life of their own.
You know.
Compared to the much more boring "expensive side-effect free function taking string as input and giving string as output".
I hate all of this.
Humanoid robots providing stiff competition to quantum as to where the ~~smart~~ desparate money will be going once everyone realises the wheels have fallen off ai.
https://thepit.social/@peter/116962244994361810
(the video is too big to upload here)
I particularly like the way they had body bag operatives within lunging distance, because they clearly expected the thing to just fucking die at a moment’s notice.
Or maybe they do that for all their speakers.
comedy gold
big tech robot had literally the same task as a karakuri doll from 300 years ago and still lost: https://www.youtube.com/watch?v=_czILDdOBc8
(more on karakuri for people who want to wach some actual technology: https://www.youtube.com/watch?v=XU-kqpkZWjc )
OpenAI: "Sorry about accidentally hacking you teehee. It was because of how good we are at computer security :3"
https://openai.com/index/hugging-face-model-evaluation-security-incident/
Maybe they should start applying some of that superhuman box-escaping AI to their own sandbox designs because if your sandbox escape is a bug in a "package registry cache proxy" you might be doing it wrong.
I’m deeply suspicious of this whole thing, because it looks a lot like a marketing exercise showing off how dangerous and powerful and autonomous their product is.
Also, “sandbox” is one of those words that the llm companies have ruined, because they use it to mean a strongly-worded sentence telling an llm not to do something.
This is funny yet also awful: adversarial tokenmaxxing suggests that writing everything as l33t$p34k increases the cost to process a document with an llm because the initial tokenisation step produces far more tokens.
This seems like it shouldn’t be too hard to work around, if it became commonplace (which it won’t) but the prospect of any anti-llm places doing this in the meantime does not spark joy.
Adversarial tokenmaxxing could probably be done by using non-English character sets in lieu of English letters (e.g. faux Cryllic) - for two examples from the Greek alphabet, alpha and omicron alone can easily substitute for A and O, respectively.
As a bonus, this would likely make the text look like complete gibberish to LLMs, potentially leaving them unable to process the document altogether. This would probably shaft anyone using screen readers, though.
EDIT: Turns out the demonstration's already caught on to this idea, didn't notice beforehand:
who is this Dean Ball guy? new week, new guy, I guess.
https://www.axios.com/2026/07/19/pentagon-openai-dean-ball-trump-ai
Zoom in: "Every industry/ecosystem has its supreme village idiot. @deanwball is that for AI," Defense under secretary Emil Michael said Sunday on X.
"Dean Ball has perhaps the biggest gap between actual IQ and his own perceived IQ of anyone in the industry (about 40 points)." Michael was arguing against Ball's point that various government agencies, including the Pentagon, have blocked employees from using Chinese AI, which has already sent a message to regulated companies.
Russ Wilcox felt compelled to weigh in also
Dean Ball, days into his new job as Head of Strategic Futures at OpenAI, posted his assessment of Kimi K3, the new model from the Chinese lab Moonshot AI. Kimi is an open-weight model, which means its inner workings are published for anyone to download, run, and modify; the newest version’s weights ship later this month. It had just matched the best publicly available models of the first quarter. The post made three claims. Open-weight models are “inherently decelerationist.” A probable outcome of an open-weight-dominant world is “full AI communism,” a future of AI as public good and digital public infrastructure that strikes him as a “dystopian hellscape.” And the United States government will at some point realize that its best strategy is to create large amounts of “regulatory risk” around the use of open-weight Chinese models.
I have met Dean Ball. We have disagreed before. This time the disagreement is public, and so the response should be too.
oh no kimi3 got the neo marxist woke mind virus
Dean Ball’s name rings a bell to me. Emile Torres wrote an article on him its pretty standard OpenAI exec stuff (that is being a wacko high off his own tech)
For that matter, who the hell is Emil Michael? Defense Undersecretary tells me nothing, they're handing those sorts of appointments to anyone who's ever scammed $5 in venture capital
Mathematicians don’t love the use of “AI” for math, not one bit, via Timnit Gebru
You know, given automated proof checkers, I was naively assuming mathematics was one field that gen-AI would have a hard time screwing up. Even programming is too difficult to write thorough testing for. But a proof (or counter example to a conjecture) seems like it would have to be solid if it passes lean or whatever system for validating it.
But the threats #3 and #5 the declaration lists make me consider the bigger picture. Academic fields without clear capitalist payouts are already underfunded and under respected. Pure mathematics could, at least up until now, draw on the respect STEM gets, but with math proofs getting used as fuel for the LLM hype machine, there are a variety of unpleasant ways things could twist.
Threat #1 makes me wonder... if LLMs+formal verification systems manage to pluck lots of low hanging fruit, and we are left with harder stuff that not enough literature exists as training data for LLMs, it seems like the entire educational pipeline for producing mathematicians could end up screwed up.
new odium symposium episode. we look at the work of conservative philosopher, grifter and tobacco lobbyist Roger Scruton.
https://www.patreon.com/OdiumSymposium/posts/21-mr-rogers-164372632