Hacker Newsnew | past | comments | ask | show | jobs | submit | glitch003's commentslogin

Author here. The short version: Qwen3.8's 248k-token vocabulary contains scraped-junk tokens that got almost no training (embedding norms down to 0.001 against a vocabulary mean of 0.899). Put one in a slot that matters, like a username, order ref, or SKU, and the model swaps in a hallucination while keeping the output schema-valid. An order ref of "Kinhted" came back as "order_id": "shelled", 16 out of 16 times at temperature 0.7.

We ran six control models. Mistral and Llama have the same class of problem and lose 5 to 7 of their 12 lowest-norm tokens. Both Gemma generations are clean, so a curated vocabulary avoids this entirely.

Happy to answer methodology questions. If you want to know whether a specific checkpoint has this problem, name it and I'll run the scan.


They can all do that, peep


she called her husband, from jail, and asked him to move them so they couldn't be used in court against her. she said "do whatever you have to do". the cops listened to the phone call, watched him load them into the car and arrested him on the drive. i think it's pretty straightforward that she was asking for his help in obstructing the investigation by hiding potential evidence against her.


Lawful investigation and prosecution would of course not hinge upon the contents of political publications. As such, any obstruction was purely aimed at the inevitable two minute hate by the addled autocrat about how antifa is hiding everywhere, making his coffee too cold and his food not taste as good as it used to.


Well regardless, the obvious answer is that you simply turn your zines in to slurry in a bucket or sink, you don’t drive them anywhere.


Maybe so, but 30 years is wildly disproportionate.


I am asking those questions in good faith, to explore the new limits on free speech, and don't see any direct answers. What happens to this guy specifically is only interesting as far as it affects everyone else in the US.


How does this compare to libp2p?


fwiw, tesla, the biggest "no physical buttons" company, does not have this problem at all. it's always very responsive, window buttons always work, there's no "startup" delay at all.

so i guess the choice is between companies that do tech poorly (legacy automakers) and one that does it right but with a big touch screen.


BMW does tech pretty damn well. Even though they're throwing everything on a screen nowadays, at least the screen is extremely responsive and (once you're used to it) quick to navigate


it's still USDS, they just changed the "D" from digital to DOGE


we get the whole US AID budget back, aside from PEPFAR. it's a good example of a case where they cut too much and undid some cuts.


Who is we? You're probably not part of the selectorate who is getting the spoils of this


surprisingly, it works excellent in rain. i was on the highway in florida in a torrential downpour, the kind where everyone turns their flashers on and goes 25mph. many people pulled over to wait it out. FSD handled it perfectly even though i could barely see. it did show a warning that it might be degraded due to weather, but there was no change in performance.

i can't speak for snow tho.


same way you do it in the iphone simulator - you hold the option key and the mouse button, and move the mouse up or down to simulate pinch to zoom.

https://discussions.apple.com/thread/1696762?sortBy=rank


Hi HN - we’re building Payout.Team for freelancers, consultants, and contractors who work on projects together.

Here’s how it works: a users sends an invoice to their client, when the client pays, the funds are automatically routed to the members of the team / subcontractors. Payout.Team takes care of all the on-boarding and tax paper work for the subcontractors, including internationally.


I do this with the sites I admin. I don't know why it's not more widespread.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: