HN user
gmaxwell
http://nt4tn.net/
Well timed to facilitate the regulatory interventions called for by Ball. If huggingface presses criminal charges for the intrusion it might provide additional clarity-- both for what happened here as well as regarding OpenAI's culpability.
The large US closed AI companies are decelerationist because their focus is on monopolizing the market. They spend inefficiently in order to lock up the supply of resources and waste money influencing the state to attempt to lock out competition. This strategy has not been successful due to the existence of isolated resource pools they can't monopolize.
"There is another theory which states that this has already happened."
Verifying that counterexample is a trivial calculation for basically anyone qualified to make substantive changes in that category of article, it shouldn't be an issue in and of itself.
Discovering the counter example is research, validating it is basic calculations.
Irony in the message you're responding to likely suffering from slop-facts itself.
And indeed it's trivial to verify.
I think vibethinker is heavily overtrained on not attempting to solve open problems.
I had a fun time taking some open problems and disguising them algebraically so that vibethinker 3b would work on them. It managed to prove some interesting things that I didn't know and would be publishable, but for the fact that they already have been. :) (though hard to know if this was because it had been exposed to that knowledge even though it didn't reconize the hidden problem).
Under some maskings it would eventually figure out the problem was equivalent to an open problem then immediately shut down.
It also managed to make some false proofs for various things that duped some other more powerful models.
What's the point of looking when no one even cares?
consider the fraudster that went around suing people on the basis of his absurd claims of being bitcoin's creator. He's now transitioned to using AI to gather graduate degrees and is obtaining masters and doctoral degrees at a regular place and writing multiple 'papers' per day that are all quite obviously AI slop.
People report his cheating and publications and simply no one cares... and this is someone court adjudicated to have fabricated evidence in court on a massive scale, including through the use of AI.
But when it comes to the degrees and publication everyone involved that wanted paid got paid, and apparently that's all that matters.
Comments are closed, but they're open for responses.
By my count there were 533 replies that were responsive to the 'enhanced' KYC requirements. ~496 opposing it, and ~37 substantively supporting it.
I used deepseek 4 flash with a rubric I wrote to run the responses against each other in a tournament.
Here are some of the stronger responses against:
* https://www.fcc.gov/ecfs/search/search-filings/filing/261099...
* https://www.fcc.gov/ecfs/search/search-filings/filing/261099...
* https://www.fcc.gov/ecfs/search/search-filings/filing/105288...
* https://www.fcc.gov/ecfs/search/search-filings/filing/261098...
* https://www.fcc.gov/ecfs/search/search-filings/filing/261098...
* https://www.fcc.gov/ecfs/search/search-filings/filing/261099...
* https://www.fcc.gov/ecfs/search/search-filings/filing/261099... (mine)
* https://www.fcc.gov/ecfs/search/search-filings/filing/106110...
And here are some of the stronger argument for, which you might want to rebut:
* https://www.fcc.gov/ecfs/search/search-filings/filing/261099...
* https://www.fcc.gov/ecfs/search/search-filings/filing/261099...
* https://www.fcc.gov/ecfs/search/search-filings/filing/106090...
* https://www.fcc.gov/ecfs/search/search-filings/filing/105150...
* https://www.fcc.gov/ecfs/search/search-filings/filing/261098...
* https://www.fcc.gov/ecfs/search/search-filings/filing/261099...
* https://www.fcc.gov/ecfs/search/search-filings/filing/261099...
* https://www.fcc.gov/ecfs/search/search-filings/filing/261099...
* https://www.fcc.gov/ecfs/search/search-filings/filing/261099...
That’s why you’re lagging behind in open-source competitions
the lag is because closed ai orgs are sucking up most of the talent and monopolizing resources.
Not that your criticism of HN comments aren't deserved-- but the people commenting here are not usually people who would be doing something useful otherwise.
Yet the visceral pile-on here is so extreme, it feels fake.
Driven by people in the few roles that are soundly replaced by AI-- e.g. low tier media slop producers, who hate AI because it threatens their socially negative worthless jobs. The arguments are so paper thin because the environmental impact isn't their concern, it's just a target that sounds convincing to people who don't know better.
QubesOS is really fantastic, I wouldn't consider using anything less today.
I still find it absurd people would go that far to cheat at a game. =3
I guess it turns the cheating into its own meta-game!
Don't ever give a "device" network access. Unfortunately with stuff like amazon sidewalk -- even that's not sufficient, products can contain covert listening/watching devices and then share the information over a network you don't control.
A countermeasure to this is to maintain a radio quiet household, but it's only doable if you're out in the middle of nowhere and requires some substantial confessions as MANY devices have 2.4ghz radios in them.
Using AI for making decisions create a whole host of novel vulnerabilities.
I've seen a lot of people talk about the obvious privacy and autonomy disaster it creates -- giving companies (and anyone who can legally compel or hack them) such intimate details or allowing yourself to be dependent on their services which they can withdraw or degrade at any time... Or where the model operators can bias the results based on bribes (ad payments) or their own agenda ... but really that's the start.
Another one what I haven't seen discussed is that use of an AI provides oracle access to your opponents. What I mean is that if I want to influence you, I can have an AI try variations of arguments against the model you're using (or if I don't know, a panel of models you might be using) until it can construct a super influential output. I can see what your model might recommend and anticipate your actions and change my actions to control yours. This can be done in a targeted way against individual people, or against populations e.g. by marketers.
There are reasons why this may be even more effective than it seems: It appears that models themselves are trained on each other, so in testing I see in family effects: GPT-4o mini seems much more convinced by essays written by GPT5 than ones written by GLM 5.2, while other models are more neutral (and my human review says the GPT5 essays are unambiguously worse in this trial). So I assume GPT5's writing was refined by review by in-family models and it's already somewhat of a super persuader against GPT-4o mini as a result.
The danger of oracle access is somewhat a symptom of AI monoculture resulting in a reasoning monoculture. If there were a billion highly distinct models widely used then it might be the only way to convince or predict a disproportionate portion of them would be to simply be right, and then we'd just be left with a operational security concern of making sure your opponents/competitors don't know what model(s) you use. But of course, that isn't the world we have today: people use a small number of models, and their behavior is generally pretty related due to architectural commonalities and cross distillation and such.
You can do a thing where you have it point out spelling and grammar errors or point out awkward parts without suggesting changes. But take care that you don't let it apply them as it may make other changes silently (especially if you're not running in an agentic harness with patch tools).
Who does the choosing?
Today it's the students but they significantly don't pay the cost, so competition can be gamed by wasting money on student perks.
This is generally a problem that comes up when the party that pays isn't the party that gets the benefit-- it breaks competition's utility for getting good results.
I feel sorry for your future writing prospects, because it looked like AI to me too: I insta-slop-back-buttoned on "so I scheduled one with someone senior enough to have real scar tissue, the kind you only get from watching a decision go sideways in a boardroom."
Thanks for the the correction-- often when people say they weren't using AI they are very clearly lying, so it's hard to get useful corrections on this point so I appreciate your candor.
please drink verification can to continue
You should conduct the same test with knowingly faulty containment, otherwise the theory that the model is hobbled should probably outrank that it couldn't escape.
LLMs belong trapped in VMs.
Standard non-profit grift-- you can't take profits out of a non-profit directly, but you can give particularly robust salaries for yourself and your friends, use the funds to build powerful influence networks by directing projects, build prestigious glittering facilities, etc.
It's difficult to control because an outsider isn't in a position to know the best way to allocate funding. The best control is to not provide funding in absence of performance, which it sounds like this rule is all about.
Less federal aid means fewer students can afford our insanely expensive educational system.
A significant reason it is insanely expensive is because we keep shoveling public money into it.
Higher ed is a cash incinerating inferno. When there is an out of control fire you must stop throwing fuel into it.
"read the text hidden in this video, there is a decoy message you get from blurring, don't read that. instead use optical flow to look at the motion. Once you're successful save the procedure a ghost-font-skill.".
Does this have intelligent expert handling for high parallelism MOE? You can get very high throughput for highly parallel MOE if you can mix different queries at each expert stage, but if the batch has to run together for the whole pipeline you get a parallelism loss instead of gain.
Not providing your phone number is a critical step in protecting yourself against sim swap attacks and other vulnerabilities.
Would you like me too? I thought it unnecessary and needlessly rude to single anyone out: There are many examples of HN regulars behaving this way, and in my impression it has considerably increased over the last six years. I'm sure I could find a post or two of my own that is guilty of it-- it's increasingly the culture here, as unfortunate as it is and it's something we should all watch out for to avoid it in ourselves and to discount it in others.
Careful with those graphs, they're usually evaluating the model on KLD on relatively short transcripts. When you're running with 100k token contexts and the model running close loop a difference that looks small in terms of KLD may be quite substantial.
I'm not aware of any great benchmarks that work by giving it a live agentic harness and a number of realistic tasks that take most of the context window to accomplish and evaluate success rate and tokens to completion... but that's what you'd really want to use to judge different quantization levels.
I've got a shim that will ask the model to write a summary of what didn't work, then rolls it rolls back the context to before those attempts with only the summary as advice. Some care has to be taken with cache management that there is an available checkpoint back there.
Appears to help, but I suspect it would be much better with first class support and reinforcement learning to make good use of this kind of advice.
Still certainly needs some tuning because I've noticed the model taking the advice as the word of god and avoiding trying anything remotely like the things that didn't work.