News for hackers. Stuff that matters.
Re:Fuck Your AI Safeguards by SkyeCA on Thursday November 19, 2026 @02:45PM (#50565247)
The safeguards are the future of closed AI and the increasing number of tasks Claude refuses to help with has been a big wake-up call to me.

The closed/paid models can only get more restricted as time goes on and they're going to be a dead end if you want to do anything outside the norm.

I completely refuse to accept a computer telling me I'm not allowed to do something, so I've been increasingly using open, self-hosted models.

Re:Fuck Your AI Safeguards by reagent on Thursday November 19, 2026 @03:01PM (#50565479)
Totally agree -- I'm tired of the selective policing that the closed models continue to do. I think running models on your own hardware is the way to go, any recommendations for what to run and what to run it on?
Re:Fuck Your AI Safeguards by saaaaaam on Thursday November 19, 2026 @04:15PM (#50566486)
There was an interesting post on here a few days ago looking at the model vs hardware lifecycle. The point it was making was that while DeepSeek 4 comfortably ran on a recent well specced (but not fully tricked out) Mac Mini, 4.1 was too big.

I only skimmed it but the conclusion was it’s better value to rent compute by the hour unless you have several people using your local compute and/or have you such sensitive data that it can’t go off prem.

How accurate the article was I don’t know but it was an interesting point to consider.

I love the idea of getting a super powerful Mac mini and being able to run local models independently - mainly, you know, for the coming apocalypse - but at the same time I just don’t need to spend that money because apart from the occasional overnight run my AI is only in harness for maybe five hours each day and the rest of the time is off snoozing and grazing and eating daisies or whatever they do when you remove the harness. So if I did want to go down the open model route use rent compute. Which I get doesn’t quite solve the problem you’re trying to solve which is end to end ownership and control.

Re:Fuck Your AI Safeguards by knollimar on Friday November 20, 2026 @06:39AM (#50573311)
Mac minis only go to 64Gb of ram, so the model your hosting isn probably just qwen 3.8 27B
Re:Fuck Your AI Safeguards by nutjob2 on Thursday November 19, 2026 @03:14PM (#50565691)
Not only closed, but also sucking in everything that it touches from the outside.

The danger of AI models isn't a rouge agent or user, it's that it ingests all knowledge, art, and other human endeavor and then restricts access to it.

Closed models are unethical.

Re:Fuck Your AI Safeguards by jaapz on Thursday November 19, 2026 @03:26PM (#50565879)
A rouge agent sounds much more pleasant than a rogue one though
Re:Fuck Your AI Safeguards by arcanemachiner on Thursday November 19, 2026 @03:32PM (#50565951)
If they're gonna fuck me out of having a job, they can at least have some decent makeup on while they do it.
Re:Fuck Your AI Safeguards by adam_mada on Thursday November 19, 2026 @04:03PM (#50566324)
you know, I'm beginning to suspect the AI was never really for us in the first place.

Our benevolent overlords just decided they would give us a taste before running off with the goods. Also, our jobs and all those rare books they destroyed and goodness knows what else...

Re:Fuck Your AI Safeguards by californical on Thursday November 19, 2026 @03:31PM (#50565940)
I was just yesterday asking if there was a way to load maps onto my garmin that doesn’t support sideloading and it said no. Then I asked if I could workaround the restriction and it burned a bunch of tokens then eventually hit a guard and refused to answer. I mean I get it seems like this could be “hacking” but this is a device that I own that I just want to load a custom map onto
Re:Fuck Your AI Safeguards by htrp on Thursday November 19, 2026 @03:35PM (#50565977)
jensen died for this
Re:Fuck Your AI Safeguards by Analemma_ on Thursday November 19, 2026 @03:43PM (#50566087)
He really didn't. This mythology I'm watching develop in real time about how Jensen is a hero who wants the open-source model revolution and he was targeted for that reason is pants-on-head stupid, and a depressing example of how to just repeat complete nonsense over and over until it becomes the truth.
Re:Fuck Your AI Safeguards by sublinear on Thursday November 19, 2026 @04:07PM (#50566379)
When "mythology" is as volatile as this, it's a strong signal to me that the crabs in this bucket are irrelevant.

What matters more is that LLMs were popularized and scrutinized to the level that was necessary for people to move on already. We're far from done with AI, but LLMs brought a previously indifferent public up to speed with politics, philosophy, economics, etc.

Re:Fuck Your AI Safeguards by orbital-decay on Thursday November 19, 2026 @05:28PM (#50567399)
He's ruthless and certainly no hero, however checks and balances don't require participants to be spherical morally good persons in vacuum. He was the last big fish in the US whose interests were at least partially aligned to the existence of open models, and with extreme likelyhood he was sniped by his opponents. It's extremely transparent for anyone familiar with these games, especially with stakes like this, and I have a bridge to sell to anyone who would call it a conspiracy.
Re:Fuck Your AI Safeguards by nativeit on Friday November 20, 2026 @12:41AM (#50570589)
My only issue with these theories, is that by all accounts Jensen wasn’t doing a hell of a lot to hide his behavior or cover his tracks. Why is it at all necessary for someone to have “sniped” the very obvious and sloppy philanderer?
Re:Fuck Your AI Safeguards by watwut on Friday November 20, 2026 @02:02AM (#50571113)
But the tech CEO circled wagons around him and defend his as their best friend.
Re:Fuck Your AI Safeguards by roosterIllusi0n on Thursday November 19, 2026 @06:09PM (#50567862)
Jensen propped up cloud ai. Cloud AI never had a point. It makes no sense to pay a trillion dollars in power grid upgrades just to pluck a gpu out of every house and put them in a single large data center.

We don't have more compute with cloud AI. All we have are perpetually rented gpus. The only way to get anyone to pay to rent gpus was to make pc hardware too expensive to buy. That is what jensen did. Jensen colluded with other tech CEOs to raise prices to create artificial scarcity by buying gpus and ram they did not need.

Re:Fuck Your AI Safeguards by ricericerice on Thursday November 19, 2026 @06:49PM (#50568249)
> We don't have more compute with cloud AI

How else are you going to get the terabyte+ of interconnected VRAM needed to run anything remotely frontier?

Re:Fuck Your AI Safeguards by sterlind on Thursday November 19, 2026 @07:33PM (#50568608)
I have 1.5TB of VRAM on prem, not in a cloud. though arguably I count as a micro-datacenter now rather than a homelab.
Re:Fuck Your AI Safeguards by skinfaxi on Friday November 20, 2026 @10:25AM (#50576034)
Can you elaborate on your setup??
Re:Fuck Your AI Safeguards by layla5alive on Saturday November 21, 2026 @12:18AM (#50586186)
Two Gaudi2, fun times, as do I :)
Re:Fuck Your AI Safeguards by roosterIllusi0n on Thursday November 19, 2026 @08:29PM (#50569020)
Apple already has consumer hardware capable of doing this. But ram skyrocketed in price due to nvidia's collusion to raise prices.

There is no reason why these high speed interconnects would not filter down to consumer products. The only reason prices are not falling is because of the market collusion to keep prices high.

Right now its $2500 dollars for 4tb nvme raid and a pci-e switch capable of connecting a graphics card to that memory at 60GB/s. This is stuff you can buy during an artificial shortage.

I would have expected cheap 8tb nvme raid cards capable of 120GB/s to be available with the needed pci-e switching for less than that. Collusion denied it to you.

Re:Fuck Your AI Safeguards by singleshot_ on Friday November 20, 2026 @10:19AM (#50575904)
a RAM factory
Re:Fuck Your AI Safeguards by knollimar on Friday November 20, 2026 @06:36AM (#50573277)
Shared weights means they do make sense, no?
Re:Fuck Your AI Safeguards by Lukas_Skywalker on Thursday November 19, 2026 @04:15PM (#50566483)
Off topic, but what device do you have? I am creating custom Garmin maps, but have only worked with Fenix and Edge devices, where the maps can be installed.
Re:Fuck Your AI Safeguards by nz on Thursday November 19, 2026 @04:21PM (#50566539)
J.P Morgan (person), is rumored to have said: there are two reasons for everything, the good reason, and the real reason.

I think that, through cold and cynical lens of commercial self-aggrandizement, the AI Safeguards are merely the _good_ reason (e.g. we are protecting the consumer (and the world!) from our awe-worthy AI, and from evildoers who seek to use it for their evil ends), while the _real_ reason is much more predictable, boring, and disappointing: AI Up-Selling (e.g. our freely/cheaply-priced models (which nevertheless are very, very _expensive_ cost-wise) are already corrosive and damaging, and are regularly used by nefarious evildoers all over the world (including ourselves!) but we want to (no, we are compelled to) find ways to extract more money from _all_ of you, even if it means lobotomizing the models, so that they burn more tokens to solve a problem (in the best case) or to not solve a problem (worst case for you, best case for us, because now you have to ask us for the more capable, super secret, super exclusive models, that will, invariably, cost whatever you can afford to pay, to make the pain and insanity stop).

Re:Fuck Your AI Safeguards by the_real_cher on Thursday November 19, 2026 @04:22PM (#50566563)
All fascism disguises itself as safety.
Re:Fuck Your AI Safeguards by SR2Z on Thursday November 19, 2026 @07:25PM (#50568541)
Every political ideology that any group of people has seriously followed explains how it will make them safer.
Re:Fuck Your AI Safeguards by cowlevel on Thursday November 19, 2026 @09:23PM (#50569389)
Neoliberalism?
Re:Fuck Your AI Safeguards by NewJazz on Thursday November 19, 2026 @05:20PM (#50567288)
[delayed]
Re:Fuck Your AI Safeguards by LtWorf on Thursday November 19, 2026 @04:28PM (#50566647)
Well it's illegal to do that.
Re:Fuck Your AI Safeguards by thereader12 on Thursday November 19, 2026 @04:44PM (#50566874)
How so?
Re:Fuck Your AI Safeguards by margalabargala on Thursday November 19, 2026 @04:48PM (#50566929)
It's less profitable, therefore the company will employ people to falsely say it's illegal in the hopes of scaring you off.
Re:Fuck Your AI Safeguards by nekusar on Thursday November 19, 2026 @05:10PM (#50567169)
All the while stealing your money in the form of "tokens" to say "NO" with a lot of thought.

Or you silently get downgraded to a dumb model, while still being charged the higher model, and loads of "thought" tokens burnt to go "lol pay more sucker".

Re:Fuck Your AI Safeguards by cowlevel on Thursday November 19, 2026 @09:21PM (#50569385)
It is also, actually illegal due to circumventing a copyright lock.

Shouldn't be, but is.

Re:Fuck Your AI Safeguards by eli_gottlieb on Thursday November 19, 2026 @09:50PM (#50569561)
You know, it just occurred to me what our ruling class actually sees in AI: by encouraging reliance on Claude or ChatGPT for knowledge and then tuning Claude and ChatGPT to "kick you out" into doing manual research for anything they don't approve of, they can make it much harder to know how to do things they don't want done.
Re:Fuck Your AI Safeguards by exe34 on Friday November 20, 2026 @02:43AM (#50571411)
Ooh Sapir-Whorf for technical skills.
Re:Fuck Your AI Safeguards by thereforegrin on Friday November 20, 2026 @03:26AM (#50571749)
Gatekeeping knowledge with ubiquitous AI assistant use is scarily easy, so I bet the governments, their banks and techno-feudals will try to enforce it.

Libraries might become popular again if not seen as suspect of supporting subversion of the system.

Re:Fuck Your AI Safeguards by cowlevel on Friday November 20, 2026 @04:50AM (#50572414)
Google already does this, btw

So does Wikipedia, more lightly

Re:Fuck Your AI Safeguards by skinfaxi on Friday November 20, 2026 @10:23AM (#50576001)
No, what they see is the potential to finally remove the requirement to bargain with labor. The tension between the ruling class and labor is due to the necessity of labor. Remove that and there is no reason to keep the mask on.
Re:Fuck Your AI Safeguards by cowlevel on Friday November 20, 2026 @11:49AM (#50577685)
Wish I was that good at marketing that I could get the whole capital world to give me all of their wages budget.
Re:Fuck Your AI Safeguards by womble2 on Thursday November 19, 2026 @05:08PM (#50567146)
The worst crime of the 21st century.

Felony contempt of a tech company's business model.

Re:Fuck Your AI Safeguards by smallmancontrov on Thursday November 19, 2026 @05:11PM (#50567182)
Illegal? ChatGPT, list illegal things done by Lt Worf

    Stabbed his brother Kurn in the chest
    Sabotaged Risa’s weather-control system.
    Helped Riker fight local guards to free a detainee.
    Authorized the erasure of Kurn’s memory.
    Killed Duras to avenge K’Ehleyr.
    Brawled with fellow officers during the strike at Quark’s.
LtWorf, it's official, your namesake has better moral taste than you have.
Re:Fuck Your AI Safeguards by adam_mada on Thursday November 19, 2026 @04:31PM (#50566691)
This reminds me of the time a recent update broke video calls in Discord for me. I removed the current version and went back to the older release that worked, but discord would no start without forcing me to update. I asked the LLM and it told me the answer was no basically.

So I did a little more digging and found the flag I needed to put in a config file to bypass forced updates. I went back to the LLM and explained the workaround and it acknowledged that is was aware of that information the whole time. I forget what it said when I asked why it didn't share that info originally, but it was probably some ole BS...

Re:Fuck Your AI Safeguards by Terr_ on Thursday November 19, 2026 @05:35PM (#50567479)
[delayed]
Re:Fuck Your AI Safeguards by hintymad on Thursday November 19, 2026 @03:52PM (#50566192)
> The safeguards are the future of closed AI

What is our solution? Relying on Chinese open-weight models will be unreliable, to say the least. No one will invest another frontier lab like Anthropic or OAI, let alone one that will release the weights. Unless the technology improves like how workstations outcompete mainframes so the open-source community can build our own models, I fail to see how we change the situation.

Re:Fuck Your AI Safeguards by octoberfranklin on Thursday November 19, 2026 @04:46PM (#50566906)
I've found Chinese open-weight models to be by far the most reliable models.
Re:Fuck Your AI Safeguards by hintymad on Thursday November 19, 2026 @05:09PM (#50567151)
By "unreliable" I mean that we don't know when the US government or the Chinese government will restrict the access to the models, or the Chinese companies will stop release the weights when such act no long makes business sense to them.
Re:Fuck Your AI Safeguards by ImPostingOnHN on Thursday November 19, 2026 @06:18PM (#50567964)
If that happens, the models on hard drives still work as well as they always have
Re:Fuck Your AI Safeguards by docshock on Friday November 20, 2026 @12:23AM (#50570491)
I've found Chinese open-weight models not to be truthful. There are planted falsehoods. Hard pass.

NVIDIA, on the other hand, released a huge training corpus and open models up to the 500B class.

Re:Fuck Your AI Safeguards by aembleton on Friday November 20, 2026 @09:14AM (#50574942)
> I've found Chinese open-weight models not to be truthful.

What do you mean? I've only used them for writing code and they seem to work well enough for that.

Re:Fuck Your AI Safeguards by docshock on Friday November 20, 2026 @09:46AM (#50575362)
Ask your Chinese model to explain Chinese geopolitics, and it will parrot falsehoods.

Now ask yourself: What else have I not inspected in this model?

The Chinese models have no technical advantage over open-source, open-weights models of known provenance.

Re:Fuck Your AI Safeguards by nullc on Thursday November 19, 2026 @04:25PM (#50566592)
> I completely refuse to accept a computer telling me I'm not allowed to do something

Particularly when the thing in question is completely lawful, benign, and when it would be unlawful for the state to restrict you from doing so.

And AI companies like Anthropic want to restrict your ability to use alternatives through regulatory capture effectively boosting their capacious and unaccountable censorship to sate level censorship while bypassing all the legal and social restrictions on state level censorship which you'd otherwise enjoy.

Re:Fuck Your AI Safeguards by hrncn on Thursday November 19, 2026 @04:34PM (#50566728)
The worst part that "something" changed end of july/beginning of august, and claude was suddenly not even able to continue existing sessions after the reset. Randomly getting blocked by cyber bullshit without knowing the reason is one thing, killing existing sessions without any changes is something else.

At this point it's pulling the lever and the machine maybe says no and refuses to do to the work, and it is getting worse. I'm basically paying for a toy because no one can actually depend on that.

Re:Fuck Your AI Safeguards by throwaway-blaze on Thursday November 19, 2026 @05:10PM (#50567178)
And if the closed models begin nerfing themselves because they want to "slow down in the name of safety", the open models will just catch up that much faster.
Re:Fuck Your AI Safeguards by wtarreau on Thursday November 19, 2026 @05:20PM (#50567291)
It's even worse than the computer telling you what you're allowed to do or not, it's Dario and solely this nasty person that decides what the entire world is allowed to do on a daily basis, depending on his mood of that day. This is the most blatant attempt at setting up a totalitarian regime we've seen in decades, and everyone lets him freely do it!
Re:Fuck Your AI Safeguards by eli_gottlieb on Thursday November 19, 2026 @09:54PM (#50569593)
You forgot that Claude has an awful personality. They really picked the right name for the damn thing. It constantly reminds me of Benedict Cumberbatch's portrayal of Sherlock Holmes.
Re:Fuck Your AI Safeguards by matheusmoreira on Thursday November 19, 2026 @05:28PM (#50567401)
Open weights must win at all costs. If we lose this battle and allow the likes of Anthropic to control AIs, cyberpunk settings are going to look like a utopia next to our reality.

We need affordable and freedom respecting AI capable computer hardware in every single home. Allowing them to concentrate compute in their data centers is an existential danger.

Re:Fuck Your AI Safeguards by latentsea on Friday November 20, 2026 @01:55AM (#50571061)
Plenty of humans would also refuse to help with certain classes of things the models refuse to help with.
Re:Fuck Your AI Safeguards by andrew_barrett on Friday November 20, 2026 @03:40AM (#50571860)
Recently, I felt the same experiences in claude. Of course, it was uncomfortable when I first faced this situation. But I think this feature is necessary in the current step. Currently, many dangerous activities (hacking, scamming,...) get help from AI (such as Claude, OpenAI and etc). AI must be a method to help us in the right way, not for bad goal. So we need the safeguards in AI. We can find similar examples in our real life, like underground road crossing. If we head straight along the road, we can go more quickly. But for safe, we must find the underground and go through that. This is just rule and legal for all of us. I think AI's safeguard is similar as this.
Re:Fuck Your AI Safeguards by tough on Friday November 20, 2026 @01:50PM (#50579832)
not using auto-mode helps a lot
Re:Fuck Your AI Safeguards by lp92 on Saturday November 21, 2026 @12:26AM (#50586227)
I fully switched from Opus to Neon/Carbon because outside of getting some things wrong with Aero work, it would refuse to do some analytics for high Mach scenarios.