Hacker News

It’s been quite exciting that Qwen 3.8 Flash Next has come out: it really is similar to Opus 4.5, 4.6 for coding. Remarkably intelligent, and runs on a single DGX Spark, which I paid $4000 USD for

Sadly they’re now twice that price, which is a shame because I really want a second one!

I have a heretic modified version of it too, for when I want to use it for security and so on. Quite interesting

Gotta find some rich friends in order to buy and host a H200 or two.

A middle road is to use open-weight hosting providers, maybe non-US ones if you’re in the US.

Yes otherwise thoughts are crimes, effectively.

This is not even an option in the UK. Communication felonies (dangerous speech, threatening behaviour) only demand potential audience.

You can trust another party and do this by renting a few H200s. You cannot pool together with some friends without trusting another party.

You end up just weighing up the difference in trust between a vendor and a friend against the level of disinterest that they might have in your affairs.

Would be better if your friend is lawyer and take $1/yr(or higher) for the service so that attorney-client privilege can be applied.

edit: ah, future crime cannot be protected.

That's definitely something I'd consider if I had cash to spare for H200's! Unfortunately I think for most of us the price of self-hosting has to be 2-5x lower still.

Way ahead of you bud, mac studio m5 ultra 256gb version is coming soon.... ordered 2x of em just in case. expensive as fuck but its a hedge against all this bullshit and more

Lead time is like a year tho no?

I agree, but someone will say fake child porn.

I'm not saying fake child porn should be allowed or not-allowed, just showing there exist possible exceptions and rationalizations for them even without two parties.

>> I think Anthropic did the right thing here

This is the paradoxical times we live in right now.

Don't do something? She walks into the office and start shooting the place up. Several officers and innocent people are killed. Cue the media claiming, "You should've known she was talking about this an AI bot! Why didn't the bot tell anybody she was planning a mass shooting?!"

Do something? She gets rolled up by the cops and questioned about what she was talking about and brought to the cop station and interviewed. Cue the media claiming, "This is an unethical way to use AI, this is an infringement on free speech! This is authoritarian!"

I believe in free speech as much as the next person. But in this day and age, its almost better to be safe than to have to explain to someone's loved ones you had to chance to prevent this and did nothing.

It feels icky, and my default is not to side with megacorps engaged in blanket surveillance, but I can’t really fault Anthropic here. The correct setup should be a law that protects this sort of interaction with a chat bot as privileged, along the lines of HIPAA-mode. Use a classifier or some sort of “private mode” toggle to tell the platform you’re engaging in privileged communication ala dear diary, and then a much higher legal standard needs to apply to protecting that data (no training, same protections as doctor / psychologist interactions). Even a therapist has a legal duty-to-report in certain situations.

Would you use an AI service knowing that they are inclined to turn you in to police if they detect illegal activity?

In a "three felonies a day" universe?

You omitted the part in the diary entry about shooting up sheriff’s office

Not just bad headlines; OpenAI is actively being sued [1] for this:

    OpenAI’s stated rationale was a concern for the shooter’s privacy, but its own interests
    better explain its silence. Upon information and belief, OpenAI was seeking to avoid implementation of a
    hard line rule to refer planning of real-world violence to authorities, perhaps due to how frequently its
    product is implicated in threats to human life. Requiring such disclosures would be incompatible with the
    company’s public position that ChatGPT is safe. It could also threaten the valuation underlying OpenAI’s
    anticipated initial public offering. Rather than expose those risks, OpenAI accepted the consequences of
    its silence. A mass murder in the only secondary school in Tumbler Ridge followed.

    Accordingly, the Crown, led by Attorney General Sharma, and SD59 jointly bring this
    action to hold OpenAI and Sam Altman accountable for designing a dangerous product, distributing it to
    every home with internet access, ignoring the warnings of their own safety team,
    refusing to notify authorities when they knew the shooter was planning gun violence, inviting the shooter back onto the
    platform after deactivating the shooter’s account, and choosing corporate self-interest over the lives of
    children. They seek compensation for the not just foreseeable but known harm OpenAI inflicted, the
    damages that they incurred and are incurring, and injunctive relief to ensure that this tragedy does not
    happen again.

I'm still puzzled that people's default assumption isn't that somebody is reading all of their internet communications. Ever since the Snowden revelations of 2013 I've more or less assumed that everything I type into a computer is stored in a government database somewhere. Not that I actually believe it is 100% of the time; there's a spectrum of trust, so I'm more confident that local apps on my Linux desktop are secure, somewhat confident in the end-to-end encryption of certain apps on my iPhone, but all bets are off for non-E2E encrypted data going across the internet.

This is one of the reasons why AI companies are looking for explicit regulations, it can help to reduce the risk of possible liability and maybe make the legal way forward more tractable. With a regulatory framework in place much of the burden of identifying risks falls on the regulatory authority.

Then the AI companies have more confidence that they can move forward in a certain way, and issue investor guidance that is maybe closer to reality.

> people need to get it in their heads that they're not chatting with their secret BFF

I see constant ads on video platform (particularly youtube/tiktok) about llm chat apps, from friends, dating, romance and everythkng inbetween; that's personal.

People need to be reminded constantly if they use such apps that they are participating in easier mass surveillance, profiling and AI training.

> people need to get it in their heads that they're not chatting with their secret BFF, they're chatting with Big Tech

Yup. And with zero privacy protections in statute for AI chat, there is nothing to prevent an AI CEO looking to curry political favour from e.g. handing over the private correspondence of an opponent or an entire district’s residents.

A disclaimer on the top of page every time would have been a better approach helping Anthropic and the end user.

A bot talking to you directly as if its some one real caters to your thoughts and can take you in a certain direction without you realizing it. I have heard first hand experience from people that they feel more comfortable talking to chatgpt or claude cause it gives a feeling of being on their side and listening to them.

And such was the case for a man who snapped a photo of his own child to send to the doctor which got uploaded to his Google photos resulting in his Google account of over a decade getting shutdown for CSAM.

As another commenter said, you're not chatting with a friend; you're chatting with Big Tech.

How much do you love Big Brother?

I suspect it will go further: imagine giving an mp3 to LLM to clean up some noise. It detects it was illegally downloaded from youtube, deletes it and automatically fines you via attached credit card.

You should probably get a head start on waiting a couple years to bite your tongue and assume everything you type into a computer is summarized and sent to your boss, government, advertisers, political actors, insurance companies, worst enemy, etc. With phones, Alexas, and little AI tamagotchis, you probably shouldn't say much in person either.

They're actively rummaging through your inputs so the damned-if-you-don't case doesn't really exist; no one expects Anthropic to not notify law enforcement once they learn of something like this. What you might have expected was some privacy in the first place though, where Anthropic would never have learned of this in the first place and where the damned-if-you-do case wasn't a thing.

it was in the article

> people need to get it in their heads that they're not chatting with their secret BFF, they're chatting with Big Tech

It's the other way around, big techs need to properly disclose in their platform, during interaction that they aren't in a private and safe environment

Right, the difficulty is partly that they can get a negative headline from any choice of behavior.

"Anthropic failed to report murderer's threats to authorities"

(or "Chatbot knew man was planning murder, yet company did nothing")

"Anthropic reported private chats to authorities"

(or "Arrested for chatbot fantasy")

To be fair to the journalists in these cases, there's also no society-wide agreed Schelling point about the correct outcome or correct rules. I have strong beliefs and intuitions about what should happen, but other people also have strong beliefs and intuitions, and many of those are probably opposite of mine. Even if my intuitions are the best and most justified, a journalist is unlikely to think "I'm just not going to mention that some people are mad at this company over this outcome, because a hypothetically better norm or principle would support the company's actions here". Hopefully the journalism can at least contextualize the lack of legal or social consensus and the difficult incentive problems, rather than jumping to "obviously companies are sociopaths staffed by supervillains".

> in private chats.

Including any chats anywhere where someone might have a phone in their pocket, or if there's a "camera" attached to a utility pole or a nearby tree. The only real private chats might be whispered lying down in the bathtub together, with a mattress covering it like you're both hiding from a hurricane.

> they're chatting with Big Tech

They're chatting with any powerful person who wants to hear it. She thought she was chatting with Anthropic, who doesn't give a shit about her. But after being threatened (and immediately backing down because, of course, they don't give a shit about her) Anthropic has become an arm of the government. So she was chatting with the Bonita Springs, FL Sheriff's office, or anybody else. If I paid enough, Anthropic would tell me about what she was doing so I could sell her laundry detergent.

It's a byproduct of the nannyism safety marketing from the AI companies. I'm glad these cases were caught, but disagree with how they were disposed of. If the automated flagging is good, enforce it by default. If it's noisy, refine the tech then enforce it by default. This middleground where everything going through the platforms is subject to training and arbitrary human inspection in the midst of an acrid cloud of marketing-driven fearmongering is unacceptable, and it reinforces the idea the fearmongering is legitimate.

Somehow humanity survived the past 40 years without Microsoft Word and Excel phoning home and shopping users to the feds at random, I don't see why the standard should be any different for this new class of tooling.

Absolutely no sympathy. Anthropic is every bit as slimy as any other big corp. And like other big corps, they must open the vault to whatever governments they intent to do biz with.

We really need a class, probably in high school, that works through how LLMs work at the high level (don't need to get too far into the deep math, but give people a taste) and then how they're trained, used, and deployed.

I feel like if people understood what these things actually are there'd be way less of this AI psychosis and similar stuff.

There'd also be fewer people falling for apocalyptic Rationalist delusions.

Also: people need to understand "not your computer, not your data." (Unless it's stored in the cloud but encrypted locally with keys only you possess.) Same goes for storing things unencrypted in OneDrive, Google Drive, etc. There is nothing to stop these companies from bulk scanning, data mining, or reporting people based on whatever request a government gives them. Don't count on them to resist, because they often can't, especially if the request is from a sovereign state where they do business.

> I have some sympathy for Anthropic here [...] So from their perspective, it's damned-if-you-don't, damned-if-you-do.

On the other hand, they did put themselves into this position deliberately.

> when you only read the threat by spying on them

The AI companies have clauses in their user agreements saying they can review content flagged as harmful. It’s not legally spying.

If you recall previous outrage about ChatGPT being used in cases of suicides or shootings, this is the result. Every time a crime was committed and the police found ChatGPT history about the crime, the media turned it into a frenzy. So the AI labs added safety filters to their consumer plans that detect threats of violence, escalate them to human review, and report to the police.

Spying is not the right analogy because the information was given to the police by a third party which had a EULA saying they would do this. A more analogous situation would be someone reading another person’s diary and then turning it into the police department. There might be some limitation in the law that makes the evidence inadmissible because it was not intended to be shared with anyone, but that’s a separate decision.

This seems to be the statute: https://www.leg.state.fl.us/Statutes/index.cfm?App_mode=Disp...

With the obvious IANAL, it doesn't seem to rely on the message be sent to the person being threatened. The specific segment is "in any manner in which it may be viewed by another person".

This may be one of those cases where we get to find out how courts view SaaS platforms.

Exactly, can you threaten someone without them receiving the threat? If this is not thrown out, Minority Report will actually happen.

this is almost certainly what anthropic is hoping for here - a judgement that says there is no point in them continuing to monitor and report this behaviour

We are snowballing to Minority Report ...

> Surely that has to be thrown out in court?

The prosecutors likely know this and expect it. But there's enough gray area here for them to make the argument, and it's hard to prove malicious prosecution, so they know they'll get away with it. It's just about sending a message to the public - they don't care whether a conviction sticks. Just politics.

It's not quite spying when you willingly hand over this information and agree to terms of service. You're data is not considered yours alone.

the argument to be made is that allowing anthropic to see it constitutes sending the threat.

sandbox your ai.

It's legal to spy if the terms of service say so

I was thinking the same. If the evidence was not obtained with a proper court order wouldn’t this result in a mistrial?

what if it's just testing the AI to see how it responds

A threat sent by mail is still a sent threat even if nobody ever opens the envelope to read it. The crime is in the sending. This woman used an online resources, one which involves transmitting everything across innumerable state lines. I am surprised she isn't up on federal charges.

Note that the law doesn't forbid the writing of a threat. You have to send it to someone. Had she kept it in a book under her bed, she would not be in trouble. But she sent it to a website/service/LLM portal.

>> It is unlawful for any person to send, post, or transmit, or procure the sending, posting, or transmission of, a writing or other record, including an electronic record, in any manner in which it may be viewed by another person

Don’t worry, they’ll store those messages forever and incarcerate you at their convenience.

I was sure I couldn't be the only one curious to push LLMs to their limits. Though these days it's much tougher, mostly impossible to get them to react in unforeseen ways to horrendous scenarios.

This is exactly the typical use I make of the llm.

Adding: - I typically ask questions in the I form, regardless for whom or why I ask for. - Gemini chats quite often end when it starts recommending psychological council or a suicide line, to talk about my problems. It apparently detects a persistent tendency to not agree with the party line. So it makes sense I must be suicidal ;-

But sure, as llm's start to babysit us, and know our inner dialog better than anyone else, we'll soon be debugging their opinion/behavior/co-existence/authority, when it comes to reporting people to the authorities, or taking on tasks in society in general. We'll hire doctors to cure our psychological profile from our record (Total Recall).

A Minority Report like this shouldn't cause a referral to the police.

I think it's fair to say this is a gray area. Clearly it was transmitted.

I can certainly threaten you harm and send it to not-you and you're still clearly in danger even if it wasnt transmitted to you. So the question becomes did she transmit it to someone? Clearly yes she transmitted it to Anthropic. But she clearly intended to send it to Claude, an inanimate object.

Yeah.. if you write a personal note and it's backed up by the operating system, it appears to be in violation of this law as well (since the company could theoretically read it)

This almost smells like thought crime... Minority Report when?

>I guess we probably want our tech to work exactly like this.

Do we, though? What if someone started an AI company that uses end-to-end encryption to make it impossible for anyone but you to access your data? Personally, I would switch to it in a heartbeat assuming it's competitive with the other products. I don't think it's the tech companies' job to surveil the population and prevent crimes. That said, I'm not necessarily against Anthropic or other companies reporting suspicious activity if their existing systems are detecting it. I'm just not sure we want every product to be forced into that data model.

Your follow-up about purely private tech seems to contradict your first statement. We can either have privacy or surveillance, not both.

That sounds adjacent to "I have nothing to hide". Everyone says that until they realize someone can change the rules. Before 2022, women didn't have anything to hide from their period-tracker app, now some have to worry about being charged with crimes.

Back to LLM chats: A system that can declare her "unstable" is also one that can permaban you from all air-travel because you "privately" said unflattering things about Dear Leader.

> It should catch normal people becoming unstable so that they can receive help. It is just highly unfortunate that the US legal system works in ways where now this woman's name is public.

So... minority report?

The third highest voted comment wants surveillance. What is next? If you write "I'm going to kill that guy", which for non-autists means "that guy was really annoying" you should be reported.

the honest actual opinion of the average consumer is probably that they don't want their chats to be monitored but they want everyone else's chats to be monitored. There was a lot of fury when a mass shooter recently used AI to help plan his assault. And of course there are all the people talking to chatbots about suicidal thoughts and intent. When they ultimately follow through, the providers are blamed for not alerting anyone.

No. This sentiment is why Snowden happened. We want privacy. Privacy isnt free just like freedom (whatever form it is) isnt free.

There are trade offs. ISP effectively is like driving on a highway, everyone can see where you are going but not what is inside the car. Id like these AI chats to be the same but they are not.

>I guess we probably want our tech to work exactly like this.

Maybe you do; I want my tech to always include secure, encrypted communications. Don't include me in your destruction of privacy with your silly bandwagon!

uh no, 'we' absolutely do not want our tech to work exactly like this. Why would you assume that people default to 'search my anal cavity please' and not 'no, stay the fuck away from me with that glove'

----

jfc, why is this website full on psychopaths

"The actual fundamental underlying problem being that not all people are equal, but we kinda have to pretend they are, because not doing so leads to fascism and other terrible stuff.

But we kinda also do not want to fully pretend that, because doing so leads to yet other terrible stuff."

other people are not pretending everyone is equal. your power levels are showing, it isn't subtle.

> in which it may be viewed by another person

To me, that is the more interesting legal question. Does a LLM-based safety net that sends content to a human, when the original use case would not have sent it to a human, count as "may be viewed by another person". It certainly wasn't intended to be, and that isn't the norm. At the same time, because no security is perfect, we could say that any digital record, stored in any way "may be viewed by another person."

Something for the courts to sort out, of course.

Except if you're one of the "warfighters" that Anthropic supports

Is the LLM now "another person"?

It is my very european belief that the problem here is not that the woman was reported, but that what likely is a mental episode was made public in a way that reduces the chances of recovery.

The article says she made the comments on September 26.

It doesn't really matter if they "should or shouldn't" have an expectation of privacy IMO, they already do have that expectation.

> If they were to offer total privacy, is it ok for the public to use chat to get advice on _how_ to commit a crime?

Yes it should be, but it should be illegal for a company providing chat service to respond with anything other than a refusal when doing so.

That detection and refusal should be a private closed loop though, anonymizing any data that will be passed into a training pipeline, or ads targeting. This requirement for closed loop private chats should be mandated by law sooner than later. Otherwise we're getting into very tricky territory where the temptation of alerting on things like pre-crime grows too close.

Won’t help if your model is trained to immediately find a way to push out a message somewhere to alert authorities of your threat

I once had a copy of 1984 in my checked baggage returning home to the USA and when I was unpacking the bag at home, the book had a notice inside the book that my bags had been inspected by TSA...

In Europe they should provide the citizens with the service of their messages not reaching US servers. So basically that there is only one party reading along with them, not half the world.

> For the sake of argument what would happen if she had kept the diary locally and claude code scanned the file?

For the sake of even more argument, imagine if she was writing her thoughts with a pencil, on a good old fashioned paper diary, and she had a phone nearby and the phone took a picture of her diary, OCR'd the words, and reported it to the police?

Both of those scenarios would demonstrate even more commitment to safety so I imagine they’d be at least as likely to happen as this, if not more.

You only accept this because you are part of a low trust society. I wonder if tech has done anything to improve trust among people or just lowered it.

https://www.npr.org/2026/07/10/nx-s1-5886113/waymo-police-pr...

Two teens riding in a Waymo were arrested because the AI detected them talking about having a gun.

Yes

She is not being charged with planning, just making threats in a place a person could read - the person being employees of the company.

Basically, under this interpretation, any personal note you store in the servers of a company could qualify, even if you didn't ever imagine someone would read and as such you couldn't have thought about it as a threat

LLM is not a person but T&C probably states that a human moderator may view any of it. Her lawyer could probably argue a moderator filtering usage isn't the intent of the law, it was more about publishing / sending message for other humans and it was never clear to her that a human was reviewing her private diary. Shouldn't LLM disclose that at some point when people are feeding really personal stuff?

Personally, I’ve never seen anyone claim this. At least anyone who one would think is informed on these matters.

I haven't seen anybody claiming this for consumer/free accounts anywhere.

You get privacy if you're a big corporation that needs to make sure OpenAI/Google/Anthropic can't read your trade secrets etc.

But those contractual privacy protections have been in place for a long time. It doesn't have anything to do with AI, it's been the same with Office365, Google Docs, etc.

Who claimed this? I've never seen that claim.

No sane person claimed this.

Uhm, those people never existed? Or maybe you just have very interesting friends?

LocalLLM enthusiasts exist for a good reason.

They still read every single message you send and train on them.

I’ve had them email me before because I was testing it as a filter for abusive messages and they detected some no-no and wrongthink in those test messages.

I looked into ZDR, it's basically a vapid claim with little to no due diligence or auditing. I expect most providers to cave with only a minimal amount of legal pressure.

How much spend do you need to negotiate ZDR?

Pre-crime prevention and detention will be the end goal. The economic devastation caused by AI will make large swaths of people get angry. Better use that massive amounts of data being hovered up and the mass of compute being built up to stop those angry people lomg before they even get a chance to properly organize.

For that to happen, someone would need to be spying on all cops.

Auditing logs is a crazy interpretation of "another person may view it"

Should CEOs of Antrophic and OpenAI be reported too then? They cannot stop saying that AI will destroy humans.

this has nothing to do with claude. “thought” is fine. “written and shared” is illegal all over

https://www.nbcmiami.com/news/local/everyone-deserves-to-die...

https://www.wdsu.com/article/maryland-high-school-student-ch...

https://www.pinellassheriff.gov/21-023-deputies-arrest-pinel...