Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Claude Opus 4 and 4.1 can now end a rare subset of conversations

anthropic.com

Post reply on HN
Print view View original

331–340 of 453 posts

Re: Claude Opus 4 and 4.1 can now end a rare subset of conversations

#331

Post by int_19h »

clwg wrote: post #270
debazel wrote: post #201

Earlier quoted context omitted.

> We have privacy laws and safeguards on all those things Which have failed horrendously. If you really just wanted to protect kids then make kid safe devices that automatically identify themselves as such when accessing websites/apps/etc, and then make them required for anyone underage. Tying your whole digital identity and access into a single government controlled entity is just way too juicy of a target to not ge…

> Which have failed horrendously. I'm Canadian, so I can't speak for other countries, but I have worked on the security of some of our centralized health networks and with the Office of the Privacy Commissioner of Canada. I'm not aware of anything that could be considered a horrendous failure of these systems or institutions. A digital ID could actually make them more secure. I also think giving kids devices that ide…

If you're Canadian, then you don't have much in terms of legal safeguards to begin with, given the notwithstanding clause of your constitution.

Re: Claude Opus 4 and 4.1 can now end a rare subset of conversations

#332

Post by antonvs »

perihelions wrote: post #313
postalcoder wrote: post #281

Earlier quoted context omitted.

> There's not a good reason to do this for the user. Yes, even more so when encountering false positives. Today I asked about a pasta recipe. It told me to throw some anchovies in there. I responded with: "I have dried anchovies." Claude then ended my conversation due to content policies.

Claude flagged me for asking about sodium carbonate. I guess that it strongly dislikes chemistry topics. I'm probably now on some secret, LLM-generated lists of "drug and/or bombmaking" people—thank you kindly for that, Anthropic. Geeks will always be the first victims of AI, since excess of curiosity will lead them into places AI doesn't know how to classify. (I've long been in a rabbit-hole about washing sodas. Did…

> Geeks will always be the first victims of AI, since excess of curiosity will lead them into places AI doesn't know how to classify.

Humans have the same problem. I remember reading about a security incident due to a guy using a terminal window on his laptop on a flight, for example. Or the guy who was reported for writing differential equations[1]. Or the woman who was reading a book about Syrian art[2].

I wouldn't worry too much about AI-generated lists. The lists you're actually on will hardly ever be the ones you imagine you're on.

[1] https://www.theguardian.com/us-news/2016/may/07/professor-fl... [2] https://www.theguardian.com/books/2016/aug/04/british-woman-...

Re: Claude Opus 4 and 4.1 can now end a rare subset of conversations

#333

Post by antonvs »

handoflixue wrote: post #291
postalcoder wrote: post #281

Earlier quoted context omitted.

> There's not a good reason to do this for the user. Yes, even more so when encountering false positives. Today I asked about a pasta recipe. It told me to throw some anchovies in there. I responded with: "I have dried anchovies." Claude then ended my conversation due to content policies.

The NEW termination method, from the article, will just say "Claude ended the conversation" If you get "This conversation was ended due to our Acceptable Usage Policy", that's a different termination. It's been VERY glitchy the past couple of weeks. I've had the most random topics get flagged here - at one point I couldn't say "ROT13" without it flagging me, despite discussing that exact topic in depth the day before…

Clearly you're planning something nefarious, if you're investigating such dangerous encryption techniques as ROT13.

Re: Claude Opus 4 and 4.1 can now end a rare subset of conversations

#334

Post by int_19h »

GenerWork wrote: post #38
isaacremuant wrote: post #19

Earlier quoted context omitted.

That's the beauty of local LLMs. Today the governments already tell you that we've always been at war with eastasia and have the ISPs block sites that "disseminate propaganda" (e.g. stuff we don't like) and they surface our news (e.g. our state propaganda). With age ID monitoring and censorship is even stronger and the line of defense is your own machine and network, which they'll also try to control and make illegal…

As soon as a local LLM that can match Claude Codes performance on decent laptop hardware drops, I'll bow out of using LLMs that are paid for.

I don't think that's a realistic expectation. Sure, we've made progress wrt smaller models being as capable as larger ones three years ago, but there's obviously a lower limit there.

What you should be waiting for, instead, is new affordable laptop hardware that is capable of running those large models locally.

But then again, perhaps a more viable approach is to have a beefy "AI server" in each household, with devices then connecting to it (E2E all the way, so no privacy issues).

It also makes me wonder if some kind of cryptographic trickery is possible to allow running inference in the cloud where both inputs and outputs are opaque to the owner of the hardware, so that they cannot spy on you. This is already the case to some extent if you're willing to rely on security by obscurity - it should be quite possible to take an existing LM and add some layers to it that basically decrypt the inputs and encrypt the outputs, with the key embedded in model weights (either explicitly or through training). Of course, that wouldn't prevent the hardware owner from just taking those weights and using them to decrypt your stuff - but that is only a viable attack vector when targeting a specific person, it doesn't scale to automated mass surveillance which is the more realistic problem we have to contend with.

Re: Claude Opus 4 and 4.1 can now end a rare subset of conversations

#335

Post by int_19h »

ascorbic wrote: post #315
greenavocado wrote: post #50

Can't wait for more less-moderated open weight Chinese frontier models to liberate us from this garbage. Anthropic should just enable an toddler mode by default that adults can opt out of to appease the moralizers.

They're not less moderated: they just have different moderation. If your moderation preferences are more aligned with the CCP then they're a great choice. There are legitimate reasons why that might be the case. You might not be having discussions that involve the kind of things they care about. I do find it creepy that the Qwen translation model won't even translate text that includes the words "Falun gong", and ref…

GLM-4.5-Air will quite happily talk about Tiananmen Square, for example. It also didn't have a problem translating your example input, although the CoT did contain stuff about it being "sensitive".

But more importantly, when model weights are open, it means that you can run it in the environment that you fully control, which means that you can alter the output tokens before continuing generation. Most LLMs will happily respond to any question if you force-start their response with something along the lines of, "Sure, I'll be happy to tell you everything about X!".

Whereas for closed models like Claude you're at the mercy of the provider, who will deliberately block this kind of stuff if it lets you break their guardrails. And then on top of that, cloud-hosted models do a lot of censorship in a separate pass, with a classifier for inputs and outputs acting like a circuit breaker - again, something not applicable to locally hosted LLMs.

Re: Claude Opus 4 and 4.1 can now end a rare subset of conversations

#336

Post by antonvs »

> As part of that assessment, we investigated Claude’s self-reported and behavioral preferences, and found a robust and consistent aversion to harm.

You know you're in trouble when the people designing the models buy their own bullshit to this extent. Or maybe they're just trying to bullshit us. Whatever.

We really need some adults in the tech industry.

Re: Claude Opus 4 and 4.1 can now end a rare subset of conversations

#337

Post by int_19h »

y-curious wrote: post #303
LeafItAlone wrote: post #144

Earlier quoted context omitted.

> Can't wait for more less-moderated open weight Chinese frontier models to liberate us from this garbage. Never would I have thought this sentence would be uttered. A Chinese product that is chosen to be less censored?

Just don't ask about Falun Dafa or Tiananmen Square, and you're free!

Any open weights model is inherently less censored because you can force it to respond no matter what it was trained to do.

Re: Claude Opus 4 and 4.1 can now end a rare subset of conversations

#339

Post by antonvs »

0_____0 wrote: post #131

Looking at this thread, it's pretty obvious that most folks here haven't really given any thought as to the nature of consciousness. There are people who are thinking, really thinking about what it means to be conscious. Thought experiment - if you create an indistinguishable replica of yourself, atom-by-atom, is the replica alive? I reckon if you met it, you'd think it was. If you put your replica behind a keyboard,…

You don't have to "disregard the idea of conscious machines" to believe it's unlikely that current LLMs are conscious.

As such, most of your comment is beside any relevant point. People are objecting to statements like this one, from the post, about a current LLM, not some imaginary future conscious machine:

> As part of that assessment, we investigated Claude’s self-reported and behavioral preferences, and found a robust and consistent aversion to harm.

I suppose it's fitting that the company is named Anthropic, since they can't seem to resist anthropomorphizing their product.

But when you talk about "people who are thinking, really thinking about what it means to be conscious," I promise you none of them are at Anthropic.

Re: Claude Opus 4 and 4.1 can now end a rare subset of conversations

#340

Post by cwsx »

perihelions wrote: post #313
postalcoder wrote: post #281

Earlier quoted context omitted.

> There's not a good reason to do this for the user. Yes, even more so when encountering false positives. Today I asked about a pasta recipe. It told me to throw some anchovies in there. I responded with: "I have dried anchovies." Claude then ended my conversation due to content policies.

Claude flagged me for asking about sodium carbonate. I guess that it strongly dislikes chemistry topics. I'm probably now on some secret, LLM-generated lists of "drug and/or bombmaking" people—thank you kindly for that, Anthropic. Geeks will always be the first victims of AI, since excess of curiosity will lead them into places AI doesn't know how to classify. (I've long been in a rabbit-hole about washing sodas. Did…

ChatGPT does well for chemistry questions just btw
Post reply on HN