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

Updates to Consumer Terms and Privacy Policy

anthropic.com

Post reply on HN
Print view View original

311–320 of 551 posts

Re: Updates to Consumer Terms and Privacy Policy

#311

Post by smohare »

diggan wrote: post #145
bigfishrunning wrote: post #131

Earlier quoted context omitted.

Except LLMs aren't a fundamental or core technology, they're an amusing party trick with some really enthusiastic marketers. We don't need them.

Personally, I'm able to write code I wasn't able to before, like functions heavy with math. For game development, this has been super helpful, when I know basically what inputs I have, and what output I need, but I'm not able to figure out how the actual function implementation should be. Add a bunch of unit tests, let the LLM figure out the math, and I can move on to more important features. For me this been a prett…

You think you are “nailing it” but also lack the background to even determine whether that is the case. I can assure you, there’s likely some fundamental flaws in what you’re vibing.

Just think about the type of code these things are trained on and the fact you’re clearly some random non-specialist.

Re: Updates to Consumer Terms and Privacy Policy

#312

Post by SoftTalker »

octagons wrote: post #288

That’s unfortunate. I believed Anthropic was playing the long game and betting on a smaller but more technically proficient userbase. I guess I’ll be canceling my subscription largely out of principal. I doubt any open-source models are capable of handling my use case as well as Claude (typically focused on getting up to speed with various ISO/IEEE standards for the purpose of security testing) but I’m sure I’ll find…

This is all you can do. Laws and regulations will not be forthcoming, and even if some are, they will be ignored or the fines paid as a cost of doing business.

Any data you give to any website or app is no longer (exclusively) yours. Use these services under that assumption.

Re: Updates to Consumer Terms and Privacy Policy

#313

Post by frereubu »

I went to claude.ai and saw a pop-over window with this choice, clicked to opt out and saw a "server error" alert. Then went to Settings > Privacy and the option wasn't there. I had to click something else (can't remember now) to see the option, which wasn't opted out (which kinda made sense because of the error message) and the option is now in Settings > Privacy. I'm definitely someone who believes that cock-up is more likely than conspiracy for things like this, but this process should have been bulletproof - no initial error message, option always in Settings > Privacy - given the sensitivities involved.

Re: Updates to Consumer Terms and Privacy Policy

#314

Post by SoftTalker »

ezfe wrote: post #236
the_other wrote: post #215

I didn't trust them much to begin with, so I generally avoid talking too much personal stuff with Claude. But I had plenty of chats with surface level discussion of topics I'm interested in and some of my relevant experience and history with those topics. So, I have deleted all my chats and am closing my Claude account (as soon as customer services get back to me; somehow the self-serve option is missing for my accou…

This change didn't grant access to your old chats and can be opted out. Good job you overreacted without reading.

For now. Once this settles in they will change the terms again, with no opt-out option.

Re: Updates to Consumer Terms and Privacy Policy

#315

Post by __MatrixMan__ »

chamomeal wrote: post #206
__MatrixMan__ wrote: post #127

Earlier quoted context omitted.

Publishing something is considered by most to be sufficient consent for it to be not considered private. I realize there's a whole legal quagmire here involved with intellectual "property" and what counts as "derivative work", but that's a whole separate (and dubiously useful) part of the law.

That is definitely normally true but I feel like the scale and LLM usage turns it into a different problem. If you can use all of the content of stack overflow to create a “derivative work” that replaces stack overflow, and causes it to lose tons of revenue, is it really a derivative work? I’m pretty sure solution sites like chegg don’t include the actual questions for that reason. The solutions to the questions are…

Stack overflow doesn't really have a legitimate claim to that data either though. Nor do the users, we're just pasting error messages and documentation. It's derivative all the way down. It'll never sit still and behave like property.

Privacy makes sense, treating data like property does not.

Re: Updates to Consumer Terms and Privacy Policy

#316

Post by riz_ »

soulofmischief wrote: post #270
riz_ wrote: post #80

Earlier quoted context omitted.

This is probably because there are laws in some countries that restrict how these buttons/switches can look (think cookie banners, where sometimes there is a huge green button to accept, and a tiny greyed out text somewhere for the settings).

Can you provide an example?

https://www.cnil.fr/en/dark-patterns-cookie-banners-cnil-iss...

Re: Updates to Consumer Terms and Privacy Policy

#317

Post by SoftTalker »

happosai wrote: post #106
I_am_tiberius wrote: post #18

In my opinion, training models on user data without their real consent (real consent = e.g. the user must sign a contract or so, so he's definitely aware), should be considered a serious criminal offense.

I think it's cute people believe companies that trained their models with every single book and online page ever written without consents from authors (and often against the explicit request of the author without any opt-out) won't do a rugg-pull and do it also to all the chats they have aquired...

You're absolutely right, but also isn't the volume of new data they are getting from chats tiny compared to what they've already trained on? I'm wondering how much difference it will really make.

Re: Updates to Consumer Terms and Privacy Policy

#318

Post by Barrin92 »

paradite wrote: post #178
sigmoid10 wrote: post #39

Earlier quoted context omitted.

The are valued at $170 Billion. Not quite the same as, but in same order of magnitude as OpenAI - while having only a single digit percent fraction of active users. They probably need to prepare for the eventual user data sellout, as it is becoming increasingly more obvious that none of the big players has a real and persistent tech leadership anymore. But millions and millions of users sharing their deepest thoughts…

Claude Sonnet 4 is the best coding model. Period. Nothing else comes close. Anthropic probably has 80% of AI coding model market share. That's a trillion dollar market.

>That's a trillion dollar market

not if you have to constantly expend enormous sums to stay ahead of your competition or otherwise you lose your edge. It's not the best coding model because they got some mystical treasure in their basement. It's so rapidly becoming a commodity that at some point Microsoft or Google will just offer just as good a model for free and like search they'll just start milking people with ads.

That's likely one of the reasons for the shifting privacy stances, not just for training but because monetization of the product itself is probably looking pretty dim in the long run.

Re: Updates to Consumer Terms and Privacy Policy

#319

Post by tjr »

chamomeal wrote: post #184
bigfishrunning wrote: post #131

Earlier quoted context omitted.

Except LLMs aren't a fundamental or core technology, they're an amusing party trick with some really enthusiastic marketers. We don't need them.

I’d say they’re a fundamental technology by now. Imagine how many people rely on them. And I’ve seen some heavy reliance.

Unless one's job expectations have been altered to demand LLM-quantity output, how could someone be reliant upon these tools now? What were they doing two years ago (or maybe even six months ago)?

I can understand becoming reliant on a technology -- I expect most programmers today would be pretty lost with punch cards or line editors -- but LLM coding seems too new for true reliance to have formed yet...?

Re: Updates to Consumer Terms and Privacy Policy

#320

Post by Deegy »

I guess I'll take the other side of what most are arguing in this thread.

Isn't it a great thing for to us to collectively allow LLM's to train on past conversations? LLM's probably won't get significantly better without this data.

That said I do recognize the risk of only a handful of companies being responsible for something as important as the collective knowledge of civilization.

Is the long term solution self custody? Organizations or individuals may use and train models locally in order to protect and distribute their learnings internally. Of course costs have to come down a ridiculous amount for this to be feasible.

Post reply on HN