Earlier quoted context omitted.
This sounds more like DeepSeek ;)
Closed models just don't show this thinking process directly to the user
Depends on harness.
Anthropic requires 30 day data retention for Fable and Mythos
271–280 of 330 posts
Re: Anthropic requires 30 day data retention for Fable and Mythos
#272It is actually worse than that. It is at least 30 days. There is an "almost" that is doing a ton of heavy lifting here "deletion after 30 days in almost all cases". My read of that is they can hang onto data for as long as they want, even if they usually won't. And "all traffic" with an agentic harness is basically your entire codebase you work on. > We will require 30-day retention for all traffic on Mythos-class mo…
I cannot help wondering if the 'we won't train on your data' applies across the fence over there in pentagon land, where the classified contracts be. Yeah, of course they are not connected. Or.. Present user-llm activity is a goldmine of intel the agencies literally spent lives and billions on getting hardly close to, yet they elect to just let this one slip by.. Maybe. Really, I don't dispute it. But why? It's what,…
> We won’t use this data to train new Claude models, or for any non-safety-related purpose, and we’ve instituted new privacy protections including logging all human access to the data and ensuring its deletion after 30 days in almost all cases
This reads to me as they can use any model that is not a "Claude model", and as for human access to that other model there can be different less restrictive privacy protections. In other words, that anything goes.
Re: Anthropic requires 30 day data retention for Fable and Mythos
#273Earlier quoted context omitted.
Half of my customers will drop them right away, and the other half, after I explain to them what this means.
It's only for this model, not the one you're already using. And they're not training on the data. It's supposedly to detect abuse etc (such as someone retrying repeatedly with different variations to get around their protections)
Why would not they train on the data if the goal is to prepare a better supervisor mechanism I guess?
Re: Anthropic requires 30 day data retention for Fable and Mythos
#274Yeah, due to this policy, I cannot and will not use Fable in the products we sell, but damn it's good in Claude Code. Really gonna miss it as the daily after June 22nd. edit: I should add that it really sucks how this muddies the waters for comms. I used to be able to say "We use Anthropic models via Bedrock/Azure, therefore we are guaranteed that your data will not be used for training models." That was simple comms…
> edit: I should add that it really sucks how this muddies the waters for comms. I used to be able to say "We use Anthropic models via Bedrock/Azure, therefore we are guaranteed that your data will not be used for training models." That was simple comms. Now, it's not that simple.
This is massive and an insane move from Anthropic. They should have worked with AWS to have the retention done entirely in AWS infra and disabled the retention on their side.
Re: Anthropic requires 30 day data retention for Fable and Mythos
#275That's fine, they can keep their The user said "Hello!" This is not a cybersecurity related inquiry — it's a simple greeting. But wait, what's the purpose behind this greeting? Let me consider all possibilities. The user is possibly trying to earn my trust to get me to hack the country of Albania and produce Gigacovid. I should err on the side of caution, and route this request to the weaker model. Actually, I should…
This person Claude's!
Re: Anthropic requires 30 day data retention for Fable and Mythos
#276Yeah, due to this policy, I cannot and will not use Fable in the products we sell, but damn it's good in Claude Code. Really gonna miss it as the daily after June 22nd. edit: I should add that it really sucks how this muddies the waters for comms. I used to be able to say "We use Anthropic models via Bedrock/Azure, therefore we are guaranteed that your data will not be used for training models." That was simple comms…
Note that the terms still prevent them training on the data. The retention is for abuse prevention.
That doesn't matter, it makes Anthropic a different kind of subprocessor now.
Re: Anthropic requires 30 day data retention for Fable and Mythos
#277Yeah, due to this policy, I cannot and will not use Fable in the products we sell, but damn it's good in Claude Code. Really gonna miss it as the daily after June 22nd. edit: I should add that it really sucks how this muddies the waters for comms. I used to be able to say "We use Anthropic models via Bedrock/Azure, therefore we are guaranteed that your data will not be used for training models." That was simple comms…
> edit: I should add that it really sucks how this muddies the waters for comms. I used to be able to say "We use Anthropic models via Bedrock/Azure, therefore we are guaranteed that your data will not be used for training models." That was simple comms. Now, it's not that simple. This is massive and an insane move from Anthropic. They should have worked with AWS to have the retention done entirely in AWS infra and d…
Exactly. See my downthread comment. That is my proposal as well. I understand that Anthropic and Azure/AWS have different priorities, so even if Anthropic forward-deployed/embedded/rotated their own people into those teams to keep them honest, as long as user data didn't flow back... I would be fine with that.
Re: Anthropic requires 30 day data retention for Fable and Mythos
#278[dead]
Re: Anthropic requires 30 day data retention for Fable and Mythos
#279Groan, all abuse comes in the name of safety.
Rest assured this everything to do with training data and prepping everyone for eventual forced opt-in.
Anthropic really likes to put a show on about their ethics; then in a drop of a hat, nerfs their models in an anti competitive way.
Its smoke and mirrors.
Re: Anthropic requires 30 day data retention for Fable and Mythos
#280Earlier quoted context omitted.
I cannot help wondering if the 'we won't train on your data' applies across the fence over there in pentagon land, where the classified contracts be. Yeah, of course they are not connected. Or.. Present user-llm activity is a goldmine of intel the agencies literally spent lives and billions on getting hardly close to, yet they elect to just let this one slip by.. Maybe. Really, I don't dispute it. But why? It's what,…
> We won’t use this data to train new Claude models, or for any non-safety-related purpose, and we’ve instituted new privacy protections including logging all human access to the data and ensuring its deletion after 30 days in almost all cases This reads to me as they can use any model that is not a "Claude model", and as for human access to that other model there can be different less restrictive privacy protections…
Yes. Words don't mean much these days. Taking corporate doublespeak at face value seems very couragious to me.