It 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 models, on both first- and third-party surfaces. 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 (see this post for further details). The data will help us defend against complex and novel attacks (including new jailbreaks and attacks that operate across many requests) as well as help us identify and reduce false positives.
A startup that uses agentic coding tools such as Claude Code or Codex is packaging up their entire codebase and sending it directly to their LM provider. Depending on their product, they might be sending it directly to a potential competitor.
people over-rate how much software/IP is useful in running a successful business. There are genuinely very few IP in this world that needs to be protected. Everyone else is running stupid CRUD apps
They also over index fear of LargeCo stealing IP from SmallCo. In fact, LargeCo is typically more scared about even the possibility of any product team looking at competitor internals due to lawsuits.
I’d be more scared of a data leak due to LargeCo being hacked than I would about LargeCo prying into the data.
What I don’t trust LargeCo with is personal information. I’ve heard too many horror stories about Govs and LargeCos swapping customer nudes or stalking ex’s to be comfortable with anything personal on those systems. But that’s a whole different topic.
Yes, it certainly is an odd situation when some people believe you cannot use Mythos-class models because security while others believe you must do code reviews with Mythos-class models because security.
I guess the better question would be if you are under and NDA and using an online model, are you already violating it but does this violate it further?
Your NDAs prohibit emailing a colleague about the e.g. project, or discussing it in a Slack DM with the client, or tracking progress on it in JIRA? You have to do NDA’d work exclusively with local tools or end-to-end encryption? Those are some difficult NDAs!
We use inhouse on-premises email, issue tracking, and messaging. Depending on the project, external communication does require E2EE email. Development happens on local hardware and software unless required otherwise by the customer.
Fortunately I can't use Fable anyway, since their hyperactive content flaggers do not let you work on anything remotely biological or medical related (i.e. parse a CSV with some medical content, nope, you're probably a bioterrorist) and you get downgraded to Opus immediately.
Yes! I have hit the same brick wall. What sort of idiots are doing this? Honestly, I have no idea. And just before their IPO. SO far Anthropic marketing has been perfect and spotless. This is serious slipup.
There were two (expensive) exceptions / alternatives so far: Bedrock and Vertex. Their Zero Data Retention was in fact contractually enforced. Now it is all f...d because of these morons at Anthropic. For now I am better off just using DS via their API.
This is just a tragic moment for Tech. We just killed AI privacy. OpenAI already follows this trend and others will do too.
I mean not just the part 30 days data retention but I think the serious trade of this product is just the token efficiency. They trade it for precision. The claims that they make that it found a 30 year software bug from millions of lines of code is just precision. To human it's looks like a lot but for it it's just the ablity to process (token processing). Let's see how long it runs. Peace.
I'm talking about scouring Twitter/LinkedIn and look at posts from employees who say SOTA model is banned. Look at what the business do. Copy it using SOTA. Call their clients with 30% discount and faster turnaround and higher quality product.
It is complicated, but I can get Private Equity of even VCs to fund this idea.
tl;dr -- I'm actually agreeing with you. Anthropic will never copy your business model due to NDA. But there are plenty of fearmongering about they copying you and because of which you won't use their models. If their models are genuinely SOTA you can use that information to your advantage and crush scaredy-cats.
All I can say to my team (and my clients): "f...k Anthropic". They've just put both Bedrock and Vertex on slippery slope of "we don't collect your prompts. period. ... comma ... except ..."
Right now we have changed the code of all our agents to data retention mode 'none' (Note: not "default" or "inherited", this is not enough now!) and we are fighting with GCP doco to set similar things for Vertex.
It 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 models, on both first- and third-party surfaces. 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 (see this post for further details). The data will help us defend against complex and novel attacks (including new jailbreaks and attacks that operate across many requests) as well as help us identify and reduce false positives.
How were they not already auditing access to customer data?
They were not keeping it beyond the timeframe necessary for the model to process it, so there wasn't access there to audit.
Half of my customers will drop them right away, and the other half, after I explain to them what this means.
Even worse when you git push something Microsoft gets all your code!
Only if you push it to GitHub.
That is why, for the last five years I have been checking in with them, code with some of the most atrocious quality. So far...its working....
The system works!
[dead]
A startup that uses agentic coding tools such as Claude Code or Codex is packaging up their entire codebase and sending it directly to their LM provider. Depending on their product, they might be sending it directly to a potential competitor.
Odd times we are living in!
people over-rate how much software/IP is useful in running a successful business. There are genuinely very few IP in this world that needs to be protected. Everyone else is running stupid CRUD apps
They also over index fear of LargeCo stealing IP from SmallCo. In fact, LargeCo is typically more scared about even the possibility of any product team looking at competitor internals due to lawsuits.
I’d be more scared of a data leak due to LargeCo being hacked than I would about LargeCo prying into the data.
What I don’t trust LargeCo with is personal information. I’ve heard too many horror stories about Govs and LargeCos swapping customer nudes or stalking ex’s to be comfortable with anything personal on those systems. But that’s a whole different topic.
LargeCo is probably struggling under the weight of technical debt and organizational challenges/politics.
I bet if you gave them the Codebase of the Gods, it’d be a heap of hacks inside a couple months.
> people over-rate how much software/IP is useful in running a successful business
Indeed, by a couple trillions...
Yes, it certainly is an odd situation when some people believe you cannot use Mythos-class models because security while others believe you must do code reviews with Mythos-class models because security.
and all their keys, because sooner or later, the harness is gonna read them
One company's irrational fear is a competitive advantage for someone else.
Not just “a startup”! Also, famously, Meta, with their famous AI usage dashboards
Related ongoing thread:
AWS Bedrock to require sharing data with Anthropic for Mythos and future models - https://news.ycombinator.com/item?id=48473166 - June 2026 (223 comments)
Google Cloud also makes you accept this safety addendum to deploy Fable 5 via their Model Garden https://cloud.google.com/terms/advanced-ai-safety-addendum
So if you are under an NDA, does this violate it?
I guess the better question would be if you are under and NDA and using an online model, are you already violating it but does this violate it further?
In the same way that using Gmail and Dropbox and iCloud and Notion violates it. (Which IANAL but for most NDAs would be not at all.)
I never had an NDA permit such usage.
Your NDAs prohibit emailing a colleague about the e.g. project, or discussing it in a Slack DM with the client, or tracking progress on it in JIRA? You have to do NDA’d work exclusively with local tools or end-to-end encryption? Those are some difficult NDAs!
We use inhouse on-premises email, issue tracking, and messaging. Depending on the project, external communication does require E2EE email. Development happens on local hardware and software unless required otherwise by the customer.
Fortunately I can't use Fable anyway, since their hyperactive content flaggers do not let you work on anything remotely biological or medical related (i.e. parse a CSV with some medical content, nope, you're probably a bioterrorist) and you get downgraded to Opus immediately.
Yes! I have hit the same brick wall. What sort of idiots are doing this? Honestly, I have no idea. And just before their IPO. SO far Anthropic marketing has been perfect and spotless. This is serious slipup.
It's good they're being overcautious here. The alternative is far worse.
the real risk is using it at all as you are already sending them your data. If you are ok with that, then this retention/review seems ok.
There were two (expensive) exceptions / alternatives so far: Bedrock and Vertex. Their Zero Data Retention was in fact contractually enforced. Now it is all f...d because of these morons at Anthropic. For now I am better off just using DS via their API.
This is just a tragic moment for Tech. We just killed AI privacy. OpenAI already follows this trend and others will do too.
The only hope now is ... tada .. Mistral LOL
Hmmm no? The only way is to deploy your own local model, using anyone else's you are at their whim on what happens to your data.
From a personal use perspective yes, the big issue here is enterprise and existing contracts as surely most companies will have signed zero retention.
I mean not just the part 30 days data retention but I think the serious trade of this product is just the token efficiency. They trade it for precision. The claims that they make that it found a 30 year software bug from millions of lines of code is just precision. To human it's looks like a lot but for it it's just the ablity to process (token processing). Let's see how long it runs. Peace.
Then don’t use it.
That’s exactly what my employer had communicated. It will not be allowed.
Step 1: Find all companies which refuses/bans to use SOTA models from irrational fear.
Step 2: Use SOTA models to copy them and crush them
Step 3: Profit.
(Yes, not every business is easily replicable, but you sure can find some)
This. And AI labs seem to be above IP / Copyright law and absolutely nothing will happen to them when they grab all the data and package it up.
Step 4, get sued because you violated an NDA or other regulation?
I'm not talking about Claude copying.
I'm talking about scouring Twitter/LinkedIn and look at posts from employees who say SOTA model is banned. Look at what the business do. Copy it using SOTA. Call their clients with 30% discount and faster turnaround and higher quality product.
It is complicated, but I can get Private Equity of even VCs to fund this idea.
tl;dr -- I'm actually agreeing with you. Anthropic will never copy your business model due to NDA. But there are plenty of fearmongering about they copying you and because of which you won't use their models. If their models are genuinely SOTA you can use that information to your advantage and crush scaredy-cats.
Can you name a single example of a business that has been replaced by another business leveraging LLMs to copy and "crush" their software?
Pretty much any Chinese business. (Except takeouts and laundries)
I mean, this is the biggest reason that's my employer's position
What an annoying company, I wish it didn't exist..
[flagged]
[dead]
I actually think that’s warranted. And if you used it to poke around, you would also agree.
> And if you used it to poke around, you would also agree.
Would you elaborate? Not sure what you're describing
All I can say to my team (and my clients): "f...k Anthropic". They've just put both Bedrock and Vertex on slippery slope of "we don't collect your prompts. period. ... comma ... except ..."
Right now we have changed the code of all our agents to data retention mode 'none' (Note: not "default" or "inherited", this is not enough now!) and we are fighting with GCP doco to set similar things for Vertex.
This is just terrible.