Are you saying you don’t trust ZDR claims from inference providers?

They don’t care about your code. There’s more and better data on the public web.

Isn't ZDR only the case if you "opt out" through a setting? Lots of opportunities to make a mistake here, or for the LLM provider to play games. You could be unaware of the setting. The provider could reset the setting upon subscription lapse/renewal, application update, model release, etc. A developer could accidentally use their personal subscription (w/ setting on) on a work codebase.

Also what's the timing on this setting's effectiveness? What if before you knew to "opt out" you used the agent for a refactor and your entire repo got sucked up in inference? But then you opt out the next day? Is it too late?

▲

kccqzy 3 hours ago | parent | next [-]

The mistakes you describe are only possible if the company doesn’t really think there’s proprietary IP in the codebase.

If a company really believed there’s valuable IP in code, there would at a minimum, ban laptops or ban code on laptops[0], provide and maintain a centrally managed LLM gateway[1] that handles authentication, billing and model choice, have MDM to prevent you from using your own Claude login, etc.

[0]: For example when I worked at Google, the proprietary google3 codebase cannot exist on laptops because there are no tools to download it to your laptop.

[1]: For example Claude code supports having a gateway: https://code.claude.com/docs/en/llm-gateway-connect

	▲	greenowl 13 minutes ago \| parent [-]
		I'd wager many software startups out there would consider their code valuable IP (even if in the AI era it's not), and don't have these types of IT controls in place. Startups I worked at gave you an email account, github access to their private repos, and that was that. Heck, you don't even need to mix up a personal and business AI subscription. For the vast majority of folks one day their IDE "updated" and a bunch of AI features suddenly became available. Free, no sub needed. A dev starts using them (because, why not?) and lo and behold the company's code got shipped out in an inference call and is now scheduled to be in the next training run. Oops.

▲

drdexebtjl 4 hours ago | parent | prev [-]

That’s the case with a personal subscription directly with OpenAI or Anthropic, but enterprise customers are opt-in, AFAIK.

And there are solutions around it from other providers and model routers. OpenRouter lets you explicitly say on a per-request basis you don’t want to be routed to a provider that trains on your input, for example.