Excited to try it. We used Claude 1.x in experimentation, but shipped with OpenAI primarily because of time and SOC 2 compliance. Anthropic has come along since then, so we'll probably experiment with Claude more with intent to take into production if it's still holding up.
Great to hear -- I just noticed that in your trust center as well! This was really the only blocker. Claude works very well, and our prompt we use with OpenAI can even just be used without any changes.
Does this page need to be updated? It say's "We're currently working towards SOC 2 and HIPAA compliance", while in another place (can't find it at the moment) it say's you're compliant.
Also, the trust portal link on that page is a link to itself. Not the trust portal.
Cool. As CISO, can you please speak to the data retention policies that I noted here? https://news.ycombinator.com/item?id=36681239 . As you can imagine, sending sensitive information to a 3rd party is impossible without explicit agreements. As you're SOC2 and HIPAA are there devices in place for us to delete data, or specify data retention as customers?
That’s awesome news - it’s be great if you also started do offer the chat style api that OpenAI now has including functions. I know it sucks to copy a competitor but it would allow us to plug in your api seamlessly and give customer base the option between multiple models which could be really cool especially with the 100k token limit.
