Hey, this is a great question - even now you can use OpenAI technologies in the isolated environments - for example in Azure you can have your own OpenAI instance, this can be included in your network and not accessible without VPN. The other potential solutions are using OpenSource models like LLama3 (check hughingface) and build your own AI server in your own datacenter (if you don't trust Azure). For certain usecases, if you have OpenAI subscription, all the data should be safe - as it is written in license agreement with OpenAI. So it's up to you which solution is best for you or your customer/employer :)
Thank You!
Loved this talk
Really Inspiring Talk! Thanks,
This was a nice talk
So how will code be protected and isolated when using the ChatGPT-wrapper? Is it safe to use on sensitive data?
Hey, this is a great question - even now you can use OpenAI technologies in the isolated environments - for example in Azure you can have your own OpenAI instance, this can be included in your network and not accessible without VPN. The other potential solutions are using OpenSource models like LLama3 (check hughingface) and build your own AI server in your own datacenter (if you don't trust Azure).
For certain usecases, if you have OpenAI subscription, all the data should be safe - as it is written in license agreement with OpenAI. So it's up to you which solution is best for you or your customer/employer :)