Hacker News

points by Shorel 4 days ago | 0 comments

You may get the same guarantees on paper, but at this point there's absolutely no reason to trust any AI provider.

At best they get a fine, and it will not make a dent in their cash reserves.

Bedrock guarantees neither your inputs are shared with the LLM providers, neither your data will be used to train the next model. I dont think AWS loves AI that much...to break their business model with a violation of terms and conditions.

But I was not even suggesting that, just run your local LLM on Cloud infra. Do you really believe they are ready to risk their Cloud business just to spy on your prompts?

[–] dmantis 4 days ago

Antrophic works with an ex-deputy of the CIA (David S. Cohen). OpenAI has an ex-director of the NSA (Paul Nakasone) in the board. Literally.

[–] Betelbuddy 3 days ago

The Bedrock models are hosted at AWS.

[–] dmantis 3 days ago

I'm sure these gentlemen will figure it out.