Nostr Web Client

Image/Video gen is an intentional omission for now. Gotta figure out how to handle really unsavory content. We don’t censor AI input or output in nostr:nprofile1qqs8msutuusu385l6wpdzf2473d2zlh750yfayfseqwryr6mfazqvmgpr4mhxue69uhkummnw3ez6vp39eukz6mfdphkumn99e3k7mf0qythwumn8ghj7un9d3shjtnwdaehgu3wvfskuep0ucrggs, so you can imagine that would create a dicey environment when adding image and video.

nostr:nevent1qqs8kwad80q9x49gj2nj84dxm3t2qwufl55ndntvq4qyaqnn8p9gmdspzemhxue69uhhyetvv9ujuurjd9kkzmpwdejhgvq6f8z

τέχνη 4mo ago 💬 1

Isn’t the censorship built into the models themselves?

Reply to this note

Please Login to reply.

Discussion

Marks 4mo ago 💬 2

For ChatGPT, Grok, and Claude the censorship happens at multiple levels.

1. Censorship at the model training level

2. Censorship at the input box

3. Censorship in hidden system prompt

nostr:nprofile1qyjhwumn8ghj7en9v4j8xtnwdaehgu3wvfskuep0dakku62ltamx2mn5w4ex2ucpxpmhxue69uhkjarrdpuj6em0d3jx2mnjdajz6en4wf3k7mn5dphhq6rpva6hxtnnvdshyctz9e5k6tcqyp7u8zl8y8yfa87nstgj2405t2shal4rez0fzvxgrseq7k60gsrx6zeuh5t only has number 1 because we use models trained by others. Over time we are providing a variety of open models so people can choose the training bias that works best for them.

We do not do numbers 2 and 3.

nostr:nevent1qvzqqqqqqypzq4maupkuu9s2qdu3vwjtk7mgp037pg8pc6x7de463sq3xj6yqexaqqsyqapu0kn3a84xscgd2n5a420m9dlavt3d8m2v75xfw4stxup6a8s85z457