Hacker News Reader

Let's use a language model trained on hundreds of gigabytes of god-knows-what, has a tendency to hallucinate, has literally never used your product, is known as a consummate bullshitter, and unleash it unsupervised on your confused users.

Sorry for the snark, but what is your plan for quality assurance that this this meeting users needs instead of just BS'ing them out of your hair?

View on HN

Ha ha ha ! You have a point, That would be the next problem to tackle. Probably try to determine if I should even try to answer a question before attempting it. And then not showing results that don't fit criteria. At a high level, it seems like some embedding works need investigation but I haven't done much evaluation there. There would be filtering at both question and answer levels I suppose.

Reply on news.ycombinator.com