How are founders finetuning AI with limited resources?

2 pointsprofsummergig2 comments

I'm seeing a lot of specialized AI's based on ChatGPT and such, which clearly needed a lot of fine-tuning to do (e.g. DeepTutor).

How are founders managing to do finetuning with limited resources? I am under the impression that finetuning is human labor intensive. It clearly isn't.

View on HN

> I am under the impression that finetuning is human labor intensive. It clearly isn't.

It can be labor-intensive and cheap. Amazon's MTurk and similar services offer tuning at wages barely above starvation level even for Global South countries.

https://dataworkers.org/

If you have any other links / articles you can share, please do so. Thanks.

Reply on news.ycombinator.com