I paid Gemini $9 to train its own replacement
I had Gemini train its own replacement for $9
A Reddit knife-enthusiast scraper was racking up Gemini API bills, so I had Gemini label 4,290 comments once for $9, then fine-tuned GLiNER large v2.5 on those labels. The result: 0.83 F1 against Gemini's annotations on a held-out set, running locally on a Tesla T4. Five of ten training runs failed, including two due to a misunderstood tensor called words_mask. Total cost: about $11.50 plus days of debugging.
If I had to choose between a better label set and an assertion on every tensor I hand-build, I would take the assertion.
- trollbridge
Request subtopic be changed to “I used Gemini to design a tool to replace specific uses of Gemini.”
- faidit
Unfortunately, advertisers are getting smarter and using bots to praise their own products on Reddit. Thanks to training on genuine comments, some models are very good at sounding like a human commenter, and can easily generate a comment history with diverse interests to appear human, making them basically undetectable. So it seems like this method will, at some point, not identify the company with the best knife, but the one with the most ad spend on bot comments.
- josefresco
The other day I wanted to gather Reddit comments about a solar panel vendor. Claude doesn't have access to I had Gemini do some "deep research". When I fed the verbose report back to Claude it basically said it was a bunch of "hallucinated bullshit".
- niekverw
It’s unreadable but then again it says it on top, but it really is so why post it
- yread
I was hoping he tricked Gemini into running the training on the cluster that Gemini itself is running on. That would be novel!