234120987654

4 ms·

1. 1.▲
   by 234120987654 2y ago

   Fine-tuning an LLM to create long-form podcast timestamps. Apparently even the best LLMs with long context are incredibly bad with this so I'm curating data and creating a service out of it.
2. 2.▲
   by 234120987654 2y ago

   100% agree. I miss the days where the title would describe the method instead of being a sales pitch
3. 3.▲
   by 234120987654 2y ago

   Great write up and clarification of some of the errors the authors made. The only thing I'm missing is the evaluation of structured generation on longer, complex JSON. In my experience this does tend to fall off depending on the model
4. 4.▲
   by 234120987654 2y ago

   > Wrong you can do most of the things the author showed with a fresh install of vim/emacs or by logging in to a fresh install of vscode/intellij - In other words no lifetime was spent on this, I like having as bare an experienc
5. 5.▲
   by 234120987654 2y ago

   Wow, I did not expect to see such negativity in this thread. Most of them read to me like the "Dropbox is just an FTP"-narrative. Yes, you and your pride can do most of these things in 0.3ms and better, but so will 1 million more
6. 6.▲
   by 234120987654 2y ago

   A17 Pro huh, that's a first for putting a pro chip in a non pro iPad, isn't it? I guess it's, as they advertise, to handle Apple Intelligence although I don't understand why they are doubling down on this _now_ while not
7. 7.▲
   Zotero 7: Zotero, Redesigned(zotero.org)

   52 points by 234120987654 2y ago|20 comments