190n

I don't understand why this issue only appeared after the upgrade, not immediately after the migration to add the column?

> And if you copy-paste the answers from LLF, I think it's only fair the end result gets flagged. You're not writing it yourself.

I wonder if it would even get flagged in that case, because wouldn't the probability distribution of a token when the LLM is suggesting an edit to your writing be different than the distribution of that token once it is in the context of the text it's editing?

Almost all of Bun's tests are written in JavaScript run in Bun itself.

> PT_DENY_ATTACH

All you have to do is attach to the process before it does that, and then prevent this call from going through.

It could be because of a Bun bug, but I don't think it's because of that one. It's a duplicate of a year-old issue, and it's specific to Bun.serve.

> *2-bit quantization produces \name\ instead of "name" in JSON output, making tool calling unreliable.

I was wondering about that statement. Shouldn't it restrict sampling to only tokens that produce valid JSON matching the schema during a tool call? On the other hand, I have heard a lot about how even production LLM providers don't always call tools accurately, so I suppose either it's hard to implement what I described or there's something I haven't thought of that makes it impossible.

Later in the post they expand on that point:

> More direct control over updates, including the ability to pause updates for as long as you need

So it does sound like you'll be able to pause updates forever and also therefore not automatically reboot.

Operating systems use less RAM than Firefox.