Glenn 'devalias' Grant

> Using `BindingIdentifier` instead of `Identifier` now For continuity, I believe this was implemented in: https://github.com/jehna/humanify/commit/1fffcde1d1bed609fb3e1b72171ea983e0c1d2bd https://github.com/jehna/humanify/blob/311758f42f1c1b4ce599b2e84116454bbfc000b7/src/plugins/local-llm-rename/visit-all-identifiers.ts#L68-L77

AIStudio isn't accessible via API as far as I know; and they've only recently started giving API access to a limited set of people applying; limited to a single Google...

These days I would probably suggest widening the scope of this issue to include multiple Gemini based models, as there are things like Gemini 1.5 Flash and similar now too:...

@jehna While it wouldn't make for the best dev workflow, I wonder if adding some basic cross-platform build tests on GitHub actions would help highlight/protect against these sorts of issues?

> Zeromq [does claim](https://github.com/zeromq/zeromq.js/tree/5.x?rgh-link-date=2024-06-19T18%3A12%3A32Z) it should work on Windows too, so not sure what the issue here is 🤔 Unsure if this is relevant, since theoretically it should have binaries...

> The prebuilt binaries only exist up to node v18 Yeah, I was just about to say, the script in the `package.json` lists the following: - https://github.com/zeromq/zeromq.js/blob/5.x/package.json#L30-L31 - ``` "prebuildify":...

> did anyone fix this? @JxxIT I believe the solution is probably "use an older version of node that the zeromq.js binaries are built for" And maybe when upstream releases...

> I'll check what the rate limits are currently Not sure if these have changed since the OpenAI dev day recently (I think they mentioned increasing the rate limits from...

> but I think it makes more sense to handle the rate limit error as it occurs @jehna Yeah, that sounds like a good approach to me. I'm not sure,...

> This is still a major limitation that requires a mitigation This would seem to be the section of code making the API call (at least for the OpenAI plugin):...