I feel like the suggestion to just delete your Meta account is unhelpful and even harmful in some situations. I know a lot of us in this field don't mind having a small social circle, but for us that struggle to even have that, or that the process of doing that, requires you to be as open and reachable as possible to make it easier to create or maintain friendships, this often requires having an account to use Messenger (I know regionally this may differ).
Being difficult in this way, when most other people, particularly non technical people, don't have a problem with this, in my experience significantly hurts your opportunities and limits the types of people you socialize with in real life, which I think can eventually be harmful.
I do some IRL activities where many of the participants use social media apps to coordinate and plan sessions. I don't have any of these accounts so while I do miss out on some of the casual things, I just show up and still get 80% or more of the interaction and all of thething I'm after in the first place: doing stuff.
If there is more friction to communicate with me I expect to hear from people less often.
As an example, my friends that only use Discord and not SMS or Signal definitely hear from me less frequently. It doesn’t mean we’re not friends, but we do communicate less frequently.
I am not a sparkly special pretty princess. My friends are where they are. In the trade off between meeting people where they're at or doing something unique to me with friction to them, I know which one wins.
(note - this is snarky and I don't actually fully disagree with you)
If this is still the same install that you've been using since 38, you might find a clean install resolves some issues (whether or not your upgrade got botched). Also helps me get rid of software I installed that I don't use anymore, which I feel is relevant to this article. But part of why I love Silverblue so much is I don't have to worry about upgrades getting botched and fwiw as well, I haven't noticed any of those bugs on 44 across several very different machines.
If Jellyfin ever fixes the mountain of bugs from the "upgrade". They aren't even acknowledging major bugs that make Jellyfin unusable for like 20% of users.
Do not upgrade Jellyfin if you have a sizeable library. Backup first if you do.
Not sure popularity necessarily suggests it's good, but possibly just what people have most heard of or is easiest to setup with. This is going to be even more true now that Claude subscriptions are going to be essentially vendor locked.
Given how far ahead OpenAI was in mindshare, monthly users, and revenue over Anthropic when Claude Code came out, I think we can conclude there's at least some substance behind claims of Claude Code's better product quality.
Comparing this project to is-odd seems very disingenuous to me. My understanding is this was the only way you could use llama.cpp with Claude Code for example, since llama.cpp doesn't support the Anthropic compatible endpoint and doing so yourself isn't anywhere near as trivial as your comparison. Happy to be corrected if I'm wrong.
That's a correct example, and I agree, it is disingenuous to just trivially call this an `is-odd` project.
Back in the days of GPT-3.5, LiteLLM was one of the projects that helped provide a reliable adapter for projects to communicate across AI labs' APIs and when things drifted ever so slightly despite being an "OpenAI-compatible API", LiteLLM made it much easier for developers to use it rather than reinventing and debugging such nuances.
Nowadays, that gateway of theirs isn't also just a funnel for centralizing API calls but it also serves other purposes, like putting guardrails consistently across all connections, tracking key spend on tokens, dispensing keys without having to do so on the main platforms, etc.
There's also more to just LiteLLM being an inference gateway too, it's also a package used by other projects. If you had a project that needed to support multiple endpoints as fallback, there's a chance LiteLLM's empowering that.
Hence, supply chain attack. The GitHub issue literally has mentions all over other projects because they're urged to pin to safe versions since they rely on it.
I would hazard a guess that it's because there's been many debates about contributing PRs that might be perceived as AI slop. Not saying that's the case here, but it's possible the fix might be a poor one, not follow the project's guidelines, or one which the contributor doesn't fully understand, but doesn't care because it fixed the issue. I would guess the better approach would be to submit a bug report with the same information the LLM used, and maybe suggest there the fix the LLM provided. Unless this really was a tiny patch and none of the above concerns applied.
The prompt processing times I've heard about have put me off wanting to go that high with memory on the M series (hoping that changes for the M5 series though). What's the average and longest times you've had to wait when using opencode? Has any improvements to mlx helped in that regard?
The M5 ultra series is supposed to have some big gains around prompt processing - something like 3-4x from what I've read. I'm tempted to swap out my m4 mini that I'm using for this kind of stuff right now!
