I don’t think those are comparable. simonw's llm Has a python SDK, but it’s very much CLI first. Light LLM is very much about the SDK. You can wrap some agent SDK’s around it, like Gemini, but that’s for agents not work flows. I can’t really think of them as in the same category.
It shows how to use it async or sync, and even handles using async in a sync context.
It's hard to write a good CLI without also writing most of a Python API, and llm went the rest of the way by documenting it. I think llm has the best docs of the Python API of the three.
Workers is a v8 isolates runtime like Deno. v8 and Deno are both open source and Deno is used in a variety of platforms, including Supabase and ValTown.
It is a terrific technology, and it is reasonably portable but I think you would be better using it in something like Supabase where are the whole platform is open source and portable, if those are your goals.
Ah, in my case, I want to just talk to a video-editing Claude, and then a sys-admin Claude, and so on. I don't want to go through a main Claude who will instantiate these guys. I want to talk to the particular Claudes myself. But if sub-agents work for this, then maybe I just haven't been using them well.
I’m still using
- free g suite
- play music
- finance
- nfc wallet is just google wallet isn’t it?
- chromecast, video and audio-only
I guess play music is now YouTube music, and doesn't have uploads, so that can be considered dead, but the others seem alive to me.
There is way too little on that site about why we should trust these tools. No way I’m going to download and run stuff from some random source that I know nothing about.
How can users be sure that their inputs are not being saved somewhere on the backend? If I paste a bunch of content into these forms, and it inadvertently has some sensitive data in it, such as a credential or key of some kind, how do I know that isn't a data breech?
I would assume it goes without saying to be careful and never use credentials or keys in an online tool that way, mine or otherwise. I don't save data but its something I'll work on to ensure.
Can't imagine what kind of stuff ChatGPT has on us too :D
If you don’t already know the answer to that question (e.g. monitoring network calls in your browser), why would you trust an answer from the same service providing the tools?
This is why web tools are typically a last resort for me, after all local options have been exhausted (usually, I can't get them working because an old binary is broken, building the code is borked, the local tool silently fails while offering no feedback, etc).
It was nowhere near as bad as the giant gym chains or any other number of businesses.
I can think of a lot of membership and subscription services that have been far harder to cancel that I wish the FTC would do something about. A few extra clicks to cancel Prime is nothing in comparison to the gauntlet required to cancel some gym memberships. I remember a story where someone forgot to cancel their gym membership before moving across the country but the gym's policy required that you cancel in-person at the gym. They had to pay the monthly fee until their next trip back home, then lose an hour traveling to the gym to fill out the cancellation paperwork.
They did that. The FTC had a simple click-to-cancel rule that was supposed to start enforcement this year. It was struck down on procedural grounds during appeal in July.
Old situation: "Android drivers" are technically Linux drivers in that they are drivers which are built for a specific, usually ancient, version of Linux with no effort to upstream, minimal effort to rebase against newer kernels, and such poor quality that there's a reason they're not upstreamed.
New situation: "Android drivers" are largely moved to userspace, which does have the benefit of allowing Google to give them a stable ABI so they might work against newer kernels with little to no porting effort. But now they're not really Linux drivers.
In neither case does it really help as much as you'd hope.
