It's useful because you can write back-end and front-end code with the same language. Code sharing should work, as long as you don't use runtime-specific features blindly. All basic language features seem to work the same regardless of runtime.
On my system it asks between a few different public registries, and dockerhub/docker.io is one of the choices.
t's all public infrastructure for hosting container images, I don't think Docker-the-company minds other software interfacing with it. After all, they get to call them 'Docker images', 'Dockerfiles', and put their branding everywhere. At this point
I've been using https://readest.com/ lately. It's FOSS and just recently got this feature. The TTS voices are pretty natural and text is highlighted one sentence at a time. Plus the design of the product is great.
I tend to agree that ActivityPub seems pretty horrifying to work with (as an outsider). At least it works, though, and is implemented by a large amount of different projects and not owned by any single one.
> seems pretty horrifying to work with (as an outsider).
As someone who both made my own implementation + hacked on others, what was/seemed to be the horrible parts? It's a pretty simple standard that is basically RSS with some added stuff (very simplified of course, before I got jumped) for facilitating the federation parts.
Ollama allows me to use a single podman command, which uses the latest version of ollama, downloads a model of my choosing, and starts a local http endpoint widely supported by different clients. I can just run this one command to chat with a local model through a web interface, get code completions in VSCode, ask about the content of my local Markdown notes.
Now, I don't use AI that much, I could totally live without this. But if it weren't for the robust one-liner I probably wouldn't use local LLMs at all.
I think in an ideal world you could groom your audience to your liking! Personally, if I could suspend my disbelief, I'd love an army of virtual techies telling me how clever my ideas are. Others might prefer a legion of other super moms praising their selfless parenting when no one else does, or like you were getting at, maybe just a peanut gallery!
Image gen on-device might be tough, we'll save that for the 1.1 release ;)
Even humorless curmudgeons (like me) can get in on the action with a huge array of people to disagree with, act snobby to, and never before heard of stupid opinions!
I'm not sure if you are sarcastic, but your definition of "ideal world" feels very different from mine. What you describe sounds as close to the scifi concept of wireheading as I can imagine to get without brain surgery.
I'm being tongue in cheek, for sure. Let's pretend I said "ideally" instead, as in, "ideally the product would function this way".
I admittedly wouldn't be the target audience for this concept, nor would I take up wireheading. When I see the huge success of Character.ai or the above SocialAI though, I'm really convinced there's a market segment of people younger than us who get a lot out of communicating with virtual "friends".
Definitely like DuckDuckGo - I have recently moved from Google to DDG as I noticed it serves the same, and even better plus privacy-first, results. Yet what you provide is not quite the same: An AI Chat Website vs a Browser Extension (Orbit).
When the Bing API went offline a short while back, the DDG website was a blank page. No fall back, nothing. A Bing wrapper with ‘trust us’ privacy marketing.
How often does the Bing API go offline? Meanwhile, every single google search I carry out includes, right at the top, 'AI' generated nonsense which is sometimes wildly inaccurate. DDG never does that.
