Tell me about it... that infinite Ctrl + Shift + L sequence circling through all credentials from all subdomains. Then you brain betrays you making you skip the right credential... ugh, now you'll circle the entire set again. Annoying.
Sounds more like a lobby thing. Once a government finds a new "recipe" to be worked out with global vendors, meaning, a new way to allocate budget with a strong social justification (e.g. protect children, fight terrorism etc.), governments from other nations jump into the matter and literally copy/paste it locally. In short, whoever comes up with a creative idea to allocate public budget will serve as the basis for others to copy.
I'm writing my first book now. It's a novel aimed at teenagers and young adults, in a technical format similar to "The Phoenix Project" by Gene Kim et al., if you're familiar with it. It explores FOSS, non-proprietary file formats, digital preservation, cryptography, and the concept of freedom as a whole. I resonate with the author of the article who discusses motivation to write and the "existential crisis" that comes and goes almost every day. I've been fighting those negative feelings by adopting the mindset that I'm writing the book for myself. It's a book I've always wanted to read, which I can then lend to my teenage children so they can read it as well. Everything else (commercially speaking) will be a nice consequence of this endeavor.
Amazing: the website's index page has the book's index in it. While this makes perfect sense, it's a kind of a feature that is becoming rare in today's tech book websites which display all sorts of marketing fluff, social confirmations etc and not the structure of the book itself.
I first read about Unpoly in Stephan Schmidt's Radical Simplicity website[1], I liked it's value prop and decided to try it. I found it bit too complex just as you said. A long time later I came across htmx and decided to try it even after reading a side comment that the library was "like unpoly". 15 minutes later I had a simple to-do list running htmx ajax calls using php and mysqlite in the backend. It was so easy I could not believe such thing could exist. Then I decided to read the Hypermedia book and never stopped using htmx in my projects.
Using metaphors is dangerous, but I would dare to say that big tech AI is like cement suppliers. It's too low level of a service. In civil engineering you have the option to contract value added suppliers that will give you prefabricated pieces in concrete or steel you could be using to build your construction.
I'm seeing a lot of AI firms building value added services on top of big tech "foundational" AI offerings. Value addition can start very early at a clear plans/billing structure, going through rate limiting, documentation and extra features that will bring stability or consistency to our AI enhanced products.
Going the other way around (I tried) and building things on top of big tech AI is challenging starting at the fundamentals as the OP described well.
Openrouter is roughly at that level of value-add. With plenty of competition now, since being able to charge 5% on your AI spend just for having sane billing, spending controls (actually enforced per-api-token budget limits!) and easy sign-up is an insanely profitable business proposition
On the other hand I think it's fair to criticize the model hosts for not offering the same
(I work at OpenRouter) Certainly for individual developers / hobby projects that's the primary value prop; super easy access to all of the models.
But there's a lot more functionality that becomes relevant when building in production. We do automatic fallbacks, route between providers based on data policies, syndicate your data to agent observability tools / your logging platform of choice, user-level and api-key-level budget management and model allow/block lists, programmatic API key management, etc, etc. More good stuff shipping all the time!
reply