This is really sad news to me. I have used Pico several times as an example of good implementation and aesthetics, so the agent could get what I had in mind. In this case, applied human intelligence and good taste had their value when I needed to interact with the LLM. Now that intelligence begins to fade-out.

It continues to be a good example of those things even if it's not maintained

As someone who lives here, I’ve noticed a pattern that keeps happening in every election since social media became a big deal in the race. Shortly before campaigns start, the TSE—Brazil’s electoral court meets with reps from the social media platforms operating in Brazil to talk about the ‘rules of the game,’ algorithms, and so on. By the time the campaigns are underway, the rules are already set and agreements have been signed (literally). Still, there’s constant monitoring of results throughout the campaigns, along with efforts to better understand what goes on inside the platforms’ so-called ‘algorithmic black box.’ The rules they set, combined with the results they see, lead to a kind of ‘fine-tuning.’ You can see this in the fact that just this week the TSE asked for more confirmation and details about the algorithms, plus other explanations that, in principle, shouldn’t even have been necessary. It seems to me that this last-minute pressure is meant to get concessions out of these platforms by threatening them and/or to get key information disclosed to the Brazilian state, which could then use it to its own advantage later on. Idk, just my gut feeling of this.

Last year (2025), I lost count of how many times I tried using an LLM to set up Docker. It was still the era of prompting, then copy-pasting the response and seeing if it worked... and it almost never did. It’s great to see this new Docker agent. For me, it would be even more appealing to see an agent and a model specialized in Docker working together, capable of proposing advanced configurations and getting them working on the first try.

Modern models should be better with mainstream technologies like docker engine or docker compose, but still to try to help them we started to work on distributing skills ( https://github.com/docker/skills ) in addition of our specialized agent (gordon) integrated in various products.

I literally built this for self hosting apps at Tarvis.io

You can drop any GitHub repo and the agent builds and deploys the app to your Tarvis workspace along with taking care of ssl domain and any other configuration and making sure the app is healthy.

<>

Oh, thank you. We need that. It's refreshing to know that we might have the option to buy a cell phone with less than 6.5 inch size.

Is there a pure CSS or CSS + JS alternative that does not rely on React?

I wish. React + Tailwind as the commonly accepted default is too bad, it sacrifices portability for convenience. It's legacy code in the making, like jQuery and Bootstrap.

Haven't use it myself, but I've bookmarked https://basecoatui.com/ to try in future.

Yeah. Neobrutalism is mostly about the drop shadows + bright color palette. You don't really need a component library for what should be some basic CSS rules.

First thing I thought. Tying it to React is a shame.

The promise of the so-called ‘malleable computer’:

a machine whose hardware parts you could change, as well as bigger chunks of its software, above and beyond end-user apps. That is possible via the integration of the OS with AI agents that do the heavy lifting for you in adapting the OS to your ‘needs.’ Users of the Pi coding agent get a better feeling for this because the tool is, by design, built to respond and adapt to user-commanded changes. This is the kind of capability the typical user won’t see in private-source OSes like Windows and iOS, especially because they have security as a concern (thus the OP’s motivation for the heads-up on security in Omarchy).

Sidenote: the "telescopic zoom" feature of the website is very interesting and __useful__. It's amazing to see how one can still bring innovation to blogs/personal websites!

Thank you!

It's interesting to see in the post that the reading experience in the web page highly contrasts with the one in the eReader.

The website shows 15 words in the first line (10 in mobile). The eReader shows 6 words in the first line. In my own experience, I found that 6 word per line is the optimal number for comfort. (age thing? maybe)

Did the company fix the high friction between any service and their models' API?

I hope so. It seems mind boggling to me that an user needs to surf around different sections (plural) of google cloud console, then this Vertex and do a dozen clicks to issue a simple key.

You can use the Gemini API which is independent of the more complex Vertex AI API. Not sure whether you still have to visit the Google Cloud UI for some things (like billing) though.