Samo Korošec

It’s not surprising or bad that this frustrates people for whom more is at stake than content. Asking them not to object that their names and work have been appropriated to sell products that purport to replace them, and that without even grasping what they do, is unreasonable and unintelligent.

The only silver lining I see is that people are realizing that LLM-based AI is a tool, yes, but also a _shitty_ one at that. And this, in no small part, thanks to Google et al pushing it as a search engine and Microsoft et al pushing it as a "business assistant".

And people _hate_ shitty tools.

Screenshot aus der Presse: Pröll: "Österreich braucht mehr Steinberger und weniger Schrems"

Max Schrems ist ein Jurist, der mit seinem Einsatz für Datenschutz- und Grundrechte international höchst erfolgreich war. Peter Steinberger reitet eine Bubble, die uns derzeit alle mit einem ganzen Bündel gefährlicher Probleme konfrontiert. Österreichische Medienpolitik so:

the most thoroughly vindicated man in hollywood.

trade disputes and conspiracies ✔️

republic subverted over sex hangups ✔️

“sand people” killed, royalty indifferent ✔️

naive idiot casts vote for tyranny ✔️

villains with silly names like “dooku” ✔️

powerful clerical order screwing up ✔️

list goes on

Klicks + Werbung = Geld für die Zeitung

Narzissismus? 🤷‍♂️

(Erklärt so vieles diese Tage.)

Slovyansk. Four FAB/UMPK glide bombs on the city centre. Europe, weak and fearful beyond any measure, looks on and keeps crapping its trousers while serving more bowls of word salad.

We’re now at a point where the Donald is happy if Iran removes the mines it may have placed in the Strait of Hormuz. Everything is going according to plan.

(I tried to make a joke so lame even AI could have came up with it.)

Fun fact: If you take one CENT out of AGENTIC you get AGI.

🤯, eh?!

everybody takes the elevator to the bright, airy 30th floor where you can't hear Ed Zitron yelling on the street corner

swipes their pass and hangs out at the water cooler coo'ing AGENTIC to each other in ASMR voice

A screenshot of the latest iteration of the upcoming https://elenajs.com documentation website.

Ok, think I’m ready to hit the publish button *soon* with this thing.

Neat! 🎉

Join us and shape how users experience our brand and products on web. Design intuitive, high-performing web journeys that turn complex geospatial tech into clear, engaging experiences. 🌍 Ready to design web where brand, UX, and tech meet? 👉 Apply here: maptiler.link/3MRENdh

Give your dashboards or delivery apps a professional polish with OpenStreetMap Dark. Built on the OpenMapTiles schema, it provides a sleek, high-contrast backdrop that fits perfectly into any modern night-mode interface!

Read more: maptiler.link/4b9H2SP

“How to tell your designer to make the logo bigger for the 4th time without hurting their feelings.”

(Suggesting this as a classically trained graphic designer. 😬)

Plus it's fun. Working with a human illustrator is so fun once you know how it works.

I think we need to do a little guide for how to work with a freelance human because a lot of people just don't know how, and a lot of the benefit of LLMs is avoiding the fear and discomfort, more than the $$.

Unacceptable!

At some point someone will have to sit down with wars, in general, and have a serious talk about their behavior.

Yeah, that’s pretty much what I am talking about. Ideally it will work as a pretty spectacular self-selection mechanism, when all is said and done. 🤞

Sound like my ex-boss. Who was also big on saying “Own your mistakes.”

(Dude “lost confidence” in working with me when telling me to own my mistakes in a call and I went through all the issues and Slack messages, with a shared screen, to look for his non-existing instructions on what to do.)

I’m kinda crossing my fingers on them blowing their own legs off with it, myself. 🤞

The only really positive thing for me is how it unmask so many people for what they are.

And the icing on the cake is how most of them are “proving” how awesome AI is by making trivial shit that no one would and will never pay them for. Literally hyping themselves out of future work possibilities.

Fail fast, fail often, but at least fail loudly! 💪
— their office motivational poster, probably

It’s a bazooka for people who are insecure about their incompetences.

That's why a lot of the impact is to extensively tarpit people who do the work and know how to do the work. Which is extremely counterproductive and rapidly burns those people out.

e.g. In the past 24 hours have watched how this is evicting the most valuable contributors from open source projects.

Can companies be trusted enough to still deliver quality and due dilligence when pressured to be ever more efficient?

For a general idea, one might ask the FAA if they trust Boeing.

The problem: This means regulatory bodies can be spammed by submissions that they don’t have the capacity to review, and that all _look_ competent. But it would be up to the companies, all under different market/investor pressures, to make sure the generated documentation is truthful.

In practice, this could mean hundreds of requirements, system specifications and even verification/validation test specifications that ”people who know what they’re doing” would review and sign, all generated within hours. Very efficient!

I worked on tools for regulated industries that pivoted hard to AI. Medical Devices were the first vertical.

The pitch _absolutely_ was to “ease the regulatory and documentation burden“, implying AI will generate any “missing configuration items” and offer them for quick review & signing-off.

Modern CSS Features You Should Know About YouTube video by Joy of Code

I went through 20 modern CSS and UI features that landed on the web platform from CSS Wrapped 2025

youtu.be/KGc5DYl6Lbs