Łukasz Tyrała ✏️

@lukasztyrala
34 Followers
171 Following
35 Posts
I design information • interfaces • interactions ✏︎ Principal Designer at Grafana ❦ UX Educator & Expert at SWPS and AGH universities ❡
Websitehttps://lukasztyrala.pl

I’ve been working on this for a while, but let’s make it official: I started a little Tumblr-like microblog about software craft and quality!

You can sign up via RSS or a weekly newsletter digest. There’s already almost two months of content in there, if you just want to check it out.

Hope you like it!

https://unsung.aresluna.org/

Unsung

A blog about software craft and quality

I played Saper on 286 with Win 3.x during a COP Museum visit during last holidays. Muscles still remember all the mouse clicks and shortcuts.

Tech and computers are cool. We must take all this back from weird billionaires, monopolies, and bros.

(Photo by https://tomekszkodzinski.com/)

Into the woods.

Exclusive and just for you, my Mastodon friends: I’ve opened a few more TestFlight spots for my upcoming Zettelkasten app, Zettel.

https://testflight.apple.com/join/G2C1wGC4

It feels a bit like shift from outcomes (or at least the framing that it’s about outcomes) to outputs again. This time though, the output MUST be AI-something.

Most employees wil be more successful doing mediocre work using AI than doing good work that cannot be framed as AI-powered.

And it is not a first time that output done using a specific tool wins over favorable outcome.

Ghimbli style is buying time for agentic AI that is buying time for AGI that is buying time for robots with AI that is buying time for quantum computing—and than finally we have a new hype cycle! Quantum!
I used – and — before ai made them popular.

New rule: if you generate something using “ghimbli” you’re obliged to watch one Miyazaki’s movie per prompt and reflect on humanity and life.

M*%!^$k fans and distant-➡️ voters are prohibited from using such prompts due to direct conflict of interest.

Proceed as advised.

ARC-AGI are fun puzzles to solve. There is no doubt that AI companies will throw resources at tweaking models to get better at the puzzles. As it happened with chess and go.

That is not the point though, as the real intelligence and creativity test is humans consistently coming up with fun puzzles hard for AI, and other humans pursuing algorithms that get better at those novel puzzles. (The former is more impressive than the later.)

Puff, I would,
But older now.
I just admire. So cool.