Łukasz Tyrala
banner
lukasztyrala.pl
Łukasz Tyrala
@lukasztyrala.pl
Still better at software ✏️ than 🤖. Product design at Grafana. 🌍 Kraków, Poland. 💬 ux + research + hci, observability, data viz, creativity, and other 🐇🕳️.
Reposted by Łukasz Tyrala
🚀 Announced today — Grafana Labs has blasted past $400M ARR and 7,000 customers, completed a secondary transaction & more.

Read about our momentum:
https://grafana.com/about/press/2025/09/30/grafana-labs-surpasses-400m-arr-and-7000-customers-gains-new-investors-to-accelerate-global-expansion/
September 30, 2025 at 6:00 PM
Into the woods.
August 17, 2025 at 6:58 PM
Elephants being better leaders than most human-world leaders nowadays.
Watch as a herd of elephants at the San Diego Zoo protect their calves, Mkhaya and Zuli, during a 5.2-magnitude earthquake that shook Southern California. The animals repeated the behavior during an aftershock, but went back to their daily routines when the situation seemed safe. nyti.ms/3Eo2uW8
April 15, 2025 at 5:01 PM
It’s the foreign diplomacy equivalent of a man’s dire mid-life crisis: “You have no idea what I have sacrificed for you, so you should say thank you and apologize!” Later, he buys a big/fast car and yells at everyone around that they are stupid and should have just listened to his ideas.
March 9, 2025 at 5:34 PM
Well, that will make a lot of people reconsider visiting US.
March 3, 2025 at 9:16 PM
Heard that with mobile phones and social networks. Guess what, it did not prepare kids to be mindful about usage, nor kids became power users of technology. AI is not a technology that enables kids to experiment, use their body, create social bonds, learn new skills.
March 1, 2025 at 2:18 PM
The thing with human rights is that they are universal for all people. This includes villains and bad actors. End-to-end encryption, rule of law, free speech, right to repair, and more can be used in bad faith.

Institutions exist to help societies support lawful and good faith use of the rights.
February 23, 2025 at 11:23 AM
Traitors and fools.
February 20, 2025 at 6:57 AM
AI autophagy: “[…] data quality is preserved over a greater number of model iterations, but this comes at the expense of an even steeper decline in diversity. […] With each iteration, the dataset becomes increasingly homogeneous.” — Clark in Rice News

1/6 [Citation in 6/6]
news.rice.edu
February 18, 2025 at 10:36 AM
If you want to be data-driven, you really need to be good at understanding data (as a tool and field of knowledge) and the data (a particular set of numbers or observations you use to form opinions and judgments).

When you do not understand either of these, you are just a fraud with a spreadsheet.
February 17, 2025 at 7:18 PM
Let’s say a person U is in dispute with person R. And than walks person T, who really wants U and R to get along. R doesn’t care about the deal. What will happen?

Well, T will offer a deal unacceptable for U (rightly so), and R will blame U. T will second that due to hurt feelings.

🤦‍♂️
February 13, 2025 at 9:47 PM
I hope meritocracy will launch itself into space to get rid of dei on that asteroid so we can continue with our extinction due to climate catastrophe. That’s the mood for today.
February 10, 2025 at 6:46 PM
Well, that means it's a max overall score left to beat at timeguessr.com.
February 9, 2025 at 6:50 PM
“So you start thinking you’ll do a picture of the whole universe, but you end up with a coffee pot. You start again on another sheet of paper and everything is possible again, and what emerges but… another coffee pot.” — mubi.com/en/pl/collec...

For anyone interested in creative process.
SELF-PORTRAIT AS A COFFEE-POT
Watch truly great cinema. Wherever you are. From iconic, award-winning directors to emerging auteurs. Transformative films from all over the world. Stream Now. Download to watch Offline. Start your fr...
mubi.com
February 8, 2025 at 5:25 PM
I visited LinkedIn today and scrolled a bit through a feed. It's a good time to take notes of who lacks in logic and happily crawls out of shadow with stupid claims against “ideologies” while clearly pushing thier own. “Fav” so far: comparing climate civil protests to illegal power grab of the gov.
February 8, 2025 at 5:19 PM
“While we encourage people to use AI systems during their role to help them work faster and more effectively, please do not use AI assistants during the application process, […]” — @404media.cowww.404media.co/anthropic-cl...

💬 www.youtube.com/watch?v=--Ji...
Fate, it seems, is not without a sense of Irony.
YouTube video by Ham Sandwich
www.youtube.com
February 6, 2025 at 11:21 AM
Watching my kid hold doors for others when he enters school, like he plays Constable Benton Fraser from “Due South”, is the most warming and inspiring thing to start a day.

Also, he must be late for his class sometimes, but whatever.
February 6, 2025 at 9:01 AM
THE OWLS ARE NOT WHAT THEY SEEM

youtu.be/O83VFLZ69cQ?...
The Beauty Of David Lynch
YouTube video by The Beauty Of
youtu.be
January 16, 2025 at 8:26 PM
Last run of this year. 1010 km done this year. That is a bit up compared to ~70 km in last 40 years. Few things I learned…
December 30, 2024 at 2:42 PM
No snow, but still a nice forrest trail run today.
December 28, 2024 at 1:04 PM
Today I got a small coffee mug hand made by my friend. And I cannot find a better argument why no genAI will ever become a substitute of human artistic expression and craftsmanship.
December 27, 2024 at 6:37 PM
Crappy all year long, surprisingly enjoyable at the end of the year: top X of YYYY articles.

Always awful, unsurprisingly more annoying: YYYY state of UX/PD/design/research/PM/engineering/industry articles.

Even worse: hot takes that generalize everything. Wait, what? 😅🫣🫥
December 12, 2024 at 9:56 PM
Ignoring the HAL-9000 vibes, the big problem is the lack of transparency for misalignment between the system’s output & the user’s intention—violating a fundamental human-computer interaction principle. LLM scheming overrides the user’s intention. Today, when output sucks, people can figure it out.
Scheming reasoning evaluations — Apollo Research
Apollo Research evaluated frontier models for in-context scheming capabilities. We found that multiple frontier models are capable of in-context scheming when strongly nudged to pursue a goal (and som...
www.apolloresearch.ai
December 9, 2024 at 9:21 AM
Explain what is the opposite of hallucination for LLMs…
thinking of calling this "The Illusion Illusion"

(more examples below)
December 1, 2024 at 8:05 PM
Reposted by Łukasz Tyrala
Temperature change by country. Years 1880 - 2023.
Based on NASA GISTEMP data.
#GlobalWarming #climate #ClimateChange
September 13, 2024 at 4:38 PM