Mattias Petter Johansson (MPJ)

Mattias Petter Johansson (MPJ)

@mpj.fff.dev

Data Engineer 💛 Creator of the YouTube show Fun Fun Function, writes weekly data development chronicle: funfun.email 🦁Previously @ Spotify & parity.io

255 followers474 following92 posts

I am reading Daniel Levintins excellent book ”The Organized Mind” and this part perfectly explains why I sometimes get stuck in multi-hour obsessive TypeScript rabbit holes to create the perfectly (unnecessary) typed expression for a problem.

God I love force directed graphs

Peter@pjpscriv.co.nz

📊 Recent little Observable play - creating an org chart viz using a force directed tree that is zoomable, pannable, collapsable and has a tooltip observablehq.com/@pjpscriv/or... Not displaying any interesting data atm, but certainly fun to play with!

💛It’s seldom one gets shoutouts for commercial partnerships, but @augmentcode.com really is that good - it’s really an incredible thing that a platform targeting large legacy codebases is free for OSS thx @adrienjoly.com

Adrien Joly@adrienjoly.com

And I would like to also thank @mpj.fff.dev for having inspired me to try it out, and for making awesome videos that I always enjoy watching! <3 If you want to check my changes and the prompts I used, here are the changes I just deployed to openwhyd.org: github.com/openwhyd/ope...

Was sent this golden thing from 3 months ago written by @peck.phd

Evan Peck@peck.phd

Trying something new: A 🧵 on a topic I find many students struggle with: "why do their 📊 look more professional than my 📊?" It's *lots* of tiny decisions that aren't the defaults in many libraries, so let's break down 1 simple graph by @jburnmurdoch.bsky.social 🔗 www.ft.com/content/73a1...

Mattias Petter Johansson (MPJ) reposted

Me in the WSJ today, arguing that the relationship between social media and mental health is not clear cut, and bans for teenagers aren't the answer. (NB I did not pick this title, I would have gone for something kinder. I understand why people are worried.) 1/ www.wsj.com/health/welln...

wsj.comEssay | Stop Panicking Over Teens and Social MediaModern life is digital. Adults need to help young people navigate the costs and benefits, not launch bans and hope for the best.

Mattias Petter Johansson (MPJ) reposted

How many % of the currently experienced value of large language models come from the fact that they ingest in a way that a traditional search engines would have been sued and blocked to oblivion if they did? And how expensive would they be if they paid fairly for the privilege?

Reuben Binns@rdbinns.bsky.social

Every major UK newspaper - left, right, broadsheet, tabloid - is running a front page campaign against copyright carveouts for AI training. I can't remember the last time any political cause had such broad support, let alone one so anti-big tech and so economically consequential.

Mattias Petter Johansson (MPJ) reposted

If you're into D3 or Observable, there is no doubt you ran into Fil at some point. I learned a ton from him, and this interview only adds to that

I'm super duper proud to officially announce the FFF partnership with @augmentcode.com, the developer AI that I've been using as my primary for the last few months. Go give it a whirl using fff.dev/augment and let me know what you think!

Augment Code@augmentcode.com

Good monday morning! New @mpj.fff.dev interview with Augment Code co-founder Guy Gur-Ari just dropped :) youtu.be/LEOSWYlQyVw?...

Mattias Petter Johansson (MPJ) reposted

New article by me! It's a common mistake to think the TFR is a prediction of the average number of children women will have over their lifetimes. But it's not. The TFR can decline even when the eventual number of children per woman stays constant or even increases; the opposite can happen too. 🧵

Very exhaustive test of Claude interpreting graph, fun to read!

Michael Friendly@datavisfriendly.bsky.social

📊 #dataviz #AI New @datavizsociety.bsky.social article just dropped: What Does AI Understand About a Graph? (coauthor: Claude Hiaku) A sort of Turing Test applied to the problem of graph comprehension,but a bit more like a cultural anthropologist meeting a new species. bit.ly/4hHOUeO

So it is trained on your uncle on Facebook?

Clare Spencer@clarespencer.bsky.social

BBC data scientist Oli Elliot analysed the news-related output of co-pilot, ChatGPT, Perplexity and Gemini and found 19% of AI answers which cited BBC content introduced factual errors including incorrect factual statements, numbers and dates. www.bbc.co.uk/aboutthebbc/...

Just because you can, you should

André (makerbymistake)@makerbymistake.com

I documented my setup to get Ollama, Open WebUI, and code suggestions in VS Code using a Raspberry Pi 5. You can use this for any Linux computer. dphacks.com/2025/02/09/r...

This was much news to me. Lots of interesting aspects with this article, but the detail that concept cells fire only after a 300 ms delay, for some reason, stuck out for me.

Mattias Petter Johansson (MPJ) reposted

Saturday ice update - #Arctic sea ice extent is currently the *lowest* on record (JAXA data) • about 790,000 km² below the 2010s mean • about 1,450,000 km² below the 2000s mean • about 2,040,000 km² below the 1990s mean • about 2,430,000 km² below the 1980s mean Plots: zacklabe.com/arctic-sea-i...

Really looking forward to listening to this new investigative pod series by @suelinwong.bsky.social. I was made aware of this fascinating (and unnerving) phenomena through @pjvogt.bsky.social on the episode pjvogt.substack.com/p/whos-behin... and this seems like a much bigger deep dive.

bsky.appbsky.app
Sue-Lin Wong@suelinwong.bsky.social

For the past year, I've been investigating the global rise of online scams - a predatory, multi-billion dollar industry that rivals the size of the illicit drug trade. Scam Inc, a new 8-part podcast from The Economist and hosted by me, is out in February: econ.st/3CoAdxJ