henrik ingo ❤️🇺🇦 ❤️

Favorite part is when one of these quiet old men says: "we just accept everyone for who they are. It's nice how we all get along even if we are so different." While this cast wouldn't exactly pass the silicon valley DEI test, it's 100% true what he says. #bastutibasel #kaj

Cheap Shaheds caused billions in damage to US and Gulf assets. A $1 billion US early warning radar was hit. At least one $300 million air defence radar was destroyed. Both were clearly visible in satellite imagery for two months before being struck. 7/

@Sukkola Sinänsä hyvä tutkimus, mutta mitään kohua tässä ei ole. Journalistiikan opiskelijoiden poliittinen jakauma vastaa täydellisesti kaikkien humanististen alojen jakaumaa. (Joka taas on suurin piirtein sama kuin nuorten naisten jakauma.) x.com/h_ingo/status/…

Suomalaisista journalistiopiskelijoista 78 prosenttia äänestää vasemmistopuolueita (vas 44 %, vihr 20 %, Sdp 14 %). Asia selviää Nordic Journalism Students 2025 -kyselystä, jossa oli mukana Tampereen yliopisto. Suomesta vastaajia oli 400. Kyseessä on kattavin aiheesta tehty tutkimus. Tutkimusta ei ole julkaistu.

...Jos eivät tiedosta niin voisiko joku järjestää heille jonkun ystävällisen insinöörin tai tilastotieteilijän pitämään vierailuluennon aiheesta?

Tiedostavatko opiskelijat itse että heidän päivittäinen työympäristönsä on näin vinoutunut? Käyttävätkö he siksi ekstra energiaa ymmärtääkseen myös miesten, maalla asuvien ja amiksen käyneiden työläisten elämää?

Noniin, oiskohan aika taas rauhoittua. Tässä tutkimustuloksessa ei oikeasti ole mitään uutista tai poikkeavaa. Tämän ymmärtämiseen taas tarvitaan joku joka olisi opiskellut yliopistotason tilastomatematiikan perusteita. (Joka taas 80% todennäköisyydellä äänestää KOK tai VIHR.)

2.5 years ago, I started a SQLite rewrite as a side quest. Today I'm backporting fixes to a release branch to address issues customers are experiencing as they work towards running it in production. Reminds me of my Scylla days and the shift when you don't just build, but also start maintaining. That time when people gradually move from trying it out to actually depending on it. Turso is growing up!

@penberg And if you want to follow these results live, directly from CI, then here are all the tpc-h queries side by side turso vs sqlite: nyrkio.com/public/https%3…

How fast is Turso? There has been a lot of discussion about SQLite and Turso performance over the past few days. Now I have this policy of not making performance claims because I represent the vendor here, and people should not really pay too much attention to vendor benchmarks. However, because Turso got confused with another SQLite rewrite (which seems pretty slow), I wanted to share something to clarify the situation. Here are the Turso vs SQLite TCP-H results. TPC-H is an industry standard benchmark for ad hoc queries and concurrent data modification. I picked to show it because it represents different types of real-world workloads decently and I believe it represents Turso's performance state well. I ran the experimental evaluation on an AMD Ryzen 9 3900XT 12-Core Processor with Fedora Linux 6.18.12-100 and the 3.49.1 version of SQLite, comparing it with Turso 0.5.0. The bars in the graph represent different query types. The y-axis shows the time to complete the benchmark, with lower values indicating better performance. If you eyeball the results, you can see that SQLite has an advantage in many of the benchmarks, but not all. You can also see that some of the benchmarks are missing from the plot. That's because they're slow in both SQLite and Turso, and we've decided to omit them from the graph. In cases where there are performance differences, it's almost entirely due to query planner and query optimizer differences. When SQLite is faster, Turso gets something wrong or is missing an optimization. Where Turso is faster, we've identified a weakness in SQLite's query planning and optimized it in Turso. Overall, Turso performs decently well, but still has some catching up to do. Of course, if you are seriously interested in performance, you should ignore what I am saying and run the benchmark yourself. With Turso, we're taking this approach of making all the benchmarks reproducible. For TCP-H, you can find the instructions on how to reproduce here: github.com/tursodatabase/…

Who are all these people wanting to talk to me about Github and CI? I unfortunately don’t have the time to take these calls, but just use @useblacksmith and be happy

Jonas Axboe takes the Pro 1500 Pole for the first race of 2026! #RadicalCupNorthAmerica / #RadicalCup / #BeRadical / #RadicalCupNA / #BecauseRadical / @hankooktireusa / @radicalmsport / @sebringraceway

@TheRickWilson The one who is in possession of the tapes

OK, MAGA. Russia is giving Iran intelligence to help target American forces. Ukraine is offering to help defend American troops from Iranian drones. So tell me again: which one is the “enemy,” and which one is the ally?

If you ever wondered how the e-divisive means algorithm works (Matteson & James, 2014), the algorithm at the core of the automatic change detection feature in Nyrkiö: Denis Shchepakin has written a summary of the Apache Otava* implementation otava.apache.org/docs/math/ * incubating

Helsinki today. Candles for Ukraine.

@__vinks__ @Bhavani_00007 I have Borland JBuilder :-D

@Bhavani_00007 Borland C++

as a dev, what was your first code editor?

@kellabyte @adithyalennzer Much respect. The rest of us usually start with hello world in a new language..

@h_ingo @adithyalennzer Yup that was the first thing I built! I have a handful of Porcupine linearizability tests checking its correctness but I’m not sure how confident I am in it yet and not sure how good it is. But it’s working 😂

The Paxos vs Raft conversation is interesting.