Jouni Seppänen

@valigo Emacs litters your directories with tilde-suffixed backup files, and the command to delete all backups is rm *~. Don't accidentally type a space between the symbols.

If you ever accidentally create a file named "~" in a local directory, PLEASE, do yourself a favor and delete it through a gui file manager. DO NOT try to be a hero. Don't ask me how I arrived to this conclusion

@AliceFromQueens One of the most loved Finnish rock songs is such an over-the-top parody of Finnish literature that you can't help but love it lyricstranslate.com/en/murheellist… (Pick English translation #1, the latest one is poor quality)

Great parodies, like this, aren't great because of a concept. They're great bc they nail the genre's details so precisely & channel its spirit so sympathetically you feel the parodist must actually love the genre. If a parody hit can't pass for a real hit, it isn't great.

@HouseChambers @AliceFromQueens Like  a hammer hits a cantaloupe!

@AliceFromQueens All the best parodies are by people who genuinely love the source material. Whether music, films, art whatever. "5, 4, 3, 1....Off Blast!" (Let's see how many people get that ref)

@almighty_lisp eww! gnu.org/software/emacs…

ADD TYPESCRIPT TO THE BROWSER: 👎 ADD LISP TO THE BROWSER: 👍 REWRITE THE BROWSER IN LISP: 💪

@vboykis There is github.com/leopiney/linus…

What I most desire doesn’t exist on the market yet: a coding agent model, but it’s only been trained on code reviews by Eastern Europeans.

@Tommi_Kinnunen Niissä on sellainen kirous, että kun toteaa että M3.5 on niin erikoinen ruuvi että heitän sen menemään, miksi minulla onkaan ollut tämä kymmenen vuotta, niin heti seuraavalla viikolla pitäisi kiinnittää ties mikä fillarin tarakka jossa osoittautuu olevan M3.5-kierre.

Olen iän myötä alkanut kerätä talteen myös puutavaran pätkät ja rakennuslevyjen palat, sillä koskaan ei tiedä, milloin tarve tulee. Nyt pelkään, että tämäkin kokoelma on vajaa eikä tarjoa turvaa tosipaikan tullen.

Jokaisessa kodissa lienee muovilaatikko, johon on säilötty ruuvit, prikat, koukut ja muut, jotta on sitten kun tarvitsee. Tänään järkytyin, kun en löytynyt kahta kahdeksan millin mutteria, vaan piti lähteä ostamaan. Varautumiseni maailman yllätyksiin onkin liian heikkoa!

@jhietaniemi @googlemaps It's a real placename but the etymology is not certain. Probably Sámi origin for some parts of the name. fi.wikipedia.org/wiki/%C3%84ter…

@googlemaps Completely bogus, prank by some Finn that probably found this unbeliavably funny. Sorry to be the killjoy.

The Finnish region, Ateritsiputeritsipuolilautatsijänkä, roughly translates to “the bog on which the storage hut standing on a single pillar and belonging to Paul, the son of Pete', the son of Andrew, stands/stood on" 🇫🇮

@AdamCSharp The Finnish word for cathedral, tuomiokirkko, means "judgment church", from the Swedish domkyrka, ultimately from the Latin domus ecclesiae. The Swedish dom also means "judgment" (cognate with "doom") but the Finnish tuomio has only that meaning.

@AdamCSharp The Finnish word for dragon, lohikäärme, means "salmon snake". Which doesn't make any sense until you realize it's from old Swedish floghdraki where flogh means "flying" (and I suppose is cognate with the English word) and draki was translated as snake.

The Finnish word for an auction, huutokauppa, means “shout market” (or “shouting shop”) * read on for more of my favourite Finnish words (a thread in honour of today being Day of the Finnish Language)…

@_chair Weird? When my billing doesn't reach 5 € in a month, it gets bumped into the next month

the Hetzner Cloud Services guide to losing a customer over four dollars and 52 cents (that any other company would have just autobilled)

@alphafox I'm sure I played it in early 1995, so it must be at least 31 years old

Descent turns 30 years old today! This innovative 3D shooter will actually play on a 486 with 4MB of RAM - talk about efficient memory usage! (I tried it)

You might have seen the WuBOT performing at the 2026 Spring Festival Gala; however, most high-dynamic extreme motions you see are executed by overfitted tracking policies. Until now, training a unified policy capable of performing various extreme motions with a high success rate remained an unsolved challenge. We spent an entire year digging into the barrier between general tracking and extreme physical behaviors. After burning through dozens of G1 robots, we finally identified the bottleneck of learning and physical executability. With these discoveries, we developed OmniXtreme: the first general policy that can execute diverse extreme motions, including consecutive flips, extreme balancing, and even breakdancing with rapid contact switches! This capability is achieved by pre-training a flow-based generative control policy and then post-training with actuation-aware residual RL for complex physical dynamics—a step we found critical for successful real-world transfer. This work is a joint collaboration with @UnitreeRobotics. Together, we are pushing the physical limits of humanoid robots. It is incredibly exciting to see a general "robot gymnast" and "robot breakdancer" come to life! It was also our first time publishing a paper with XingXing, which was an enlightening experience. The model checkpoints are now released—we welcome you to play with them! 📦 📄 Paper: arxiv.org/abs/2602.23843 🌐 Project: extreme-humanoid.github.io 💻 Code: github.com/Perkins729/Omn…

@OUusi @juhakala3 Muistan nähneeni noista yhden ja ajatelleeni, että Puuilo on keksinyt uuden mainoslauseen.

@juhakala3 Ei välttämättä ihan paras strategia.

Huomenta. En nyt halua mainostaa mutta nyt on mainostoimistossa välähtänyt älynystyröissä 🤭

@keysmashbandit Not usually a fan of Grokipedia but it does much better here. (Further down the page it becomes unreadable because of LaTeX errors.)

This is ridiculous and would be 100% unpublishable if it were in a textbook. Contender for the most messy incoherent Wikipedia introduction standing, but it's in a tight race with every other (mathematics) article

@bfitzx @doffcocker Apparently the TNG remaster was a financial failure so they aren't likely to do DS9 and Voyager. There is a DS9 fan project ds9redefined.wordpress.com/about/

CBS spent $12 million remastering Star Trek The Next Generation

@euxoa @dioscuri New opportunities for Batesian mimicry through hash collisions with dangerous species

@jks @dioscuri that and a 32-bit hash, and if there’s a collision, well who said biology has no exceptions

I think we really need some kind of criminal trial for the people who decided hominina, hominini, homininae, hominidae, and hominoidea should all refer to different taxonomic ranks of apes 🤬😡

@euxoa @dioscuri Agree that IP-like codes are a stupid idea. OIDs are much more extensible. Or why not go full LDAP dn: subspecies=apus,species=apus,genus=Apus,tribe=Apodini,subfamily=Apodinae,family=Apodidae,superfamily=Apodoidea,suborder=Apodi,order=Apodiformes,class=Aves,kingdom=Animalia

Or, from Gemini: Order: Apodiformes Suborder: Apodi Superfamily: Apodoidea Family: Apodidae Subfamily: Apodinae Tribe: Apodini Genus: Apus Species: Apus apus Subspecies: Apus apus apus (Common Swift) IMO the criminal investigation should be directed towards those who (1) describe new genera, for that often changes inflection of the species name too, so search etc. impossible; (2) oppose replacing this mess with numerical IP-like codes.

@htmx_org odds the product team gets bonuses based on time spent on site?

odds the "product team" ever hears about this? odds the "product team" ever asked any users if this was a good idea?

@emollick There's at least evidence.guide, 20¢ per paper, import into an R data frame to get started with your meta analysis

GPT-5.2 Pro is good enough to check reproducibility & robustness of academic papers across many fields (given the data, can you get the same results? are the statistics brittle?). At scale, this would have a big impact. It can't do an independent replication with new data, yet.

@cagdasyelen @paulg And not all text it's trained on is from the Internet. I met a guy who writes high-school/college level math problems and solutions in Finnish for training data. I assume someone else is writing emoji-riddled "diversity isn't just a buzzword—it's our greatest asset" corporatese.

@paulg This is most likely because of RLHF (post training). Humans probably vote up answers that are bulleted, easy to skim and hyper-safe. Models aren't mimicking humans, they are optimizing for a reward function based on clarity/readability.

Presumably ChatGPT is trained on corpora of things written by humans, and yet it doesn't sound like any human I know. Is there a population somewhere of people who write in chirpy bulleted lists that I've somehow managed to avoid?