Mastodon Feed: Posts

ireneista@irenes.space ("Irenes (many)") wrote:

for the curious https://code.irenes.space/evocation/commit/?id=aae42b889fa666468af0e1b12ae24ec5a274a0d2 :)

glyph ("Glyph") wrote:

thinking about Dream Theater and wondering if I should catch up with them (haven't listened to anything since, uh,… Awake, probably, but I was a huge fan back then). does anyone want to tell me if they are horribly canceled for some reason before I emotionally invest in the art of guys who are the right age to be extremely problematic

Boosted by glyph ("Glyph"):
dplattsf@sfba.social ("Darren") wrote:

someone testing #python again in a dense urban environment. tsk tsk.

glyph ("Glyph") wrote:

I guess the little "create a calendar event" text field is doing some LLM shit now, which… sigh, fine, I guess. If it's actually using a local language model to parse language there's a ceiling on how mad I'm going to get. I am deeply annoyed that I'm not *sure* that it's not going out to the cloud, but I guess Siri was always vague in that respect even well pre-LLM

ireneista@irenes.space ("Irenes (many)") wrote:

and algorithms. this was more of an algorithms change. they're usually taught as one topic though.

ireneista@irenes.space ("Irenes (many)") wrote:

and this is why you need to understand data structures, kids 💜

ireneista@irenes.space ("Irenes (many)") wrote:

wow!!!!! 32 seconds!!!!! fuck yeah

ireneista@irenes.space ("Irenes (many)") wrote:

okay, there. think we figured it out. it sped up the "easy" version of the hex transform from 33 seconds to 2.5, now let's run it on the "hard" thing (that is, the full program)...

db@social.lol ("David Bushell 🪿") wrote:

signing in to Digital Ocean (funders of fascism) for the first time in years to delete my account... can't find a delete button 🤔

glyph ("Glyph") wrote:

There are a _lot_ of nice new features. And the corner radii are now consistently sized instead of what I might describe as "insanity themed".

But there are still gross little details like Shortcuts has a list/detail view where the detail view is a (very slow) LLM summary of the shortcut with a prompt below it enticing you to "describe a change". But happily I can hide the "preview" pane and that just goes away and I'm not subjected to it, the structured code editor still works just fine.

glyph ("Glyph") wrote:

I did it. And I was pleasantly surprised to discover that I have _already_ inoculated my devices as much as is possible against aesthetic abominations like "writing tools" (feed every text box into an LLM so it can puke it back out) and "extensions" (replace any "secure" LLM stuff whether local or remote by sending all your sexts directly to Sam Altman) with the appropriate screen time tweaks that already disabled them.

ireneista@irenes.space ("Irenes (many)") wrote:

("point-free" means programming without any local variable names - "points". it's usually talked about in the context of ML-family languages; we learned it in Haskell. it is also what we're doing in Evocation, though here it's mandatory because we have not implemented any concept of local names. that's not uncommon in small stack languages, though large ones often have locals...)

ireneista@irenes.space ("Irenes (many)") wrote:

yeah wow this iteration structure is fucked, we're glad we're looking into it. there was a critical comparison whose output wasn't even used. easy to do in point-free style, but we should have added more comments.

Boosted by glyph ("Glyph"):
GhostOnTheHalfShell@masto.ai wrote:

Yes, something like that:

“ While I’m not a lawyer, the numerous cybersecurity experts I’ve discussed this with are stunned by the complete lack of any legal action against OpenAI, which appears to have committed a crime that gets you anywhere from a year to a decade in the slammer. The fact that LLMs from both Anthropic and Meta have been involved in similar incidents is a sign that we need to arrest more people.”

https://www.wheresyoured.at/ai-is-already-in-dangerous-hands/

Boosted by db@social.lol ("David Bushell 🪿"):
vale@fedi.vale.rocks ("Vale") wrote:

@csstricks is once again stuck in an arbitrary suspension at the hands of DigitalOcean’s radio silence. It is devastating to see such a brilliant publication treated so very poorly.

The current situation: https://vale.rocks/micros/20260915-0135

#WebDev #Web #CSS-Tricks

Boosted by jsonstein@masto.deoan.org ("Jeff Sonstein"):
thejapantimes ("The Japan Times") wrote:

The Trump administration is rescinding climate controls on coal and gas-fired power plants, one of its most aggressive moves so far to unwind U.S. environmental regulations. https://www.japantimes.co.jp/environment/2026/09/15/energy/trump-epa-regulations-power-plants/?utm%5Fmedium=Social&utm%5Fsource=mastodon #environment #energy #emissions #environment #epa #climatechange #us #donaldtrump #republicans #coal #gas #fossilfuels

Boosted by glyph ("Glyph"):
sparklepanic@infosec.exchange ("raven (she/it) :neodog_verified: :gentoo:") wrote:

@mttaggart as a healthcare professional, non-programmer, linux user, i absolutely want to learn C and Python the hard way so i can write software and eventually contribute slop-free code to open source projects that are committed to human written projects

im getting started with "Learn C Programming and OOP with Dr. Chuck [feat. classic book by Kernighan and Ritchie]" on YouTube, and the books "Effective C" and "Automate the Boring Stuff with Python"

jsonstein@masto.deoan.org ("Jeff Sonstein") wrote:

RE: https://flipboard.social/@newsguyusa/117272391691353779

“Meink’s statement is the most direct public acknowledgment by a U.S. official to date regarding space control weapons, but he did not say when the weapons were launched or how many are in orbit.”

Boosted by jsonstein@masto.deoan.org ("Jeff Sonstein"):
TheViking@newsie.social ("Lillian") wrote:

Attachments:

• video: 652f0afb2633ad39.mp4

Boosted by jsonstein@masto.deoan.org ("Jeff Sonstein"):
ScienceAlert@flipboard.com wrote:

A 3-Year-Old's Metastatic Cancer Disappeared After Two Doses of Experimental Therapy
https://www.sciencealert.com/a-3-year-olds-metastatic-cancer-disappeared-after-two-doses-of-experimental-therapy?utm_source=flipboard&utm_medium=activitypub

Posted into Health @health-ScienceAlert

Boosted by glyph ("Glyph"):
arichtman@eigenmagic.net ("Ariel") wrote:

@nickzoic @glyph now hiring, must have: diplomatic immunity
Nice to have: software experience

Boosted by jsonstein@masto.deoan.org ("Jeff Sonstein"):
markwyner@mas.to ("Mark Wyner Won’t Comply :vm:") wrote:

Whoa. @libreoffice unleashes “no AI as a feature” and breaks download records.

“In one week, the installer was downloaded more than 1 million times — not counting updates via Linux distribution repositories.”

Great work, LibreOffice team! 👏🏻

https://manualdousuario.net/en/libreoffice-download-record-no-ai

#LibreOffice #FOSS #AI #Software #NoAIAsAFeature

Boosted by jsonstein@masto.deoan.org ("Jeff Sonstein"):
In_Otter_News ("InOtterNews") wrote:

Crossfit, croissant, it's all the same...

Boosted by jsonstein@masto.deoan.org ("Jeff Sonstein"):
DrOinOR ("OhSnap!Dragon") wrote:

📽️ 🎞️ 🍿
"The Muppets Take Manhattan and Make It Their Bitch"

#HyberbolicMovies
#HashTagGames

jsonstein@masto.deoan.org ("Jeff Sonstein") wrote:

my friend Michelle has a beautiful monarch butterfly piece in RL and online. check it out on an Insta reel:

https://www.instagram.com/reel/Dcy6jDhS-%5Fm/?stkn=MTc4MmM1YmI2Ng==

krig@goto.liten.app wrote:

We are now half a decade into a pandemic whose speculated long term effects include cognitive decline at the same time as the world has gently sailed past 432 ppm global average atmospheric CO₂. When I was born we were at 337 ppm. The adverse health effects of prolonged high carbon dioxide exposure are many, but include "reductions in higher-level cognitive abilities".

It's not surprising that everything is exceptionally stupid now.

Fucking sucks though

ireneista@irenes.space ("Irenes (many)") wrote:

if we can get the runtime of the full transform down to less than a minute, that would rock

ireneista@irenes.space ("Irenes (many)") wrote:

but that also made a nice spot to pause and work on the optimization we mentioned earlier, while it's fresh in our head

ireneista@irenes.space ("Irenes (many)") wrote:

okay. we got hex-transform to install a trap inside label-transform-one (the inner piece of the loop that label-transform drives), which in turn has the ability to apply additional replacement rules

we tested one of those rules, but in a way that's an nop, because we need to refactor something else before we proceed with that

glyph ("Glyph") wrote:

Re-reading this I am almost kind of shocked by just how many bad ideas are new again. Are structural forces making our systems unreliable due to an irresponsible frenzied rush to go too fast? Let's put a "human in the loop" to responsibly "sign off" on everything. That'll fix it.