yencabulator

It's at most a Terms of Service violation escalated into political theater because checks notes China bad.

Get elected president?

icache pressure was the bane of my optimization work around 2002. Mostly because it fluctuated unpredictably even from fairly small fairly faraway changes. The system got large enough, with a long enough hot code path that was very hot, that it was almost pointless to benchmark a subsystem in isolation. Simplistic attempts to make any subsystem faster typically made the whole system slower.

In general, whatever subsystem took the cache misses showed up as the slow part, and I owned the very first component in the pipeline... It was fun to compile time disable my component to demonstrate that nothing got faster, though.

Just like xAI was regulated to not run more than 15 gas turbines? US lets companies get away with breaking things and punishes wrongdoing at most with a fine that's less than the money that was made by breaking the rules, so US regulations sort of don't matter.

https://www.theguardian.com/technology/2026/jan/15/elon-musk...

https://en.wikipedia.org/wiki/Colossus_(data_center)

It wouldn't be the linker that has to be unsafe, it'd be the "and now execute!" jump. And that could be abstracted as memfd+execveat, which are fairly normal operations.

OP's argument is roughly "doings things with computers has to be unsafe to be useful", which is.. uninteresting.

They said "Bosch washing machine and dryer". It's a combo unit that dries after washing.

https://lucumr.pocoo.org/2026/7/4/better-models-worse-tools/ says already too much to use foreign tools.

Go doesn't help with memory safety in concurrent settings, and a kernel is very much that.

What else would you want to see in this project?

Absence of spamming mentions of it everywhere.

How about we make browsers not report mouse movement to the page? This is getting way too creepy.

OpenCode and Codex are also crap vibecode:

https://github.com/anomalyco/opencode/issues/10939

https://news.ycombinator.com/item?id=48626930

That part makes sense, but the implementation is weird.

It was decided by Claude based on an ambiguous high-level goal-oriented prompt, don't expect it to make sense.

For anyone else looking for the source of that quote, it seems they edited the text slightly:

https://www.crowdsupply.com/open-tools/open-printer

Open Source

Open Printer will use the Creative Commons BY-NC-SA 4.0 license for all of its files, including electronics and mechanical design files, firmware code, and the bill of materials.

Calling CC-BY-NC "Open Source" is a blatant lie.

My high school physics teacher made a point of doing math on the size of the moon with pi=3. Close enough.

The same plain-language interface that answers questions is the interface that changes the system.

Surely nothing will go wrong with this.

Outside of some specific areas with very humid climates, needing a dehumidifier is even more rare than needing an air conditioner.

Especially so in the winter, which is dryer than summer in many places.

Wasmer, huh?

https://news.ycombinator.com/item?id=24900186

https://news.ycombinator.com/item?id=30758651

187,084 homeless in California in 2024.

California has a significantly better climate to be homeless in, so you'd want a bus ticket from Minnesota or whatever to there too if you were homeless.

- The car that parked nearby last night coincided with a "Derek's Galaxy Buds" device with MAC address aa:bb:cc:dd:ee:ff. The buds also drove by briefly the previous night.

Are there still lots of devices that don't do randomized MACs? Apple and Android are not that easily trackable anymore.

an ambitious DHT project

https://nlnet.nl/project/FogDHT/

links to a completely empty repository and the project name has zero hits outside of NLNet announcements -- what's going on here?

https://codeberg.org/FogDHT/FogDHT

but blades did fell out of fashion,

I prefer to think of that was "datacenter weren't ready for denser workloads"; blades were ahead of their time, and their adoption wasn't great because of that. Since then, new datacenters have been built with more ambitious power/cooling/weight budgets.

I'll go further. Linux heavily uses a form of garbage collection that cannot be implemented in typical userspace (without awkward & slower additions to the consistency algorithm).

https://en.wikipedia.org/wiki/Read-copy-update

I understand the general area of "small imaging and manipulation" is something IBM has been working on for a very long time. This is STM from 1989:

https://sciencephotogallery.com/featured/ibm-atom-manipulati...

I sometimes use LLMs as search engine replacements for finding libraries for a specific niches. Gemini is the only one that seems to routinely hallucinate Rust crates from thin air, where otherwise Google is the one most willing to let their LLM peek at a search index to have up-to-date ground facts. It's puzzling. Definitely feels like it's not "organizing the world's information" there.

The stupidest thing is that according to that logic, ncurses is a game engine too.

The problem is that it often causes tumors.

This short video talks about telomere-snipping transcription factors that make seemingly-dead cells revert to normal or even to stem cells, but the factors only act when under artificially-maintained conditions:

https://www.youtube.com/shorts/k49bGcbOKwg

I’ll concede that “identity theft” could conceivably have a reasonable meaning in the context of e.g. Estonian digital identities where you could in a sense steal someone’s private key.

Stealing someone's private key still wouldn't be "identity theft", they'd just need to bootstrap a new key the same way they got the government to trust the previous one (and revoke the old one).

It's just victim-blaming and PR attempting to make companies not liable for their bad decisions.

GLM-5.2 cost a fraction as much. Opus finished in half the time and shipped a cleaner game.

The only thing you measured was a single tunable of how much work should be done on a vague prompt. Now make the prompt be something that causes GLM-5.2 to cost 4x of previous budget, to get something comparable.

(And the wallclock time measures the inference provider, not the model.)

Running builds on FUSE all the time is likely more wasteful than cloning a typical repo once per session.

Totally breaks the permission model in Claude Code.