tensegrist

given they turned off the refrigerators (instead of something evil like turning them down for just long enough to allow bacterial growth and then turning them back up) customers would probably have noticed their ice cream being liquid, the frozen meat not being frozen, etc

i know some of us are thinking "if you consistently fail to ensure your ability to make high-severity security patches reach critical internet-exposed systems on time you lose your ability to operate the business and probably get jail time"

but what about the deranged (and very prediction market–brained) version of this idea: the government develops ransomware, in a manner functionally similar to towing cars or confiscating illegal items, and uses it against companies in proportion to their revenue and obviously puts the money towards improving the ransomware

"the r in rsi stands for…"

i'm not sure it's something i want to see, but i don't know if it's impossible for it to actually come to pass or merely unlikely

(of course, cross-government versions of this are much more likely to come about but not as ransomware)

> ParticleDB’s home page advertised that ParticleDB was “formally proven” with a TLA+ proof. However, during our collaboration no TLA+ specification was available. It is unclear what properties might be proven, or whether they described ParticleDB as built.

perhaps i was wrong

https://news.ycombinator.com/item?id=50272530

in my over a decade of using nix i have never really managed to get into using the repl because every time i have it's been … weird in some way or another

php2shell was overblown perhaps but this is being blown quite correctly

perhaps you could add optional magnetic rails for extra support

or even 3d-printed "grids" of arbitrary shapes, with magnets embedded at the junctions, to interface with the keys' magnets while being separate from them

i suspect there is an article under all of the ads but i don't have the patience to look for it

for what, chess.com?

doctolib mcp

i was at the manchester art gallery last weekend and they had a bunch of lovely t&h (and other) books on colour in the gift shop

that's astrish not claudish

beside-device inference

that's how you know it's good. bad apps have rather smaller daily user-bashing numbers

possibly enabled by lsp (Adw.foo -> "foo found in GI.Awd, import it with the name you used?")

> Caecilians

with that capitalization i had 200ms of my brain going "what part of the bible is that"

this is related to, but not the same as, the question i had in the original thread: https://news.ycombinator.com/item?id=49719097

noting that input:output:cached is 10:50:1 for astra and luna but 10:50:0.5 for sol. this doesn't mean a lot without "tokens per task" information but it's still interesting for there to be a "dip" like this instead of a monotonic change in one direction or the other

asimov, "the feeling of power"

[2] https://mypowergrows.substack.com/p/posters-in-profile-davie

where's the flash model? it's out already isn't it

i thought the computation was within the font itself

like

https://github.com/RoelN/SansBullshitSans https://typographica.org/typeface-reviews/chartwell/ https://github.com/trishume/numderline

which can be taken rather far

https://www.coderelay.io/fontemon.html

what is the…epistemic status, for lack of a better way to put it, of the probabilities? what do they mean? what (probabilistic) guarantees do we have about, say, the responses to

- is the capital of france paris?

- it is august. is it raining in paris?

(forgive the examples; they're probably not semantically the sort of thing jev is trained to work on. but i figure the point translates to various kinds of questions that come up in "inner loop of agentic pid controller" contexts)

a normal text-generating model if asked to produce a number will also do that just fine. i assume in jev's case it was actually rled to essentially learn to express priors over things using its implicit world model, which definitely ought to help, but can we say more?

"i'm sorry, i left the task 半done"

    In keeping with our commitment to user responsibility, following the official launch of V4.1 Flash and prior to the release of V4.1 Pro, all requests to the Pro model will be routed to V4.1 Flash and billed at Flash's price.

just in terms of user perception when selling this sort of service, this is what they call a "good look"

say with "finite" before it. it's a thing

how do you think "referencing" happened when "referring" was already around

or conference, difference, arbitrage…

this reminds me that i'm long overdue for a weekend spent studying firecracker, the userfaultfd trick [0], and whatever zeroboot does

[0]: https://codesandbox.io/blog/how-we-clone-a-running-vm-in-2-s...

    claude -p "${curl $url}\nreply only with a substantive comment in 'statement. question?' form" | dd conv=lcase

as a service

i wonder how hard it would be to write a decent classifier for individual comments like this (in practice you can just look at the account history which would make it an easy problem)