nilstycho

https://developer.nvidia.com/blog/inside-nvidia-groq-3-lpx-t...

https://xkcd.com/1179/

O algorithm, algorithm! all men call thee fickle.

Better to read the original story from Politico.

https://www.politico.com/news/2026/01/27/cisa-madhu-gottumuk...

I believe it's "TextSynth".

https://bellard.org/ts_server/

It was Jack Lance, who wrote Enigmash. Tragically, he died in 2023 at the age of 25. Jack Lance superlatively creative. I cannot find the words to express how much the world lost. I do not know of a finer puzzle designer.

https://jacklance.github.io/games.html

Oops, you're right. I read that factoid [1] while searching for information about the 2025-077C, but it was about a different 2021 launch.

[1] https://spaceflightnow.com/2021/06/15/three-nro-satellites-l...

I think that Centaur was sent into solar orbit.

I think there are only two US rocket stages in LEO that were launched in the past 21 months:

1. 2024-125H — Firefly Alpha FLTA005 Stage 2, launching demo cubesats for NASA

2. 2025-077C — Orbital Minotaur IV (8) Stage 4 (Orion 38), launching a few spy satellites

Fun fact: the first stage of the latter rocket was manufactured in 1966.

“The trick, William Potter, is not minding that it hurts.”

Do you have any tips for how to work on this with kids?

Second, Starship is much more expensive for each launch attempt than Falcon 9 ever was.

The launch cost of a Starship today is high, especially if you include development costs, but Musk's goal is a marginal launch cost of ~$1M. A Falcon 9's launch price is ~$70M; Musk claims a "best case" marginal Falcon 9 launch costs ~$15M.

The distributions in TFA are accurate. Compare with the shape of the 2015 distribution here: https://ourworldindata.org/the-history-of-global-economic-in...

The x-axis isn't "income", it's "log income".

My partner, who was raised conservative Mennonite, tells me this is exactly how pastors are chosen today. About three men are nominated, then they draw lots.

Here's a link to the diff between first publication and today: https://www.diffchecker.com/CJz1Bn51/

Was it this? On August 27, 2020, Natalie Escobar for Code Switch interviewed Vicky Osterweil about her book In Defense Of Looting. The segment was titled "One Author's Argument 'In Defense Of Looting'", and was subsequently retitled "One Author's Controversial View: 'In Defense Of Looting'".

https://www.npr.org/sections/codeswitch/2020/08/27/906642178...

https://www.hachettebookgroup.com/titles/vicky-osterweil/in-...

The Pritzker Prize leans modernist. If you prefer classical architecture, check out the Driehaus Prize [1]. The 2025 laureate is Liam O'Connor.

[1] https://driehausprize.nd.edu

I got lost at (?R). IIUC, this is a feature of the "regex" module that "tries to match the entire regex recursively". Fun.

I like the idea of union names [1].

If you're AC and your partner is BD, then on marriage you choose a new name X. You become AXC and your partner BXD. Your child is EX. Then each child has equal connection to both parents. The cost is that you lose the deep history of names, but that history only existed for a single lineage anyway, so it's not as important as it seemed.

(1) https://nothingismere.com/2013/11/12/solve-surnames-with-uni...

For about $6,000 a pair of Nvidia Project Digits offer about a combined 2 petaflops processing and 256 GB VRAM.

2 PFLOPS at FP4.

256 GB RAM, not VRAM. I think they haven't specified the memory bandwidth.

Are you comparing tinybox red with 738 FP16 TFLOPS at $15K to Project Digits with 1 FP4 PFLOP at $3K? Or did they announce the Project Digits FP16 performance somewhere?

Far less than 10%. During a heavy downpour, by volume, about one part in a million is liquid. In a cubic meter of heavily rain, there are only a few tens of raindrops.

Yes, that would be my preferred answer!

What's the name of the women's liberal arts college in Cambridge, Massachusetts?

Wait, what is the correct answer? “Radcliffe College”?

With a large sample size, it's astronomically unlikely for there to be a confounding trait that is important to the outcome and is also widespread in only one of the experimental arms. If you were to write a simulation showing the issue I might be able explain more specifically why I think the simulation doesn't reflect reality.

In a sample of 10,000, the number who have trait T9 is about 20. The chance of all 20 being in a randomly chosen half of the sample is about 1 in 500,000. It's quite unlikely that trait T9 is randomized this way, and it's also quite unlikely that the effects of just 20 T9 individuals on the outcome of interest are so strong that a typical analysis would find a significant difference between the treatment and control arms. The chance of both these things being simultaneously true is negligible. Do I misunderstand you?

Based on the flow diagram, that doesn't seem to be the same thing. Webwright seems to be a shell as a tool for me, enhanced with AI features. I'm suggesting the shell as a tool for AI.

Webwright is a front-end shell that presents to me; I'm suggesting a back-end shell that presents to Claude.

It doesn't appear that Webwright enables tool-use. In other words, there's no task-oriented feedback loop between AI-provided shell commands and the results of those shell commands. Please correct me if that's not right.

IIUC, Claude's "Computer Use" is roughly a remote desktop, which is a superset of a remote shell. I don't think I'm proposing anything with a greater risk than already exists.

What are the risks? Isn't this a strict subset of the risks of full desktop access? Claude can just open a GUI terminal with Computer Use. (I think.)