I am a big fan of WUFFS but some recent developments by people assisted by AIs writing format parsers in other languages has made me question whether it will be maintained or forgotten.
Everything at Google rolls out incrementally to different populations of users so no matter what feature they are announcing there is a good chance you can't see it yet. When they announced Gemini 4 Argon it was 2 days before I could use it.
Live state of IMEI corresponding to every cell would be technically easy. Most of the stuff that goofballs imagine the NSA doing are technically impractical or even thermodynamically impossible.
It isn't "fuckery" as such. There's just an industry of selling dossiers on individuals, groups, or populations that is way, way more intrusive than what people imagine about online advertising, and they have been doing it for decades. It's data on the open market.
The article could have been interesting but by fundamentally misrepresenting the structure of the American federal government they made it impossible to take seriously.
Fission is unmatched in its ability to abandon projects after spending tens of billions of dollars. I don't see how any rational person can look at the history of nuclear and say "damn, this thing can drop electrons on short notice." That's just a ridiculous take.
Distributed solar is cute and it sometimes benefits some people but it is pointless for national policy to pursue it. As you pointed out, the cost structure of small distributed solar is dominated by office overhead, sales, management, insurance, litigation, etc etc.
Perhaps they are planning to launch application features that require or benefit from the A20. Anyway it looks like the base model of the new iPad is using binned CPUs that were not quite good enough for the iPhone 18: 5 cores instead of 6, 1 neural engine instead of 2.
Polymarket isn't binary so you are understating how wrong they were. Polymaket gives outcomes as percentage odds and less than 12 hours ago they had Bolsonaro by 5-to-1.
"Data centers run flat out around the clock", a statement that is not true, has never been true, has been advanced by know-nothing writers in non-technical outlets, and I guess is now encoded in Claude or whatever slop cannon created this site.
There was an official 1394 card for the O2 that only supported DV (not the usual block device protocols).
SGI was super weird with 1394... the Origin 3400 I-brick had a 1394 port that it used for its internal CD-ROM via a 1394-to-IDE adapter, which is freak show stuff.
It definitely took me back when the presenter came within half a pixel of logging out instead of opening the application. That's why they made the O2 small enough to throw across the room.
That doesn't explain anything in this article. All the time series go back to 1990 when interest rates were significantly higher any way you look at them.
This seems to raise the question: is libdeflate fast? My overall impression of libdeflate is that it produces a similar output size to zstd but takes 40 times longer to do it. I realize these are different algorithms of course, but they aren't that different, both being in the Huffman line of the LZ family. So why is libdeflate considered a performance benchmark?
This is true for things other than LLMs. Outside contributors have been credited with hundreds of security fixes to macOS and iOS that were found with fuzzers and sanitizers. Just off the top of my head CVE-2019-8605. Clearly, Apple has access to AddressSanitizer and libfuzzer, but outsiders still find these things. Not only Apple: Google can find flaws in iOS but outsiders also find flaws in Chrome, so the phenomenon is widespread.
The site is roughly here, for anyone who wants to judge for themselves whether clear-cutting forests is a local custom or whether every Finnish tree is sacred.
That isn't how history has unfolded. What has actually happened is that operating systems ship with extremely bad allocators unsuitable for any application and due to some kind of emergent collective problem they are never improved.
The operating system could effect this by not providing any page tables to the CPU, so every memory reference faults and the OS has to handle them all.