hajile

69 ms·

901.
by hajile 4y ago
I believe that was shown in 2011 (though as this follow-up states, there was a slight mistake that needed to be corrected). https://accodeing.com/blog/2015/css3-proven-to-be-turing-com...
902.
by hajile 4y ago
We wouldn't be in this situation if Brendan Eich had been allowed to ship a scheme like he was hired to do. Had that happened, we'd be free of the perpetual JS feature churn (and all the weirdness). Performance would have gotten b
903.
by hajile 4y ago
Few alternatives (when looking for equipment beyond a basic tractor) and most are just as bad.
904.
by hajile 4y ago
But if you chain both of those, the `.compact()` will use Lodash iterators while the builtin will repeatedly iterate over the array.
905.
by hajile 4y ago
Drought resistant not a silver bullet. It can help, but only up to a point (and that point seems lower than the sales guys and sponsored studies under optimal conditions might indicate). Edited because part of this wasn’t factually correct.
906.
by hajile 4y ago
What does GMO do to solve “we didn’t get any rain”? They’ll get sold into patent slavery they can’t afford to solve the wrong problem.
907.
by hajile 4y ago
If you want a truck for looks, EV is nice enough. If you want a truck to haul stuff, EV is terrible . Put a trailer on your EV and they have basically no range at all. If you need a truck to haul stuff reasonable distances, EV isn't f
908.
by hajile 4y ago
If I recall correctly, Rivian is hugely invested into by Cox (specifically Cox Automotive started with a 350M investment), Amazon (750M in 2019 alone), and even Ford (around 500M investment in 2019).
909.
by hajile 4y ago
It looks at existing data to statistically identify the most probable answers. From this, you can infer that angry, intolerable developers who believe their language of choice is the best are most likely to program C++.
910.
by hajile 4y ago
It's not quite as bad as you might think (though it is bad). TSMC N5 is 0.021um per SRAM cell in high-density configuration (still going to be 5-10x faster switching than DRAM). That's about 47.62 bits/um. With 1,000,000 um2
911.
by hajile 4y ago
Cache size is constrained by different things at different levels. Cache uses 6 transistors per bit while normal RAM uses 1 transistor and 1 capacitor. You can buy a LOT more RAM per dollar than you could buy cache. This has been further ex
912.
by hajile 4y ago
That's only partially true. SRAM doesn't need to be refreshed constantly like DRAM, so SRAM will use LESS power when not switching. We think it's different because something like L1 cache is CONSTANTLY switching. It should be
913.
by hajile 4y ago
No matter how you slice it, Java enterprise would only cost $180/yr which is almost 10x better than your best estimate Let's say you have 25 devs on a 5-year project. 5 of them are mobile focused, 5 are windows focused, 5 are mac&
914.
by hajile 4y ago
$15/mo or $180/yr. Lispworks for that same platform support would cost about $20,000 per dev the first year then a few thousand every year after. You could buy 100 years of enterprise Java support for Jay one year of Lispworks.
915.
by hajile 4y ago
They know their locked-in customers, but they also know they don’t get many additional customers at those prices either.
916.
by hajile 4y ago
Professional starts at $1500 for the 32-bit edition. That price doubles for 64-bit. Need to connect to a database and use their ORM? That’s another $1500. Want to do a cross platform mobile app? Add another $2000. Need windows and Mac? Th
917.
by hajile 4y ago
> Electrons are relatively speaking more like the nice little billiards ball thwacking away at the SI that we like to imagine. They are as long as you keep a close eye on them…
918.
by hajile 4y ago
Because it's a bad deal. N3 isn't a good node and offers ZERO SRAM scaling. It's a stopgap for N2 GAA. At the same time, chip shipments have dropped like a rock as everyone braces for recession. Nvidia allegedly even tried to
919.
by hajile 4y ago
Process node is HUGE here. M2 is pretty close to just M1 with higher clocks and higher power consumption. This is largely because they didn't get to add transistors. The N3 shrink doesn't help SRAM one bit, but it does give them m
920.
by hajile 4y ago
N3 scaling? There's literally zero scaling. N3 is a stop-gap as TSMC is behind on their N2 GAA process. It's not so different from Intel's 14nm++++++ where it was dramatically better than before, but still not enough.
921.
by hajile 4y ago
I think TSMC N3 will be underwhelming. It offers a shrink of logic transistors, but ZERO shrink to SRAM meaning one of Apple's biggest advantages (massive caches) isn't going to have much room to improve.
922.
by hajile 4y ago
MS had to add content to the page then render that new content. That would count as a derivative work which would be covered by copyright. Normally, when a user does this themselves, it isn't for commercial purposes which puts it in a
923.
by hajile 4y ago
Interestingly, the Google download page is copyright by Google. Microsoft changing that page for their profit is 100% NOT fair use. Google should simply sue them for all those counts of willful copyright infringement at $180k each.
924.
by hajile 4y ago
It remains to be seen how well RDNA3 actually works. Their biggest changes around making culling and intermediate ray calculations faster require software changes. Honestly though, I feel like most people would rather get higher framerates.
925.
by hajile 4y ago
Interested in a GPU under $800, then RDNA2 is the latest and greatest. Interested in the newest available mobile chips? RDNA2 is the latest and greatest. Steamdeck? Game console? It's the most recent there too. That said, they've
926.
by hajile 4y ago
This is huge for CPU workloads that need intermittent parallel processing. A great example would be JSON parsing in WASM apps where SIMD algorithms result in massive speedups. Another example would be a WASM software decoder could allow AV1
927.
by hajile 4y ago
> Added support for WASM SIMD. This is huge for the progress of WASM > Added support for growable SharedArrayBuffer. Not a headliner feature, but a pretty important one for JS/WASM interactions. > Added support for AVIF on mac
928.
by hajile 4y ago
Everyone thinks they have big data and 99.99% of the time, their entire DB could be served from a single 100GB SQLite file on a single SSD on a random dev's laptop with better performance than the expensive AWS deployment they've
929.
by hajile 4y ago
The Haskell type system is WAY more simple than the type soup TS allows you to make. I'd gladly take Hindley-Milner typing and typeclasses (or better yet, module typeclasses) over the current TS situation.
930.
by hajile 4y ago
TS shouldn't have embraced all the JS quirks. I want a type system that embraces the correct way to to do things instead of allowing anything you want no matter how bad it actually is. I want my type system to inform me that my code is

More ›