As far as I can tell, power isn't actually a major part of the expense, it's dwarfed by the capex. Just the amortization on the GPU will be an order of magnitude higher than the cost of the power to run the GPU at 100%. (Assuming a 5 year depreciation period.)
I mean, the claim is certainly nonsensical in the sense that this isn't something Wärtsilä just "realized". They have been in the power plant business for decades. In the oldest financials they have online (the annual report for year 2000) their power plant sales are larger than their marine engine sales.
Really makes me wonder about anything else I've read on Semianalysis. Like, it is such an insane thing to claim and so easy to check. And they just wrote it anyway, like some kind of pathological fabulists.
But what's the part that seems like a "big reach"? Are you saying they didn't sign those contracts? That their customers are making a mistake?
So that is what the classic "Cheney on MTA" Lisp implementation paper was named after. It felt like a reference, but I never had an idea about what it was referring to. Thanks!
It was for a dynamically growing ring buffer that also did short-object optimization. The natural implementation was to have the capacity and the offsets stored in fixed locations and with a fixed width, and have the variable part be a union of pointer or inline byte buffer.
Depending on the element width, you'd have space for different amounts of data in the inline buffer. Sometimes 1, sometimes a few more. Specializing for a one-element inline buffer would be quite complex with limited gains.
In retrospect trying to use that as a running gag for the blog post did not work well without actually giving the full context, but the full context would have been a distraction.
There are plenty. But it's not the comparison you want to be making. There is too much variability between the number of tokens used for a single response, especially once reasoning models became a thing. And it gets even worse when you put the models into a variable length output loop.
You really need to look at the cost per task. artificialanalysis.ai has a good composite score, measures the cost of running all the benchmarks, and has 2d a intelligence vs. cost graph.
The _Science_ paper linked is paywalled, is anyone aware of a preprint?
I find it a bit curious that they've chosen to use SMS verifications as a proxy for the difficulty of creating an account, when there are similar marketplaces for selling the actual end product of bulk-created accounts. Was there some issue with that kind of data? SMS verification is just one part of the anti-bulk account puzzle, for both the attacker and defender.
The infinite grid is a cool idea, but the game as a whole did not feel very engaging. There was no way I was finishing even one full screen of this, which is the bare minimum threshold for getting any value from the infinite grid. I'm not a word search guy, but it feels like the idea might work better with a different puzzle type.
The drag to pan, hold-then-drag to mark felt really clumsy with a mouse. The hold delay is just too long. Maybe consider drag-with-left-button to mark, click/drag-with-right button to pan.
The word list seems odd, half the words I tried were rejected. You don't need many failures to accept a reasonable word to lose faith in the game. Is the idea that only the words at the bottom are in the dictionary? Doesn't feel like it can be, given it only shows about 20 words, is not scrollable in any way I can see, and many words not shown in that list were actually accepted.
reply