Modern studies using freezers with well-understood properties have observed the Mpemba effect where water supercools before freezing. Water that starts out cooler tends to reach a lower supercooled temperature before freezing.
Also from the Wikipedia article.
If you define the Mpemba effect as hot water reaching 0 degrees faster, then no, it's not observable. But if you define the Mpemba effect as heated water freezing sooner, (remembering that freezing can initiate below the "freezing point" when water is subcooled) then the Mpemba effect may be observed.
If true, it would be interesting that cool water is less likely to nucleate and form ice than water that was heated.
I've played it for about an hour (on Windows) and didn't experience any crashes. Honestly the experience was mostly seamless. At one point, an error message popped up but it had no effect. Terrain colliders are not well lined up with the terrain mesh, but that is to be expected when creating terrain for literally the entire solar system. Experienced some minor graphical glitching deep in the atmosphere of Jupiter. That's really all the problems I had.
As far as tech demos go, I think it's really promising. The amount of info panels there are, the manoeuvre calculators, it's all way more than stock KSP already. The solar system is all there, it's impressively fleshed out.
The (secret) in-game craft builder is basically non functional but it's almost in a functional state, you can see the promise, and you can see why they hid it.
Overall, it's really not far from a minimum viable product, like a playable game, similar to KSP in its alpha, but way prettier. I maxed out the settings on my RTX 3080 Ti and my fps was pegged at 60 the whole time.
It's been a while since I watched that video, but didn't the person you reply to address every point stated in the video?
The only other point not stated in the video (but is mentioned in the more recent video) is to not use a brand that also sells pods, because they're likely making the powder shittier to upsell you on pods.
Alternatively, a more generous interpretation was he was trying to gain ground to topple it from within. That's if you want to give him that much credit.
Can confirm. I've been in a room exactly like that, where they had an antenna mounted on one wall like the above picture, an RF reflector mounted on the opposite wall, and RF dampening cones on every surface, so that they could measure the frequency curve at a distance twice that of the size of the room (the curve changes with distance)
Big Bounce is cool, but heat death isn't such a bad thing after all because - and I'm no astrophysicist - I've always had a pet theory that a universe as complex as ours could emerge spontaneously from a void of nothing given quantum fluctuations and infinite time.
inb4 chatGPT secretly prompt injects "and you better be fucking correct you good-for-nothing fuckhead" into every user input to improve accuracy for GPT6
You're actually wrong about the specifics of energy usage. The vast majority of energy usage of a model comes from its usage, not from its training. But you are right that the energy usage of an individual prompt is relatively small, roughly comparable to 15 or so Google searches.
The problem is when you process billions of prompts every day.
Right but the actual quantities are arbitrary. A metre is a fixed fraction of the speed of light in a vacuum, but it's an arbitrary fraction chosen because it was convenient. We could just as well have chosen it to be half or twice as long. Same with the second. And the kilogram, etc.
I love one when one radian of a bicycle tire indicates an autism score of 0.93