Awesome work
HN user
desperate
I think this is a good video to share to that end https://youtu.be/242pqLSFzh4
Wow, I've been wanting an article on this topic for a while and this one really delivered that and more. Thank you.
It's just plain scary that Apple talks about the government as temporary and thinks it can delay complying with the law until something gives.
This is horrible news. The USCSB does incredible work.
Same. Pretty much upvoted because I didn't understand but wanted to, and thought other people might be able to provide more insight.
Yeah but money can't pay for skill that doesn't exist anymore.
^^^^ this is so frustrating to me.
Your speculation is pretty accurate afaik. Mark commented that releasing open source models is what is needed to retain the best so it is what they will do (paraphrasing) on his interview with Lex Friedman.
Afaik you're describing something akin to a recurrent neural network, and the problem with that is that it doesn't parallelize well to modern hardware. And vanishing gradients.
Did you watch the video? The author has nothing against trucks themselves, it's SUVs (light trucks) in suburbia and cities that he argues is the problem. Specifically, the many cases where people never even use the capabilities of the larger car (off roading, towing, stuffing it full of shit) and instead the choice of a larger car becomes an overall negative for society (more dangerous for the other guy in any collision, harder to see around, harder to see kids or pets as the driver, headlights more blinding to oncoming traffic die to their greater elevation). Anyhow, you sound very angry and closed minded, maybe give other opinions a chance, you don't have to agree but you also don't have to be an ass about it
How would the energy cost of the two methods compare?
To me this is proof that infinity is something only present in our math and not in the universe.
Infinity is a nice approximation but it feels like wishful thinking that our universe or anything in it is infinite.
Happy to hear disagreements tho.
“I don’t think there’s any argument that this is a worse console, at least for thermals and for cooling,” says Evans.
Worst quote they coulda pulled. Literally sends the opposite message than the one its meant to convey.
Unless im even worse at english than i thought.
Can you explain? I'm (semi) familiar with the heap and definitely hated on first fit and next fit in favor of beat fit. I get that its gonna take significantly less time to find a block to allocate but why does listing files benefit so much from it and not run into a memory usage problem. Unless your splitting the block on allocation but then i would think you still have a peoblem cause you'd have a high degree of external fragmentation.
I typed a lot so i can be better corrected, not because I understand what is going on.
Tesla valve?
Don't you think these fourth buyers apply more to supporting a baseline price on Monero and not Bitcoin?
Not sure how relevant this is since it's a third party app, but using Harmonic on mobile the original link would allow me to view the PDF without downloading it (albeit after hitting an empty HN page) while the direct link requires me to download the PDF before viewing.
Tbh I'm just happy to have hacker news, continue as you are.
For those interested Ted-Ed did a neat video on the topic of grocery bags.
Exactly my thinking. My first thought was 'Without someone to enforce them, laws are merely suggestions'.
My bad, I'll call it off right now. Sorry for the inconvenience.
Another great video on it, this one from TED-Ed
Not yet anyway...
How come the search is somewhat stochastic? Deleting the "e" from "cake" and retyping it causes ~ 1 in 4 searches to display pure "cake" entries while the rest show things like "Baked praline ice cream cake".
I don't think you understood what your parent meant. In their case the GPU was detrimental cause it was reducing the ram available to the rest of the system to 5gb.
"The vast majority of the world’s chips are made in China, while the U.S. is the second biggest producer. "
I guess CNBC doesn't recognize Taiwan?
Thanks for this, incredibly informative.
Slightly off topic but I just realized that not only does optimized code = less processor cycles but less cycles = less power used.
This brings two rhetorical questions, How much more effective compute would we have if all code running today were perfectly optimized, and how much less energy would be needed under the same circumstances?
The computer gains could only be used in networks where unused compute can be scheduled / sold to other tasks. The energy gains could only be reaped if processors had proper low power modes they could fall into or some other method of conservation.