Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
platz
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
94 ms
·
481.
▲
by
platz
5y ago
Maybe we shouldn't build bsl4 labs in big cities
482.
▲
by
platz
5y ago
Maybe their teenagers are just out having fun surfing on a cosmic gamma ray blast and we happen to be in it's way.
483.
▲
by
platz
5y ago
What if they are like a space fungus.
484.
▲
by
platz
5y ago
We don't know that the universe is infinite. Also, even if space was infinite, that doesn't mean that low-entropy energy/matter (to extract energy/work from) is also infinite.
485.
▲
by
platz
6y ago
I believe he said the wealther you got, the bigger burden you owe to athens, because without Athens you wouldn't have gotten rich, and we want there to be rich people in the future in athens as well.
486.
▲
by
platz
6y ago
david cay johnston on the moral basis of progressive taxation (3 min) https://youtu.be/8Qsl6QUlWVc
487.
▲
by
platz
6y ago
Imagine a stock market without shorting
488.
▲
by
platz
6y ago
> It would be pretty easy even with today's tech to stimulate neurons similarly to braille touch which we know works Citation needed
489.
▲
by
platz
6y ago
And theres a big difference between between a completely controlled environment like a chip and an uncontrolled environment relative to engineering like a brain. I think your missing the point about what I'm highlighting needs to happ
490.
▲
by
platz
6y ago
Do you really randomly poking some electrodes into your brain tissue is going to give you enough fidelity to query google. That's all it is, you know. Some wires, randomly placed in tissue. The cells around those wires don't kn
491.
▲
by
platz
6y ago
https://16colo.rs/
492.
▲
by
platz
6y ago
The actual interactions seem rather limited. Rotations, translations, associations. not sure how much fidelity that actually gives you. How many problems can you actually solve with that?
493.
▲
by
platz
6y ago
Which is what I demonstrated the parent poster did.
494.
▲
by
platz
6y ago
it's not good to cherry-pick paragraphs from the whole artile. > But as the Brookhaven team accrued 10 times more data, their measurement of the muon’s g-factor stayed the same while the error bars around the measurement shrank. The
495.
▲
by
platz
6y ago
Does anyone know how close Falcon 4.0 (or Falcon BMS) got in terms of the flight model? Do the DCS games have a good flight model?
496.
▲
by
platz
6y ago
Scarcity of precisely what is the question. In this case, a small amount of metadata. That metadata is inherently worthless, because the metadata is just information that can be copied, and a new supposedly "scarce" token can be
497.
▲
by
platz
6y ago
> what's considered "an original", and I think that's what's happening with NFTs today An original what? Remove everything off-chain from the conversation. All you have left is a small amount text (which itself
498.
▲
by
platz
6y ago
I realize now that calling it an edition was an error on my part, because in irl an edition is a physical copy of the work itself which can be made scarce, whereas a token is just a pointer with a url to some infinitely-duplicable bits, so
499.
▲
by
platz
6y ago
An NFT does not solve or prevent the artist from selling additional editions of their work. They can always create more tokens.
500.
▲
by
platz
6y ago
please show me all the tokens on openseas that are licenses that grant commercial distribution rights.
501.
▲
by
platz
6y ago
> tokens of ownership They are not even "tokens of ownership"; they're just tokens that one owns , nothing more, but the semantic distinction here is significant, because a token cannot own anything outside of itself.
502.
▲
by
platz
6y ago
> Even though it's not physical, it's exactly like owning a baseball card or a work of art. you're confused. You don't own anything with an NFT other than a pointer. so it's not like owning a work of art at all.
503.
▲
by
platz
6y ago
even with art, an NFT doesn't confer copyright rights, so you don't even own the art. > But as it turns out, you can make digital things scarce, through cryptography. The art isn't stored on-chain, so it's not scarce
504.
▲
by
platz
6y ago
https://opensea.io/assets/0x495f947276749ce646f68ac8c2484200... > Link to a print-ready high-resolution image file will be unlocked on purchase (2048 x 2048 pixels). LOL
505.
▲
by
platz
6y ago
even with art, an NFT doesn't confer copyright rights, so you don't even own the art. > a way for the artist to generate artificial scarcity The art isn't stored on-chain, so it's not scarce either. All the NFT is is
506.
▲
by
platz
6y ago
What is you pivot to?
507.
▲
by
platz
6y ago
I definitely would
508.
▲
by
platz
6y ago
Why not just use a real recording
509.
▲
by
platz
6y ago
well, the orion image you linked doesn't look like it's been over-sharpened, in contrast to the OP image. I'm not saying there's a way to tell just by eyeballing it, but that's what it smells like.
510.
▲
by
platz
6y ago
It can be true that it's both based on real data and that there is an aggressive amount of post-sharpening applied.
More ›