many worlds bros...

@georgia in another world: who keeps flipping these bits??

@snacks https://blog.google/technology/research/google-willow-quantum-chip/

@georgia I’ll believe it once it actually does something useful, but it’s interesting to see that they are making progress.

@lain @georgia if you could do even the most basic math using that type of speed couldn't you just brute force RSA

@sun @lain were gonna need longer passwords...

@georgia RCS means they're just sampling results from quantum circuits, this is a completely expected "speedup" for a 100-qbit system given that's a 2^100 element vector to simulate. that doesn't mean it's doing anything actually meaningful. also no idea where they're getting MWI, quantum computers just do that no matter what

@sun @lain @georgia yes

@arcana @lain @georgia my pt is just if it did anything useful they'd be saying what useful thing it did

You can just tell investors you don't have enough secondary/cache memory or it decays too fast -- yet if we get more funds we can fix that -- when you can't do something useful.

"Also we need to start running a fission reactor as part of the datacenter anyway, so could use the money."

@arcana @georgia @lain I'm not quite a quantum computing skeptic yet but based on vibes

@sun @arcana @georgia
> Having said that, the biggest caveat to the “1025 years” result is one to which I fear Google drew insufficient attention. Namely, for the exact same reason why (as far as anyone knows) this quantum computation would take ~1025 years for a classical computer to simulate, it would also take ~1025 years for a classical computer to directly verify the quantum computer’s results!!

https://scottaaronson.blog/?p=8525

@chjara Copenhagen bros...

@lain @arcana @georgia the cool thing about factoring keys is you can tell in a couple microseconds if it worked

@sun @arcana @georgia @lain I think like ai it's not cracked up what people think it is

@sun @arcana @georgia yeah, I wonder if some of these quantum computation benchmarks are like the instantaneous effects that happen during entanglement, where they might be happening at higher than light speed but you can’t actually do anything useful with it.

@snacks @arcana @georgia @sun no idea but apparently not

@lain @arcana @georgia I trust aaronsen at least, I don't actually know jack shit about quantum computing so I can't even be properly skeptical

@snacks @lain @georgia @sun for verification you could try to crack passwords I suppose

@snacks @arcana @georgia @sun actually looks like that’s what they did lol, see the article

@Ree @lain @georgia @sun AI is lol

@snacks @lain @georgia @sun lmao, isn’t that like saying we can’t verify any of the massive prime numbers because they can be reached via none computer assisted arithmetic?

@arcana @lain @georgia @snacks I think the issue is quantum calculation has to do a bunch of error correction?

@sun @lain @georgia @snacks possibly yes

@sun @lain @georgia Yes you can break RSA, and basically any cryptographic algorithm based on the exponential behavior of factorizing large prime numbers, with quantum computers. Technically. Not yet in reality.

The current research is mostly theoretical and the current record that could break RSA is a 50-bit integer I think. It's still years if not a decade or more until it becomes feasible at a larger scale. Still a threat right now though, if you care about data retention for feature breaking.

AES fairs much better and the last time I checked ~2 years ago the best theoretical algorithm halved the effective key size.

entropy in a nutshell!

@georgia The fact that this is possible scares the shit out of me and my gut tells me that we should stop all r&d on quantum computing immediately and back away slowly

@sun @arcana @lain @georgia I studied the basics in college and have a somewhat informed take:

That line about the multiverse is REALLY stretching it, and I wouldn't hesitate to call it disingenuous. What people call "quantum parallelism" is sometimes described as computation occuring in multiple parallel universes all at once, but it's more like quantum states are probability distributions that you can reshape using unitary operations (which are just special kinds of matrices). Then when you observe a quantum system, it's like your sampling from that distribution. In a sense it does let you work "in parallel" across a bunch of different states, because that is how linear operations work, but saying that this behavior (which is consistent with almost a century of experimental observation, as far as I know) lends credence to the multiverse theory is ridiculous.