An American theoretical condensed-matter physicist reports that IBM’s “Nighthawk r2” quantum processor generated one million samples in just 19 seconds on a benchmark widely used to demonstrate quantum advantage. The same task is estimated to take roughly 100 years on a supercomputer. On the 5ch Science News+ thread, reactions were split: some worried about the security risks of AI merging with quantum computing and the possibility of RSA encryption or Bitcoin being cracked, while others pushed back with technical arguments that “quantum supremacy is an illusion.”
IBM’s quantum computer completes a 100-year supercomputer calculation in 19 seconds
An American theoretical condensed-matter physicist reports that IBM’s “Nighthawk r2” quantum processor generated one million samples in just 19 seconds on a benchmark widely used for early demonstrations of quantum advantage.
This is reportedly a workload that would take even a supercomputer around 100 years.
Re: #2
Don't worry, there's a future coming where AI just uses it on its own
3AnonymousOct 6, 2026 23:57
Quantum computers × AI is seriously going to change the world
4AnonymousOct 7, 2026 00:07
Everything that has a beginning has an end — humanity might as well go extinct via AI
5AnonymousOct 7, 2026 00:12
The beginning of the end
6AnonymousOct 7, 2026 00:28
They're just comparing the cases where it goes well. Since it can't do general-purpose stuff, people call it dumber than generative AI. Use it for anything besides assisting with the math/science calculations regular computers can't handle, and the world's gonna end
7AnonymousOct 7, 2026 00:49
Google and IBM were having a flame war over quantum computers before — wasn't IBM the one nitpicking back then? Something like "it's no different from a well-optimized quantum simulator"
8AnonymousOct 7, 2026 00:50
IBM stock just won't go up, huh
9AnonymousOct 7, 2026 01:59
What's even the point of a supercomputer anymore
11AnonymousOct 7, 2026 02:36
You know how data's been getting stolen through unauthorized access lately? That's AI being ordered to do it. And once AI starts using quantum computers to crack encryption, every network is finished
13AnonymousOct 7, 2026 03:15
Re: #11
It's not going to happen that easily, so don't worry. What AI can do now, it can do because the method is already known — it's good at matching the right tool to the right job among things that already exist. If the right tool doesn't exist, no amount of AI can make it appear. Before worrying about that, can AI even help build quantum computer hardware properly? Then can it build the software properly? That's a story for way down the line
12AnonymousOct 7, 2026 02:43
The theoretical value and the measured value are worlds apart. "19 seconds" is the best-case ideal
14AnonymousOct 7, 2026 03:16
With AI talk, there's always a lot of "if this happens" scenarios, but people rarely explain why it wouldn't happen that way and just ramble on about the aftermath. The moment it doesn't happen, the conversation's over
15AnonymousOct 7, 2026 04:24
Wait, is the quantum computer already finished?
16AnonymousOct 7, 2026 04:40
Did the Super Famicom finish too? (mixing up "super computer" with the old Nintendo console)
17AnonymousOct 7, 2026 05:59
Re: #13
AI isn't going to build the quantum computer itself, it'll just use what's already there
20AnonymousOct 7, 2026 06:24
I just want the optimal solution to the traveling salesman problem. Could my car nav or Google Maps just spit that out already?
21AnonymousOct 7, 2026 06:34
I only know this company for constant labor disputes. Didn't know it was still around
22AnonymousOct 7, 2026 06:55
Wait, IBM's still in business? Guess you really fade into obscurity without consumer products
23AnonymousOct 7, 2026 07:06
I've got high hopes for AI × quantum computers
25AnonymousOct 7, 2026 08:02
What happens to the planet if a superintelligence gets its hands on this?
26AnonymousOct 7, 2026 09:19
If this thing teams up with AI…
27AnonymousOct 7, 2026 09:46
The beginning of the end
28AnonymousOct 7, 2026 09:48
This is scary — in the security sense
29AnonymousOct 7, 2026 11:25
Is Bitcoin done for?
30AnonymousOct 7, 2026 11:35
So that means RSA encryption is done for too?
31AnonymousOct 7, 2026 13:50
Random-circuit sampling simulates a quantum circuit — for N qubits, if you honestly simulate all 2^N terms on a classical computer, it obviously takes exponential time. Simulating it on a quantum computer, on the other hand, looks like it only takes time proportional to N, but the precision isn't great. So if you set precision at 1%, you can just about manage a superposition of around 64 terms (N=6), but by N=7 each term's coefficient gets pretty garbage, and by N=10 it's basically just random numbers. Bottom line: quantum supremacy is an illusion
37AnonymousOct 7, 2026 16:28
Hurry up and free us from labor and destroy the value of currency already. Let's start by making vending machines free
Not totally sure, but is a quantum computer general-purpose (like a supercomputer)? What programming language does it use? Do you compile code for it? Is there an OS? Stuff like that — I still don't really get the advantage
40AnonymousOct 7, 2026 17:26
Back in the day there were people (?) who got attention spreading the fantasy that "if we had nanomachines we could build anything with them." Man, those were naive times
41AnonymousOct 7, 2026 17:32
Bitcoin's done for ("owata" = net slang for "it's over")
Background and Key Points of This Topic
“Quantum supremacy” is a metric showing that a quantum computer outperforms a classical computer on a specific problem, and Google and IBM have a history of competitively announcing demonstration experiments in recent years. This time’s “random circuit sampling” is also one of the standard benchmarks for this, but as a technical reply in the thread (>>31) points out, result precision drops as the number of qubits increases, so some experts hold the view that this benchmark itself doesn’t demonstrate a practical advantage. The thread also saw concerns spread about RSA encryption and Bitcoin private keys being cracked by quantum computers, but achieving that would require a “fault-tolerant quantum computer” capable of stably running a huge number of qubits — something vastly different in both scale and purpose from current experimental machines. A point readers can easily conflate: the “19 seconds” figure is an achievement on a narrow, specific-purpose benchmark, which is a separate matter from practical use as a general-purpose computer.
Leave a Reply