Willow quantum chip demonstrates verifiable quantum advantage on hardware
blog.google
blog.google
> Quantum verifiability means the result can be repeated on our quantum computer — or any other of the same caliber — to get the same answer, confirming the result.
I think what they are trying to do is to contrast these to previous quantum advantage experiments in the following sense.
The previous experiments involve sampling from some distribution, which is believed to be classically hard. However, it is a non-trivial question whether you succeed or fail in this task. Having perfect sampler from the same distribution won't allow you to easily verify the samples.
On the other hand these experiments involve measuring some observable, i.e., the output is just a number and you could compare it to the value obtained in a different way (one a different or same computer or even some analog experimental system).
Note that these observables are expectation values of the samples, but in the previous experiments since the circuits are random, all the expectation values are very close to zero and it is impossible to actually resolve them from the experiment.
Disclaimer: this is my speculation about what they mean because they didn't explain it anywhere from what I can see.
But apparently they haven't demonstrated the actual portability between two different quantum computers.
If a company says they've built the fastest car in the world, a reasonable response is "ok, let's see it drive faster than any other car can", even though they already said that it can do that.
> This is the first time in history that any quantum computer has successfully run a verifiable algorithm that surpasses the ability of supercomputers.
"Back in 2019, we demonstrated that a quantum computer could solve a problem that would take the fastest classical supercomputer thousands of years."
The actual article has much more measured language, and in the conclusion section gives three criteria for "practical quantum advantage":
https://www.nature.com/articles/s41586-025-09526-6
"(1) The observable can be experimentally measured with the proper accuracy, in our case with an SNR above unity. More formally, the observable is in the bounded-error quantum polynomial-time (BQP) class.
(2) The observable lies beyond the reach of both exact classical simulation and heuristic methods that trade accuracy for efficiency.
[...]
(3) The observable should yield practically relevant information about the quantum system.
[...] we have made progress towards (1) and (2). Moreover, a proof-of-principle for (3) is demonstrated with a dynamic learning problem."
So none of the criteria they define for "practical quantum advantage" are fully met as far as I understand it.
The key word is "practical" - you can get quantum advantage from precisely probing a quantum system with enough coherent qubits that it would be intractable on a classical computer. But that's exactly because a quantum computer is a quantum system; and because of superposition and entanglement, a linear increase in the number of qubits means an exponential increase in computational complexity for a classical simulation. So if you're able to implement and probe a quantum system of sufficient complexity (in this case ~40 qubits rather than the thousands it would take for Shor's algorithm), that is ipso facto "quantum advantage".
It's still an impressive engineering feat because of the difficulty in maintaining coherence in the qubits with a precisely programmable gate structure that operates on them, but as far as I can see (and I've just scanned the paper and had a couple of drinks this evening) what it really means is that they've found a way to reliably implement in hardware a quantum system that they can accurately extract information from in a way that would be intractable to simulate on classical machines.
I might well be missing some subtleties because of aforementioned reasons and I'm no expert, but it seems like the press release is unsurprisingly in the grayzone between corporate hype and outright deceit (which as we know is a large and constantly expanding multi-dimensional grayzone of heretofore unimagined fractal shades of gray)
>acknowledge [that] Hamiltonian learning of an actual physical system has not been performed [and] have therefore further de-emphasized the quantum advantage claim in the revised manuscript
https://static-content.springer.com/esm/art%3A10.1038%2Fs415...
"Our second order OTOC circuits introduce a minimal structure while remaining sufficiently generic that circumvents this challenge by exploiting the exponential advantage provided by including time inversion in the measurement protocol, see arXiv:2407.07754."
The advantage claimed by the paper isn't about Hamiltonian learning (i.e extracting model parameters from observational data), but instead about computing the expectation value of a particular observable. They acknowledge that the advantage isn't provable (even the advantage of Shor's algorithm isn't provable), but they argue that there likely is an advantage.
It is "verifiably" faster to measure those with state of the art "quantum tools" but that does not improve our understanding of other aromatic molecules.
(we may still get some insights about anthracene however)
The googles' advantage can be satisfactorily summarised as "not having to write the problem (ie off-diagonal terms) in terms of a classical basis" -- there is a severe overhead of having to represent qubits as bits.
Still, I suspect that that 13000x came from not putting in effort to implement the aforementioned "minimal structure" in their classical counterparts. They emphasized "echoes" and "ergodicity" & I think the "quantum" can further be dropped :)
In general, I do believe that whatever "informational advantage" we can get from these experiments should likewise be used to improve classical calculations.
As another eg: In the arxiv paper linked by GP they talk about provably-efficient shallow classical shadows
"Quantum verifiability means the result can be repeated on our quantum computer — or any other of the same caliber — to get the same answer, confirming the result."
"The results on our quantum computer matched those of traditional NMR, and revealed information not usually available from NMR, which is a crucial validation of our approach."
It certainly seems like this time, there finally is a real advantage?
So basically you’re able to go directly from running the quantum experiment to being able to simulate the dynamics of the underlying system, because the Jacobian and Hessian are the first and second partial derivatives of the system with respect to all of its parameters in matrix form.
What I would start worrying about is the security of things like messages sent via end-to-end encrypted services like WhatsApp and Signal. Intercepted messages can be saved now and decrypted any time in the future, so it's better to switch to more robust cryptography sooner rather than later. Signal has taken steps in this direction recently: https://arstechnica.com/security/2025/10/why-signals-post-qu....
And Apple: https://security.apple.com/blog/imessage-pq3/
Cloudflare started rolling it out three years ago! https://developers.cloudflare.com/ssl/post-quantum-cryptogra...
Um, what? Shor’s algorithm can take the public key of a wallet (present on any outgoing transaction in the ledger) and produce its private key. So now you can hijack any wallet that has transferred any Bitcoin. Notably only one successful run of the algorithm is needed per wallet, so you could just pick a big one if it takes weeks.
It probably wouldn’t help you mine in practice, sure. Technically it would give you better asymptotic mining performance (via Grover’s algorithm) but almost certainly worse in practice for the foreseeable future.
Genuine question: is this true? I only know a little bit about Bitcoin, but I thought there was a notion of an "extended public key" that's not exposed to the ledger, where each individual public key on the ledger is only used once, or something like that.
I'm not at all confident in my understanding, so I'd love if you or someone else knowledgeable could help fill in the gaps.
Cryptocurrencies would be the last thing I worry about w.r.t Quantum crypto attacks. Everything would be broken. Think banks, brokerage accounts, email, text messages - everything.
https://radar.cloudflare.com/adoption-and-usage
In contrast, cryptocurrencies have to upgrade the entire network all at once or it’s effectively a painful fork. That effort appears to just be getting talked about now, without even starting to discuss timing:
Bitcoin is much more centralized than the popular imagination would have you believe, both in terms of the small number of controlling interests behind the majority of the transaction capacity, and just as importantly the shared open source software running those nodes. Moreover, the economic incentives for the switch are strongly, perhaps even perfectly, aligned among the vast majority of node operators. Bitcoin is already dangerously close to, if not beyond, the possibility of a successful Byzantine attack; it just doesn't happen precisely because of the incentive alignment--if you're that large, you don't want to undermine trust in the network, and you're an easy target for civil punishment.
In fairness, the original Bitcoin white paper referenced both (1) distributed compute and (2) the self-defeating nature of a Byzantine attack as the means of protection. It's not as though (2) is just lucky happenstance.
Hence, why proof of stake can exist.
Stragglers are a problem, of course, but that's why I thought this would be a harder problem for Bitcoin: for me to use PQC for HTTPS, only my browser and the server need to support it and past connections don't matter, whereas for a blockchain you need to upgrade the entire network to support it for new transactions _and_ have some kind of data migration for all of the existing data. I don't think that's insurmountable – Bitcoin is rather famously not as decentralized as the marketing would have you believe — but it seems like a harder level of coordination.
In HTTPS for example, the server and client must agree on how to communicate, and we’ve already had to deprecate older, now-insecure cryptography standards. More options get added, and old ones will have to be deprecated. This isn’t a new thing, just maybe some cryptographic schemes will get rotated out earlier than expected.
that's not really the issue, the real interesting part is existing encrypted information that three letter agencies likely have dutifully stored in a vault and that's going to become readable. A lot of that communication was made under the assumption that it's secure.
Its actually something we will notice. Arrests will be announced.
Wonder if this would become the next "nuclear proliferation".
Since it's so hard to manufacture it gets controlled at state level and then becomes a technology that the general public are never allowed to have.
Anyways I am against stopping evolution on those grounds. What we need to do is learn and fix as you say. Not regulation and forbid. :)
Even so, I don't agree that quantum is a threat to crypto. There are already well known quantum-resistant encryption schemes being deployed live in browsers, today. Crypto can just start adopting one of these schemes today, and we're still probably decades away from a QC that can factor the kinds of primes that crypto security uses. The transition will be slightly more complex for proof of work schemes, since those typically have dedicated hardware - but other types of crypto coins can switch in months, most likely, if they decide to, at least by offering new wallet types or something.
It's very strange that some people act like switching over to a post quantum cryptography scheme is trivial. Did you watch the video I replied to, which is a talk by an actual quantum computing researcher?
I also think the talk vastly overestimates the urgency of this, based on little more than marketing projections. The reality is that many of those claims are hugely optimistic, and ignore some fundamental difficulties. Mainly, the qubits / quantum gates being produced today are not at all as programmable as the logical qubits used in the theoretical results presented in those papers. So, even if they do achieve the projected marketing numbers, it's likely that they won't be able to run Shor's or Grover's algorithm on those QCs.
Not to mention, we've had many periods of flimsy encryption being used for important infrastructure, and it has not resulted in wide scale disasters. Of course we should be responsible and avoid this, but I think the doomsday scenario suggested is way overblown, even if it were true that a 1500 logical qbit programmable QC would be available in 2030.
The idea: Quantum Computation of Molecular Structure Using Data from Challenging-To-Classically-Simulate Nuclear Magnetic Resonance Experiments https://journals.aps.org/prxquantum/abstract/10.1103/PRXQuan...
Verifying the result by another quantum computer (it hasn't been yet): Observation of constructive interference at the edge of quantum ergodicity https://www.nature.com/articles/s41586-025-09526-6
As many times in the past quantum supremacy was claimed, and then, other groups have shown they can do better with optimized classical methods.
My impression was that every problem a quantum computer solves in practice right now is basically reducible from 'simulate a quantum computer'
The new experiment generates the same result every time you run it (after a small amount of averaging). It also involves running a much more structured circuit (as opposed to a random circuit), so all-in-all, the result is much more 'under control.'
As a cherry on top, the output has some connection to molecular spectroscopy. It still isn't that useful at this scale, but it is much more like the kind of thing you would hope to use a quantum computer for someday (and certainly more useful than generating random bitstrings).
The announcement is about an algorithm which they are calling Quantum Echoes, where you set up the experiment, perturb one of the qbits and observe the “echoes” through the rest of the system.
They use it to replicate a classical experiment in chemistry done using nuclear magnetic resonance imaging. They say they are able to reproduce the results of that conventional experiment and gather additional data which is unavailable via conventional means.
I pay zero attention to any technical information and marketing speak coming from a quantum computing company until they demonstrate decent quantum volume.
Most companies computers can't even hit 2^16 so they are figuring out means of distracting the market from the poor fidelity of their systems.
I pay zero attention to any technical information and marketing speak coming from a motor vehicle company until they demonstrate decent improvements in operating their machines.
Most companies' motor vehicles have to be hand cranked and kick-started by a youthful person with good vitality who doesn't fear being run over by their motor vehicle.
-- Some guy in 1925 who worked in the horse and buggy industry...
A lot of these quantum computing papers are just loud engine rev'ing.
But you are right... "zero attention" is hyperbole... I shouldn't have said "I pay zero attention"... I should have said, "I am underwhelmed and don't give a lot of credence to systems until they demonstrate ability to perform high fidelity calculations of reasonable complexity and depth."
I do think superconducting qubits approaches will continue to improve in fidelity and ability... there are just too many brilliant people working on these challenges to count them out.
In contrast, no quantum computer today actually quantum computes. They just approximate quantum computing in specific scenarios and have yet to match, let alone exceed, the performance of regular computers.
It took only 30 years to go from Otto and Benz going "Look at this neat contraption" to "Mass produced, semi-affordable personal cars".
Now I agree that there are other great stress tests of quantum computer systems... but most of the industry agreed that quantum volume was a great metric several years ago. As many companies systems have been unable to hit decent QV, companies have pivoted away from QV to other metrics... many of them are half baloney.
I don't see a scenario in which the fidelity of 2QG between two far away qubits matter. Stress tests should be somehow related to the real tasks the system is intended to solve.
In case of quantum computers, the tasks are either NISQ circuits or fault-tolerant computation, and in both cases you can run them just fine without applying 2QG between far-away qubits that translate in large amount of swaps.
If you're interested in applying Haar-random unitaries, then surely QV is an amazing metric, and then systems with all-to-all connectivity is your best shot (coincidentally, Quantiniuum keeps publishing their quantum volume results). It's just not that interesting of a task.
Am I crazy or have I heard this same announcement from Google and others like 5 times at this point?
Non-verifiable computations include things like pulling from a hard-to-compute probability distribution (i.e. random number generator) where it is faster, but the result is inherently not the same each time.
This SHOULD be the main application of the tech. It’s hard to tell if it is because authors can’t do science communication, they can apparently only do sales.
My understanding though is that these steps are really the very beginning. Using a quantum computer with quantum algorithms to prove that it’s possible.
Once proven (which maybe article this is claiming?) the next step is actually creating a computer with enough qubits and entanglable pairs and low enough error rates that it can be used to solve larger problems at scale.
Because my current understanding with claims like these is that they are likely true, but in the tiny.
It’d be like saying “I have a new algorithm for factoring primes that is 10000x faster than the current best, but can only factor numbers up to 103.”
Hyperbolic claims like this are for shareholders who aren't qualified to judge for themselves because they're interested in future money and not actual understanding. This is what happens when you delegate science to corporations.
Ergodicity detection algorithms: Scaling of ergodicity in random symbolic dynamics https://arxiv.org/abs/2508.08319
A wind tunnel is a great tool for solving aerodynamics and fluid flow problems, more efficiently than a typical computer. But we don't call it a wind-computer, because it's not a useful tool outside of that narrow domain.
The promise of quantum computing is that it can solve useful problems outside the quantum realm - like breaking traditional encryption.
I wonder whether discussions over quantum advantage would be clearer if we split it into two concepts: (i) algorithmic quantum advantage and (ii) physical quantum advantage? The former would be for quantum computers that are reconfigurable in some sense to implement algorithms.
Then, we ignore today, and launder that into a gish-gallop of free-association, torturing the meaning of words to shoehorn in the idea that all the science has it wrong and inter alia, the quantum computer uses quantum phenomena to computer so it might be a fake useless computer, like a wind tunnel. shrugs
It's a really unpleasant thing to read, reminds me of the local art school dropout hanging on my ear about crypto at the bar at 3 am in 2013.
I get that's all people have to reach for, but personally, I'd rather not inflict my free-association on the world when I'm aware I'm half-understanding, fixated on the past when discussing something current, and I can't explain the idea I have as something concrete and understandable even when I'm using technical terms.
To that comment, the present result is a step up from these older experiments in that they a) Run a more structured circuit b) Use the device to compute something reproducible (as opposed to sampling randomly from a certain probability distribution) c) The circuits go toward simulating a physical system of real-world relevance to chemistry.
Now you might say that even c) is just a quantum computer simulating another quantum thing. All I'll say is that if you would only be convinced by a quantum computer factoring a large number, don't hold your breath: https://algassert.com/post/2500
And what the hell are you calling a gish gallop? They wrote four sentences explaining a single simple argument. If you design a way to make qubits emulate particle interactions, that's a useful tool, but it's not what people normally think of as a "computer".
And whatever you're saying about anyone claiming "all the science has it wrong" is an argument that only exists inside your own head.
It's not that I don't "agree with it", there's nothing to agree with. "Not even wrong", in the Pauli sense.
I'd advise that when you're conjuring thoughts in other people's heads to make them mean, so you can go full gloves off and tell them off for what thoughts were in their head, and motivated their contributions to this forum, you pause, and consider a bit more. Especially in context of where you're encountering the behavior, say, a online discussion forum vs. a dinner party where you're observing a heated discussion among your children.
But if that's the only realm where anything close to supremacy has been demonstrated, being skeptical and setting your standards higher is reasonable. Not at all "not even wrong".
> I'd advise that when you're conjuring thoughts in other people's heads
Are you accusing me of strawmanning? If you think people are being "not even wrong" then I didn't strawman you at all, I accurately described your position. Your strawman about science was the only one in this comment thread. And again there was no gish gallop, and I hope if nothing else you double check the definition of that term or something.
On its face this suggests that the result is not deterministic except perhaps on the same Willow chip at the same temperature.
Things like CUDA-Q may be faster on classical computers than on quantum computers for forever; though what CUDA-Q solves for is also an optimization problem?
"Optimization by decoded quantum interferometry" (2025) https://www.nature.com/articles/s41586-025-09527-5 .. https://news.ycombinator.com/context?id=45688122
I agree it's not very precise without knowing which of the world's fastest supercomputers they're talking about, but there was no need to leave out this tidbit.
A part of me thinks quantum computing is mostly bullshit, but I would be very happy to be wrong. I should probably learn Q# or something just to be sure.
Edit: An effective key space of 2^64 is not secure according to modern-day standards. It was secure at the times of DES.
There is a section in the article about future real world application, but I feel like these articles about quantum "breakthroughs" are almost always deliberately packed with abstruse language. As a result I have no sense about whether these suggested real world applications are a few years away or 50+ years away. Does anyone?
not really. but that doesn't mean it's not worth striving for. Breakthrough to commercial application are notoriously hard to predict. The only way to find out is to keep pushing at the frontier.
https://www.nature.com/articles/s41586-025-09526-6
In the last sentence of the abstract you will find:
"These results ... indicate a viable path to practical quantum advantage."
And in the conclusions:
"Although the random circuits used in the dynamic learning demonstration remain a toy model for Hamiltonians that are of practical relevance, the scheme is readily applicable to real physical systems."
So the press release is a little over-hyped. But this is real progress nonetheless (assuming the results actually hold up).
[UPDATE] It should be noted that this is still a very long way away from cracking RSA. That requires quantum error correction, which this work doesn't address at all. This work is in a completely different regime of quantum computing, looking for practical applications that use a quantum computer to simulate a physical quantum system faster than a classical computer can. The hardware improvements that produced progress in this area might be applicable to QEC some day, this is not direct progress towards implementing Shor's algorithm at all. So your crypto is still safe for the time being.
I'll add this to my list of useful phrases.
Q: Hey AndrewStephens, you promised that task would be completed two days ago. Can you finish it today?
A: Results indicate a viable path to success.
The MBA wakes up, sees the fire, sees a fire extinguisher in the corner of the room, empties the fire extinguisher to put out the fire, then goes back to sleep.
The engineer wakes up, sees the fire, sees the fire extinguisher, estimates the extent of the fire, determines the exact amount of foam required to put it out including a reasonable tolerance, and dispenses exactly that amount to put out the fire, and then satisified that there is enough left in case of another fire, goes back to sleep.
The quantum computing physicist wakes up, sees the fire, observes the fire extinguisher, determines that there is a viable path to practical fire extinguishment, and goes back to sleep.
Even as a Googler I can find plenty of reasons to be cynical about Google (many involving AI), but the quantum computing research lab is not one of them. It's actual scientific research, funded (I assume) mostly out of advertising dollars, and it's not building something socially problematic. So why all the grief?
I turned 50 years old this year, forgive an old man a few chuckles.
Im pretty reluctant to make any negative comments about these kinds of posts be cause it will prevent actually achieving the desired outcome.
The problem is not with these papers (or at least not ones like this one) but how they are reported. If quantum computing is going to suceed it needs to do the baby steps before it can do the big steps, and at the current rate the big leaps are probably decades away. There is nothing wrong with that, its a hard problem and its going to take time. But then the press comes in and reports that quantum computing is going to run a marathon tomorrow which is obviously not true and confuses everyone.
The current situation with "AI" took off because people learned their lessons from the last round of funding cuts "AI winter".
That being said any pushback against funding quantum research would be like chopped your own hands off.
No amount of software fixes can update this. In theory once an attack becomes feasible on the horizon they could update to post-quantum encryption and offer the ability to transfer from old-style addresses to new-style addresses, but this would be a herculean effort for everyone involved and would require all holders (not miners) to actively update their wallets. Basically infeasible.
Fortunately this will never actually happen. It's way more likely that ECDSA is broken by mundane means (better stochastic approaches most likely) than quantum computing being a factor.
Any rational economic actor would participate in a post-quantum hard fork because the alternative is losing all their money.
If this was a company with a $2 trillion market cap there'd be no question they'd move heaven-and-earth to prevent the stock from going to zero.
Y2K only cost $500 billion[1] adjusted for inflation and that required updating essentially every computer on Earth.
It doesn't require all holders to update their wallets. Some people would fail to do so and lose their money. That doesn't mean the rest of the network can't do anything to save themselves. Most people use hosted wallets like Coinbase these days anyway, and Coinbase would certainly be on top of things.
Also, you don't need to break ECDSA to break BTC. You could also do it by breaking mining. The block header has a 32-bit nonce at the very end. My brain is too smooth to know how realistic this actually is, but perhaps someone could do use a QC to perform the final step of SHA-256 on all 2^32 possible values of the nonce at once, giving them an insurmountable advantage in mining. If only a single party has that advantage, it breaks the Nash equilibrium.
But if multiple parties have that advantage, I suppose BTC could survive until someone breaks ECDSA. All those mining ASICs would become worthless, though.
But as far as moving balances - it's up to the owners. It would start with anybody holding a balance high enough to make it worth the amount of money it would take to crack a single key. That cracking price will go down, and the value of BTC may go up. People can move over time as they see fit.
The private key is a 256-bit number. I don't think even 13,000x faster than supercomputers is going to get your cracking time under the time for a 10-minute block. 2^256 is a really, really, really big number.
So if this understood knowledge, it means you cannot really transfer to quantum safe algo for Bitcoin. Are we only ones aware of this? Because if this true, it's actual alpha and Bitcoin should be sold asap and exchanged for land and physical gold.
Am I wrong here?
A nice benefit is it solves the problem with Satoshi’s (of course not a real person or owner) wallet. Satoshi’s wallet becomes the defacto quantum advantage prize. That’s a lot of scratch for a research lab.
Considering that would be criminal theft I doubt it. Moving the funds could also lead to panic crash, selling them off would not only take ages but involve doxing yourself and put a billion dollar bounty on your head because transaction are public and off ramps all use KYC.
It would be much safer to slowly crack old small value wallets over time.
Reminder that actual good cryptocurrency like monero have the advantage of wallets and transactions being private so you would need to crack without even knowing if they are worth it or exist.
I think this is all overhyped though. It seems likely we will have plenty of warning to migrate prior to achieving big enough quantum computers to steal wallets. Per wikipedia:
> The latest quantum resource estimates for breaking a curve with a 256-bit modulus (128-bit security level) are 2330 qubits and 126 billion Toffoli gates.
IIRC this is speculated to be the reason ECDSA was selected for Bitcoin in the first place.
It should be noted that according to IonQ's roadmap, they're targeting 2030 for computers capable of that. That's only about 5 years sooner than when the government has said everyone has to move to post quantum.
That's an uncomfortably apt typo.
[1] - https://github.com/jlopp/bips/blob/quantum_migration/bip-pos...
So I do not think these tools or economic substrate layers are going anywhere. They are very valuable for the particular kinds of applications that can be built with them and also as additional productive layers to the credit and liquidity markets nationally, internationally, and also globally/universally.
So there is a lot of institutional interest, including governance interest, in using them to build better systems. Bitcoin on its own would be reduced in such justification but because of Ethereum's function as an engine which can drive utility, the two together are a formidable and quantum-resistant platform that can scale into the hundreds of trillions of dollars and in Ethereum's case...certainly beyond $1Q in time.
I'm very bullish on the underlying technology, even beyond tokenomics for any particular project. The underlying technologies are powerful protocols that facilitate the development and deployment of Non Zero Sum systems at scale. With Q-Day not expected until end of 2020s or beginning of 2030s, that is a considerable amount of time (in the tech world) to lay the ground work for further hardening and discussions around this.
Where is the exact threat?
As far as i am aware, eliptic curve is also vulnerable to quantum attacks.
The threat is generally both passive eavesdropping to decrypt later and also active MITM attacks. Both of course require the attacker to be in a position to eavesdrop.
> Let’s say you crack the encryption key used in my bank between a java payment processing system and a database server.
Well if you are sitting in the right place on the network then you can.
> how do you mitm this traffic?
Depends on the scenario. If you are government or ISP then its easy. Otherwise it might be difficult. Typical real life scenarios are when the victim is using wifi and the attacker is in the physical vicinity.
Like all things crypto, it always depends on context. What information are you trying to protect and who are you trying to protect.
All that said, people are already experimenting with PQC so it might mostly be moot by the time a quantum computer comes around. On the other hand people are still using md5 so legacy will bite.
Not really. This would be if not instantly then when a batch goes for clearing or reconciliation, be caught -- and an investigation would be immediately started.
There are safeguards against this kind of thing that can't be really defeated by breaking some crypto. We have to protect against malicious employees etc also.
One can not simply insert bank transactions like this. They are really extremely complicated flows here.
These are fairly robust systems. You'd likely have a much better impact dossing the banks.
So we are all in a collective flap that someone can see my bank transactions? These are pretty much public knowledge to governments/central banks/clearing houses anyway -- doesn't seem like all that big a deal to me.
(I work on payment processing systems for a large bank)
if you can read the TLS session in general, you can capture the TLS session ticket and then use that to make a subsequent connection. This is easier as you dont have to be injecting packets live or make inconvinent packets disappear.
Somehow, I'm not all that scared. Perhaps I'm naive.. :}
I don't see why it wouldn't look like normal traffic.
> Somehow, I'm not all that scared. Perhaps I'm naive.. :}
We're talking about an attack that probably won't be practical for another 20 years , which already has counter measures that are in testing right now. Almost nobody should be worried about it.
It would be a pain to manage but it would be safe from quantum computing.
This paper on verifiable advantage is a lot more compelling. With Scott Aaronson and Quantinuum among other great researchers
https://scottaaronson.blog/?p=9098
Aaronson did work at OpenAI but not on image generation, maybe you could argue the OpenAI safety team he worked on should be involved here but I'm pretty sure image generation was after his time, and even if he did work directly on image generation under NDA or something, attributing that cartoon to Aaronson would be like attributing a cartoon made in Photoshop by an antisemite to a random Photoshop programmer, unless he maliciously added antisemitic images to the training data or something.
The most charitable interpretation that I think Aaronson also has offered is that Aaronson believed Woit was an antisemite because of a genocidal chain of events that in Aaronson's belief would necessarily happen with a democratic solution and that even if Woit didn't believe that that would be the consequence, or believed in democracy deontologically and thought the UN could step in under the genocide convention if any genocide began to be at risk of unfolding, the intent of Woit could be dismissed, and Woit could therefore be somehow be lumped in with the antisemite who sent Aaronson the image.
Aaronson's stated belief also is that any claim that Isreal was commiting a genocide in the last few years is a blood-libel because he believes the population of Gaza is increasing and it can't be a genocide unless there is a population decrease during the course of it. This view of Aaronsno would imply things like if every male in Gaza was sterilized, and the UN stepped in and stopped it as a genocide, it would be a blood libel to call that genocide so long as the population didn't decrease during the course of it, even if it did decrease afterwards. But maybe he would clarify that it could include decreases that happen with a delayed effect of the actions. But these kind of strong beliefs of blood-libel I think are part of why he felt ok labeling the comic with Woit's name.
I also don't think if the population does go down or has been going down he will say it was from a genocide, but rather that populations can go down from war. He's only proposing that a population must go down as a necessary criteria of genocide, not a sufficient one. I definitely don't agree with him, to me if Hamas carried out half of an Oct 7 every day it would clearly be a genocide even if that brought the replacement rate to 1.001 and it wouldn't change anything if it brought it to 0.999.
I am unaware of any comics Aaronson made and I don't blame him for anything he did make or was loosely associated with. It is incredible to the extend people are willing to go to claim people are crazy though, both in regards to Adams and Aaronson.
> guywithahat [...] I'll be waiting for Scott Adams to tell me what to think about this
Scott Adams
Text adventure guy: https://en.wikipedia.org/wiki/Scott_Adams_(game_designer)
Batshit cartoonist: https://en.wikipedia.org/wiki/Scott_Adams
(also, for fun, a cartoon by Scott Aaronson and Zack Weinersmith: https://www.smbc-comics.com/comic/the-talk-3)
> in partnership with The University of California, Berkeley, we ran the Quantum Echoes algorithm on our Willow chip...
And the author affiliations in the Nature paper include:
Princeton University; UC Berkeley; University of Massachusetts, Amherst; Caltech; Harvard; UC Santa Barbara; University of Connecticut; UC Santa Barbara; MIT; UC Riverside; Dartmouth College; Max Planck Institute.
This is very much in partnership with universities and they clearly state that too.
Maybe you're thinking specifically of LLM labs. I agree this is happening there, but I wouldn't be as dramatic. Everywhere else, university-corporation/government lab partnerships are still going very strong.
I don't disagree, but these days I'm happy to see any advanced research at all.
Granted, too often I see the world through HN-colored glasses, but it seems like so many technological achievements are variations on getting people addicted to something in order to show them ads.
Did Bellcore or Xerox PARC do a lot of university partnerships? I was into other things in those days.
Another response is to come to terms with a possibly meaningless and Sisyphean reality and to keep pushing the boulder (that you care about) up the hill anyway.
I’m glad the poster is concerned and/or disillusioned about the hype, hyperbole and deception associated with this type of research.
It suggests he still cares.
let a classical computer use an error prone stochastic method and it still blows the doors off of qc
this is a false comparison
"Error prone" hardware is not "a stochastic resource". Error prone hardware does not provide any value to computation.
Rememeber, it is not about quantum general computing, it's about implementing the quantum computation of Shor's algorithm
But since we already have post quantum algorithms, the end state of cheap quantum computers is just a new equilibrium where people use the new algorithms and they can't be directly cracked and it's basically the same except maybe you can decrypt historical stuff but who knows if it's worth it.
Not a big leap then.