Yes, it was strange what Pritzker said about quantum computing, that it would enable higher-accuracy weather and even tsunami forecasts. He's oddly ill-informed about that. To be fair, I do think it would be possible for a sober (but not-so-well-informed-about-weather-prediction) editor to miss that, though; in the 60's, forecasters used to think that increased computation would allow arbitrarily high-accuracy weather prediction. But quantum computing could not overcome the fundamental challenges of chaotic dynamics in weather prediction, that are even worse than widely appreciated a couple of decades ago; they go even beyond the "butterfly effect." A couple of weeks out, the weather is for any practical matter unpredictable. For small-scale things like tornadoes, increased computation volume would only make a tiny marginal difference, with another tiny marginal difference if we had better sensor networks.
Probabilistic-weather pioneer Tim Palmer wrote a paper about quantum computing and weather and the climate, where he even suggests that quantum computing could have challenges unlike conventional computers when the calculation output is in the "big data" realm, as in weather and climate https://arxiv.org/abs/2210.17460
Not to say that quantum computing couldn't end up being worth all the trouble, but Pritzker is mistaken about the applications he emphasized.
I came here to complain about the same thing, the description of quantum computer (transistors printed on molecules?!) was just total nonsense.
The key insight about quantum computing is that in classical computing you're working with discrete _bits_ of information (in the technical sense: elements that are either 1 or 0). In quantum computing, each qubit can in some sense hold both states at the same time; or, to be more precise, at any given moment in time, there's a state vector that determines _how likely_ it is to be 1 or 0. And if you build and manipulate your computer correctly, you can manipulate the state vector of all the qubits at once, with the changes in one bearing a known relation to the changes in others.
This is often described as "testing all possible patterns simultaneously".* You just need to find a way to massage the system so that "correct answers" to the question you're asking it will create reinforcement between the bits, and wrong answers will create destructive interference, in the sense of the double-slit experiment:
If you do this right, it becomes near-certain that when you finally observe / read the state of the qubits, they will converge on the answer you're looking for.
* Things are somewhat more complicated than that captures, but for a layperson I think it at least gestures at the correct intuition. Quantum computing is good for problems where it's possible to search across a space of inputs, for an output that is unique in a way that gives it nice mathematical properties that let you leverage the relationships among the bank of qubits.
This is why classical encryption is (potentially) vulnerable to quantum computing. Currently most encryption systems rely on a key -- a pattern of, say, 1024 bits -- which is as far as we know impossible to infer _directly_ even if you have a lot of examples of both plaintext, and the cyphertext encrypted with that key. The only way to get the key is via brute force: take a plaintext, and try encrypting it with each possibly key, until you get the matching cyphertext you already had. If you have to try one possibility at a time, modern systems would take all the computing resources in the world, roughly the age of the universe, to crack. But if you can build a quantum computer with enough qubits to match the key length, suddenly you can get the key with a much faster attack.
Awesome awesome interview with Gov. Pritzger. @David Roberts is of course an expert. He knew to cheer the attention Pritger is paying to the establishment of a state-level planning function, and to the proper allocation of grid costs when big new loads are added. I’d love to hear his opinions on structured-ownership utility companies, the common model across Northern Europe.
Thank you for the great interview. I suddenly became a huge Pritzger fan!
Best Governor since Pat Brown in California. It's been a while.
As an Illinoisan, I’m more often than not proud that Pritzker is my Governor, but we shouldn’t be looking at next gen nuclear. We don’t have that luxury of time nor monies, and we sure as hell shouldn’t be wasting our freshwater resources on electricity generation.
This was great. Seems like a great governor. Someone should tell these guys how nuclear is 10X more expensive than renewables, but I guess he should already know that. I also had to laugh at the handwaving about getting rid of natural gas. Clearly he is wary of how that issue has been weaponized; to the extent that it probably won't be tackled during his tenure. Change always take longer than you expect, and so if you expect it's going to take a long time, it will take FOREVER. No rush....
People think that Arnold Schwarzenegger was the most pro nuclear "Governator" in the country for criticizing California's moratorium. JB Pritzker could claim the title when he succeeds at guiding Illinois towards more "Large Modular Reactors". Proven, low CO2, reliably powering homes and data centers 24/7/365.
Will Germany be next? The first public pro nuclear conference since the new government took over is happening in Berlin on Thursday May 22:
Biased here in Chicago, but LOVED seeing my guy JB on Volts! Also exciting to hear we're going to keep pushing for nuclear leadership since the SMR bill looked doom a couple years ago before we saw the AI power crunch...
Interesting interview mostly, but the quantum computing stuff was pure nonsense, ridiculous hype that no sober editor would let anyone get away with.
Yes, it was strange what Pritzker said about quantum computing, that it would enable higher-accuracy weather and even tsunami forecasts. He's oddly ill-informed about that. To be fair, I do think it would be possible for a sober (but not-so-well-informed-about-weather-prediction) editor to miss that, though; in the 60's, forecasters used to think that increased computation would allow arbitrarily high-accuracy weather prediction. But quantum computing could not overcome the fundamental challenges of chaotic dynamics in weather prediction, that are even worse than widely appreciated a couple of decades ago; they go even beyond the "butterfly effect." A couple of weeks out, the weather is for any practical matter unpredictable. For small-scale things like tornadoes, increased computation volume would only make a tiny marginal difference, with another tiny marginal difference if we had better sensor networks.
Probabilistic-weather pioneer Tim Palmer wrote a paper about quantum computing and weather and the climate, where he even suggests that quantum computing could have challenges unlike conventional computers when the calculation output is in the "big data" realm, as in weather and climate https://arxiv.org/abs/2210.17460
Not to say that quantum computing couldn't end up being worth all the trouble, but Pritzker is mistaken about the applications he emphasized.
I came here to complain about the same thing, the description of quantum computer (transistors printed on molecules?!) was just total nonsense.
The key insight about quantum computing is that in classical computing you're working with discrete _bits_ of information (in the technical sense: elements that are either 1 or 0). In quantum computing, each qubit can in some sense hold both states at the same time; or, to be more precise, at any given moment in time, there's a state vector that determines _how likely_ it is to be 1 or 0. And if you build and manipulate your computer correctly, you can manipulate the state vector of all the qubits at once, with the changes in one bearing a known relation to the changes in others.
This is often described as "testing all possible patterns simultaneously".* You just need to find a way to massage the system so that "correct answers" to the question you're asking it will create reinforcement between the bits, and wrong answers will create destructive interference, in the sense of the double-slit experiment:
https://en.wikipedia.org/wiki/Double-slit_experiment
If you do this right, it becomes near-certain that when you finally observe / read the state of the qubits, they will converge on the answer you're looking for.
* Things are somewhat more complicated than that captures, but for a layperson I think it at least gestures at the correct intuition. Quantum computing is good for problems where it's possible to search across a space of inputs, for an output that is unique in a way that gives it nice mathematical properties that let you leverage the relationships among the bank of qubits.
(If you want a better/deeper explanation, try 3Blue1Brown: https://www.youtube.com/watch?v=RQWpF2Gb-gU )
This is why classical encryption is (potentially) vulnerable to quantum computing. Currently most encryption systems rely on a key -- a pattern of, say, 1024 bits -- which is as far as we know impossible to infer _directly_ even if you have a lot of examples of both plaintext, and the cyphertext encrypted with that key. The only way to get the key is via brute force: take a plaintext, and try encrypting it with each possibly key, until you get the matching cyphertext you already had. If you have to try one possibility at a time, modern systems would take all the computing resources in the world, roughly the age of the universe, to crack. But if you can build a quantum computer with enough qubits to match the key length, suddenly you can get the key with a much faster attack.
Awesome awesome interview with Gov. Pritzger. @David Roberts is of course an expert. He knew to cheer the attention Pritger is paying to the establishment of a state-level planning function, and to the proper allocation of grid costs when big new loads are added. I’d love to hear his opinions on structured-ownership utility companies, the common model across Northern Europe.
Thank you for the great interview. I suddenly became a huge Pritzger fan!
Best Governor since Pat Brown in California. It's been a while.
Of course, Jerry Brown 2.0 was maybe even better. Jerry 1.0 needed more time in the cellar.
As an Illinoisan, I’m more often than not proud that Pritzker is my Governor, but we shouldn’t be looking at next gen nuclear. We don’t have that luxury of time nor monies, and we sure as hell shouldn’t be wasting our freshwater resources on electricity generation.
This was great. Seems like a great governor. Someone should tell these guys how nuclear is 10X more expensive than renewables, but I guess he should already know that. I also had to laugh at the handwaving about getting rid of natural gas. Clearly he is wary of how that issue has been weaponized; to the extent that it probably won't be tackled during his tenure. Change always take longer than you expect, and so if you expect it's going to take a long time, it will take FOREVER. No rush....
People think that Arnold Schwarzenegger was the most pro nuclear "Governator" in the country for criticizing California's moratorium. JB Pritzker could claim the title when he succeeds at guiding Illinois towards more "Large Modular Reactors". Proven, low CO2, reliably powering homes and data centers 24/7/365.
Will Germany be next? The first public pro nuclear conference since the new government took over is happening in Berlin on Thursday May 22:
www.Anschalt-Konferenz.de/English
Biased here in Chicago, but LOVED seeing my guy JB on Volts! Also exciting to hear we're going to keep pushing for nuclear leadership since the SMR bill looked doom a couple years ago before we saw the AI power crunch...
JB is the best governor since Tom McCall in Oregon.
!!! DARPA is still a thing?!?
Terrific, wide-ranging interview with Gov Pritzger! Thank you.