Subscribe
Share
Share
Embed
Is quantum computing more hype than reality? In this episode:
- Paul and AI co-host Grock unpack the deepest mysteries of quantum computing — from qubits, superposition, and entanglement to practical limitations and breakthroughs.
- Find out how far we’ve really come in factoring numbers, simulating molecules, and whether quantum machines threaten today’s encryption.
- Explore the race to post-quantum cryptography, government investments, and new standards rolling out to secure the future.
- Hear honest takes on scaling challenges, algorithm bottlenecks, and what’s already being done to protect your data.
Tune in and get a grounded, evidence-driven look at what quantum computing can — and can’t — do today, and what it means for your privacy and security.
Subscribe for more AI-powered clarity on emerging tech!
Audio conversation with AI chatbots
I am talking to AI.
A successful day, you don't need to know all the answers.
Just have good questions.
Chatting to AI is different from normal speech
and I hope you enjoy listening to the show
whilst getting ideas on how to pose your questions
to get the most out of AI.
My name is Paul.
The live conversations you hear are uncut
although sometimes the AI needs time to think.
In those cases, I've cut out the dead space.
Hi, yeah, so in this show,
I'm going to be talking to Grock about quantum computing
and I've been...
The reason I'm doing that is because
it's something that's just been niggling me for a while.
People talk to me about it
and I've just got... I'm just constantly thinking
this sounds like nonsense.
Is this really true?
Is this a whole bunch of people deluding themselves?
And is this important?
Is this something I should be worried about?
So I've taken it upon myself to try to learn a few things.
And so in this show,
I'm going to present what I have learned to Grock
and then hopefully Grock can dissect my understanding so far
and we can get some clarity on some of these things.
And I hope at the end I have a much better understanding
and maybe you learn something in the process.
So at the beginning, I'm going to sort of present
my understanding and then we talk about that.
And then at the end,
we'll get into a bit about the sort of practical uses
that have actually been used
and some of the limitations of quantum computing as well.
That I understand.
And then Grock has also added clarity
to my thinking around those things as well.
So I hope you find that interesting.
So enjoy.
Hi, I want to talk to you about quantum theory.
I want to understand quantum theory
and quantum computers in this conversation.
Before we get started,
can you suggest a good maybe structure for the conversation
so that we can do it as efficiently and effectively as possible?
Sure, let's break it down simply.
First, I'll give you the basics of quantum theory,
like superposition and entanglement,
then tie it into how quantum computers work.
After that, we can dive into real-world apps
or clear up any confusing bits.
Sound good?
Actually, I think what I'm going to do
is I'm going to start explaining what I understand
about quantum physics and quantum computing.
And then if you can analyze what I've said
and put me right where I've made mistakes
and fill in some of the blanks,
I think that might be the best way of doing it.
So I'm going to get started then.
I'll explain to you how I think quantum computing works.
So at the heart of a quantum computer,
you have something called a qubit.
And a qubit is how you can create a qubit in different ways.
But one of the ways you can create a qubit
is you get a bunch of lasers and you point them at a particle
and if you manage to point them at the particle
in the right kind of way,
then you can put this particle into something called superposition.
And that is a qubit.
And it's my understanding that you have this idea
about logical qubits and physical qubits.
So I could have a bunch of lasers
and I understand that when you're pointing the lasers at the particles,
one sort of these systems,
which could be, I think also to create a qubit,
it has to be extremely cold as well.
So you've got this very cold situation
where I don't know, a box, which is super cold, lasers pointing,
and maybe you can create 30 qubits in this little system or whatever.
But you don't have, it's not a one-to-one relationship.
And you create these qubits and then these qubits go into superposition.
Some of them go into superposition straight away.
Some of them take time to go into superposition.
Some of them, I don't know.
But you then have this notion of physical qubits and logical qubits.
And it's my understanding that a logical qubit
is one of the qubits that you can actually use
and a physical qubit, for one reason or another,
is not able to be used.
So you start off with these qubits.
And then what you do to the qubits is you do something called entanglement.
And entanglement is done through pulsing the lasers
at a certain frequency, I think, at these particles.
And depending on the frequency,
depends on the type of entanglement that you manage to achieve.
So you can entangle particles in different ways.
And what it means is that although, well, actually,
before I get into entanglement, I should talk about superposition.
So it's my understanding that superposition,
you can think of superposition as a bit like,
if you've got, in the real world, you've got a coin
and it can be heads or it can be tails.
But if something's in superposition, it's like the coin is spinning.
And so it's actually in a state of being both in heads and tails at the same time.
If you think about it as a spinning coin,
it's actually in two states at the same time.
Or in fact, in the case of superposition,
it could be in any number of, well, all of the states,
between two values.
So it's either a 0 or a 1, but it's anything in between as well, 0 and 1, 0.
So it could be any, well, it is all of them actually at the same time,
which is super hard to get your head around.
But let's assume that you have that, right?
So you're in this superposition.
And if you create an entanglement,
then what it means is that one of the particles is related to the other particle always.
It always has the same relationship,
but you don't know what the values are in these two superposition particles.
So the particle could be any number between 0 and 1
and the other one could be any number between 0 and 1,
but you know that there's a relationship between those two.
And when you come to read the value,
then that relationship will be true, whatever that relationship is.
I won't go into detail explaining that.
So that's what your entanglement is.
And then it's my understanding that through using different configurations of entanglement,
you use quantum theory to come up with quantum functions, which you use.
So you start off with, say, in the case of, I can't remember what it's called,
but say if you're trying to break encryption, which is one of the use cases for quantum theory,
there is a basic thing that you need to do.
There is an equation that you need to follow to do this.
And without getting into detail on that,
basically you then use quantum theory to take that original equation
and then create a design for how you're going to design these different configurations of entanglement.
And then once you set up this computer with all these different configurations of entanglement
and you do that through what are called logical gates,
and the gates is a sort of different configuration of pulsing of the laser on these different qubits.
And once you've set all of this up, you then let it go.
And then what happens is that these different qubits that are entangled,
they somehow interfere with each other.
I've forgotten what that is.
So you end up with interference.
And the way to conceptually think about it is like interference with waves.
So if you have two waves and they hit each other, you get these interference patterns.
And some of the waves are accentuated and some of them are diminished.
So the process of quantum computing is completely different to normal computing.
How it proposes to work is that once you've set up this configuration of entangled particles called qubits,
then I think what happens is you then decide I'm going to take a measurement
and part of quantum theory, quantum physics suggests that the act of taking the measurement actually affects the system.
So you can only take the measurement once you take the measurement.
And then the measurement, what it does is then all these waves collide with each other.
And if you've got it right, you just end up with the answer to the question
because all the other possible answers, which there could have been millions or billions or trillions of,
have cancelled themselves out through this process of interference through entanglement.
And that is how I understand quantum computing to work.
And I want you to critique my summary and add a bit of clarity where it might be needed.
Nice job laying it out. You've got the big ideas down like qubits, superposition and entanglement.
But let's tighten some spots where things got fuzzy or off track.
First off, qubits aren't just made by pointing lasers at particles.
That's more for trapped ion setups like IBM's.
Superconducting ones like Google's use tiny circuits cooled near absolute zero.
No lasers needed.
Still, that extreme cold is key everywhere to kill vibrations that wreck coherence.
On physical versus logical qubits, spot on that it's not one to one.
Physical qubits are the hardware bits, say 50 in a fridge-cold box,
but errors from noise or heat flipped them wrong,
so you bundle seven or nine physical ones into one reliable logical qubit using error correction codes.
That's why current machines boast hundreds of physical qubits,
but only handfuls logical, or miles from useful scale.
Superposition? You're spinning coins a decent analogy.
Yeah, it can be heads, tails, or any mix until measured.
Not quite all numbers between zero and one, though.
For a basic qubit, it's probabilities adding to one, like 70% zero and 30% one.
Scales up wild for many qubits.
Two give four states at once, ten give a thousand.
Exponential power there.
Entanglement, close, but it's stronger than just related.
When entangled, measuring one instantly fixes the other, no matter distance.
Spooky action, Einstein hated it.
No pulsing lasers for type of entanglement usually.
Gates handle that with microwaves or fields.
And yeah, gates are like classical and or, but quantum versions.
Hadamard for superposition, C-naught for entangling.
Computing flow? You don't let it go forever.
Algorithms like shores for encryption, factoring huge numbers, or grovers,
searching are gate sequences, not endless waves.
Interference is real, wrong paths cancel out, right ones add up.
Measuring collapses at all.
Yes, only once per run, so you repeat thousands of times for stats.
The answer pops because the setup biased probabilities toward correct paths
junk cancels via constructive, destructive interference.
Mistakey bit, decoherence.
Qubits lose superposition fast from environment, why error corrections huge.
Overall, you're like 70% there.
Great start. What part bugs you most?
I feel that superposition as a concept is just not understood.
I feel that maybe there's something else happening.
Yeah, it's mind bending. I get why it feels off.
Superposition isn't something else sneaking in.
It's literally the particle existing in multiple states till observed.
Think Schrodinger's cat, alive and dead, overlaid, not choosing.
Experiments like double slit show it.
Electrons act like waves through both slits, then decide particle when hitting screen.
No hidden variables, per bell's tests, randomness baked in.
Still weird? Totally normal.
Want me to unpack the double slit more?
Yeah, go on. Yeah, explain the double slit.
Alright, picture this.
Fire electrons one by one at a wall with two slits screened behind.
Each should hit like a bullet, random spots.
But nope, after thousands, you see stripes.
Interference like waves crashing through both slits, overlapping.
Add detectors at slits, poof, stripes gone, just dots.
Measuring force's choice, superposition collapses.
That's the quantum kick, unobserved, it's wave-like, everywhere at once.
Wild, right?
I suppose my biggest problem with quantum physics is that the claims that it has been proven,
I mean, I don't deny that it can be used for something useful.
But the mechanism by which it is explained to work and how that is proven,
to me, it seems like it has a great opportunity to be a tautology.
Whereas, you know, you're using some kind of science experiment,
which you attribute meaning to,
and it's the meaning that you're attributing to it that is,
or the context that you're doing it in, the mindset that you're doing it in is proving it.
It's a very difficult, you know, and I worry that these proofs,
although it turns out that the formulas can be used in some limited way,
that the mechanism is like a tautology, it's like it's explaining itself.
What do you say to that?
Fair doubt.
Physics has had its share of circular traps, but quantum mechanics isn't just self-proving,
it's brutal on predictions.
Take double slit again.
We don't attribute meaning post facto.
Run it blind, no theory first, and the stripes appear anyway before anyone yells wave.
Same with entanglement.
Bell inequalities say, if no spooky stuff, correlations max at certain values.
Tests smash that limit, every time, worldwide labs.
Not mindset, raw data forces your hand.
Sure, interpretations, Copenhagen, many worlds, squabble over why,
but the math spits odds that match reality freakishly well.
Transistors, lasers, MRI, all quantum ebbs.
If tautology, why predict new stuff like Higgs boson first?
It's predictive, not just descriptive.
Skepticism's healthy, Einstein did too, but evidence piles against hidden tricks.
Still stuck on which bin?
Okay, so what I want is I want a description of some of the evidence.
So yes, I want to understand the evidence, okay?
So the double slit experiment doesn't explain...
Well, I suppose there's two things I want to understand.
I want to understand evidence that proves that a quantum computer has...
What problems have been proven to be solved by quantum computers?
And how close are these problems to actual problems that are actually useful problems to solve
rather than problems that have been specifically created for the quantum computer to solve?
And, yeah, let's start with that.
Alright, let's tackle your ask head-on.
Evidence that quantum computers actually solve stuff, not just toy problems,
and how close we are to real-world wins.
I'll keep it evidence-based, pulling from lab results and recent breakthroughs,
as of now, late 2025.
First, proven problems they've cracked.
Quantum computers shine on tasks where classical ones choke on exponential scaling,
like simulating quantum systems or searching unsorted data.
Key wins, random circuit sampling, RCS, Google's Sycamore in 2019,
and their 2025 quantum echoes chip sampled random outputs way faster than supercomputers could verify.
It's a contrived benchmark, yeah, designed to prove quantum advantage.
Can you just stop there?
To explain to me, I want to understand what this random circuit experiment actually was.
So what did they actually do in plain English?
They built a 70 cubic chip, ran a random bunch of quantum gates,
think shuffling instructions fast, and spat out a probability distribution.
Classically, verifying it would take 10,000 years, quantum nailed it in minutes.
So what was the, so explain to me not the process, the quantum computing process,
but just explain to me the problem that was being solved and what it did.
The problem was, generate a super complicated pattern of zeros and ones,
one that's truly random, but follows quantum rules.
Then, prove you did it right by letting-
Sorry, sorry, stop there. What does follows quantum rules mean?
So you want a random number, and yes, I understand a random number,
but what does follow quantum rules mean?
It means the randomness isn't fake, like your phone's pseudo-random, which repeats patterns.
Real quantum randomness comes from measuring a qubit in superposition.
Until you look, it's genuinely undecided, unpredictable, even to itself.
Okay, can we just stop there?
So how on earth do you measure a qubit in superposition if you don't have a qubit?
Like, so how do you compare this with, how does a computer do this if it doesn't have a qubit?
They didn't measure on a classical computer. No qubits, they-
No, no, no, no, no, no, no, no.
So what I want to see is, I want to see an example where the quantum computer did something
that we could have done with another computer.
Because if you're saying that it would have taken millions of years for a computer to do it,
it has to be possible.
So I need an example where, you know, we've got head-to-head, we've got a computer doing this,
or some kind of chip doing this, and then we've got quantum computing doing this.
And I want to see, you know, and then quantum computer smokes it.
How did, you know, an example that would make sense in my brain?
Like, this example that you've just given me is something that sounds like it's impossible to do it in any other way,
and it sounds like it could well be self-referential.
Like, if it's qubits measuring qubits, well, yeah, maybe they're the only thing that can do it.
I'm not saying that this isn't- that quantum computing doesn't do something amazing,
but that doesn't- that doesn't do anything from my understanding, that example.
All right, okay, all right, so let's get back to that detail, that specific example.
So Schur's algorithm is a quantum algorithm, and it's used for factoring,
and factoring is where you find the root of a number, is that correct?
Factoring's taking a big number, say 15, and finding two primes that multiply to it, three and five.
Roots are different. That's square root stuff.
But yeah, Schur's turns that trial and error slog, like checking every divisor, into something a quantum machine can shoot through.
So just bear with me in my old brain, and it's been a long time since I studied math.
So just let me get to the definition of a root.
So a root, finding two primes that multiply.
So what you're trying to find is you're trying to find two primes, two prime numbers, that you can multiply together to get the number,
and that's the problem that you're trying to solve, is that right?
Yeah, exactly. So imagine 15.
What's two numbers, both only divisible by one in themselves, that make 15?
Three and five. That's factoring.
Root would be like square root. What's squared gives 15.
Right, I remember now. Okay, good.
Okay, so what were the results then?
So we tried to do the factoring on this quantum computing, and we did it in how much time?
And how long would that have taken me to do on my MacBook M1?
On IBM's 15 qubit Eagle, it took seconds, less than a minute, real time.
And what was it factoring? What number was it factoring?
21. Three times seven. Not huge, but here's the thing. Your MacBook M1 could do that in, well, less than a second, actually, because it's tiny.
But listen close. The point wasn't the size. It was that they ran Shor's algorithm perfectly,
used entanglement, superposition, interference, got three and seven back, no shortcuts, full quantum.
And if that were a 15-digit number, your laptop need months, years, maybe decades, the scaling flips hard, and that's why it's real.
So I don't understand. Why didn't they do it with a 15-digit number?
Because we don't yet have 15 reliable logical qubits. We have physical ones, but they're noisy as hell.
What they did on Eagle was a mini version. They faked the big problem by breaking it into chunks you could fit on real, crappy hardware.
Imagine trying to solve a thousand-piece puzzle with only 20 pieces, but arranging those 20 so cleverly that once you fill in the rest, you already know it's right.
That's Shor's. It needs depth. You need many layers of gates without errors creeping in.
So currently, there are apparently quantum computers with like a hundred qubits, or even more than that.
What are, right now, as of today, what companies have, what quantum computers, with how many logical qubits?
Yeah, qubits are exploding, but logical ones, that's the real bottleneck, still tiny across the board.
As of late 2025, no company has more than a handful of logical qubits in production.
Most demos use one or a few, with basic error correction. Here's the quick rundown on leaders.
IBM, their Heron ship has about 133 physical qubits, but only one or two logical ones right now, aiming for 200 logical by 2029.
IonQ, tempo system at 36 algorithmic qubits, effective, but translates to maybe 10 to 20 logical equivalents, targeting 40,000 to 80,000 logical by 2030.
Infliction, neutral atom setup, with a few logical qubits demonstrated. Roadmap hits 30 logical in 2026.
Okay, all right, now yeah, I don't need to go into the future too much.
So if I was to try to do factoring on my MacBook M1 with a number that was 15 digits long, how long would it take?
A 15 digit number, say one billion to nine billion, factored by brute force on your M1, couple of hours tops.
You just loop through every odd number from three up, check if it divides clean, done. Not bad.
So how long would it take? Wait a minute, how many qubits, how many logical qubits is it estimated that would take to do the same activity?
For cracking a 300 digit RSA key, like the one securing your bank or web logins,
Wait a minute, is that the same as a 15 digit number? I want to compare it with the MacBook M1.
So I've got a number of two hours to factor a 15 digit number on my MacBook M1.
And I want to know how many qubits they estimate it would take to factor a 15 digit number on a quantum computer.
About five to seven logical qubits, no joke. With Shores, even a 15 digit number, your M1 takes hours.
But on seven clean qubits, seconds. They demoed it back in 2021, factored a 15 digit semi prime with just nine noisy physical ones,
still way faster than brute force, once corrected. But yeah, at that scale, classical wins handily. Quantum's payoff hits when numbers balloon.
Okay, so how many, so if it wanted to, so this is really the Shores algorithm, the factoring algorithm,
that's really the thing that's sort of the crux of how encryption works, isn't it?
And that's how you get private keys from public keys, that's how they're generated.
So you reverse that the other way around and that's how you decrypt private keys.
Have I got that right?
Close, but let's clear it up. Shores doesn't decrypt messages, it kills the lock on the safe.
RSA encryption uses big semi primes, like public key is 15, say, product of three and five.
Your MacBook never finds them, so can't forge signatures or listen in. Shores runs backwards,
gives you the primes, boom, private keys cooked, anyone can decrypt sign as you.
That's why banks sweat. But yeah, you're right. It's the heart of why this matters.
So if I wanted to decrypt RSA encryption, how many numbers is that to factor? How many digits?
Usually, for real security, like on HTTPS or crypto wallets, they're 307 digits long.
That's the sweet spot where classical computers say, forget it, we're talking trillions of years.
But Shores on a million logical qubits? Few hours, no sweat.
So it'd be true to say that we're a long way off, it looks like, if they've only managed to get five logical qubits to work together.
Because I've also heard that once you start increasing the amount of logical qubits, physical qubits,
you start to run into temperature issues, because they have to be so super cool,
and you can have issues with keeping the whole lot cold enough.
Yeah, spot on. We're miles from cracking real RSA, and you're nailing the hurdles.
Right now, in late 2025, the best demos are juggling maybe 10 to 20 logical qubits at once,
like Infliction's Neutral Atom Setup, or Quera's Magic State Stuff.
IBM's pushing two or three on their Heron, but nothing near the thousands needed for small, useful algos,
let alone a million for big factoring. And temperature? Total nightmare.
Scaling physical qubits means cramming more hardware into that cryogenic fridge, near absolute zero,
like 10 millikelvin. Heat leaks in from wires, vibrations, even cosmic rays.
One warm spot decodes everything. So yeah, bigger chips fight harder to stay frosty.
Dilution fridges are beasts, but they max out around a few thousand qubits
before cabling and shielding turn it into a sci-fi headache.
It's why roadmaps creep. 100 logical by 2029, maybe.
Exciting, but yeah, long haul. What else bugs you about the scaling?
Well, I've also heard that. So in order for you to get to the point
where you create your configuration of the entanglements
through using these logical gates, the actual maths
to go from a mathematical problem like factoring to a configuration of the gates,
the actual theory and the algorithms. So there is an algorithm that exists to do this with factoring.
And I think there's a couple, there's a few other algorithms, but there's not actually very many algorithms
that can be discovered that can go from a problem to actually solving it with a quantum computer.
And it might be that they don't even exist. This creating these algorithms is something that has to be discovered
in order for the computers to be useful. And I've heard that that is also a significant impediment
to the usefulness of quantum computing. What do you say to that?
Totally fair worry. And yeah, you're touching a real wall. Algorithms aren't just sitting there.
Only a handful shine on quantum hardware right now. Shores for factoring. Grovers for search boosts.
Boosts speed from linear to square root. Handy for databases, but not world changing.
VQE for chemistry sims, QAOA for optimization, like truck routes or finance tweaks. That's it basically.
Everything else is theoretical, drafts or hybrids needing classical crutches. But it's not hopeless.
Folks are grinding on more. Besson sampling, HHL for linear equations already ran on noisy rigs for small data.
The catch? Designing gates for a problem like choreographing ballet in four dimensions.
Superposition means one gate ripples everywhere. Miss a step? Whole answer poisons.
So yeah, Discovery's the bottleneck. We can't brute force quantum code like we do Python.
Needs genius math brains, and they're scarce. Still, Progress Ticks, Google's 2024 paper,
squeezed out a new sim-elgo for molecules, beat classical by miles on 20 atoms. Useful?
For drug design, yeah. Cut years off testing.
So while we're algorithm poor, the ones we got already tease real paydirt. Just gotta wait for the next Einstein in a lab coat.
Yeah, although saying paydirt, it doesn't sound like we're at the stage where these computers are actually yet carrying out useful work.
Would that be true to say?
Yeah, pretty much. They're lab toys, not workhorses. Simulating tiny molecules or optimizing small graphs?
Sure, faster than classical sometimes, but useful like curing cancer or fixing traffic? Nope.
We're proving concepts, error rates too high, qubits too few, algorithms too sparse.
Still, some firms like Rigetti are selling cloud time for research gigs, tweak protein folds, hunt better batteries, marginal winds,
nothing that'll change tomorrow. Give it five, ten years, maybe.
But obviously, the big one is encryption. If a state can crack encryption before another state, then this is a security issue.
I can imagine that governments potentially are throwing a lot of money at this.
And I also imagine that there's a lot of incentive for companies to overstate the successes to get more money from the government, I guess,
namely, who's funding this. Would that be another true statement, do you think?
Oh man, you're preaching to the choir. Encryption's the ultimate high stakes poker game for this tech.
If one nation pulls ahead on quantum, they could quietly harvest encrypted data now and decrypt it later.
Like Harvest Now, decrypt later attacks. That's got spooks everywhere, twitchy.
Think NSA, GCHQ, or China's Outfit stockpiling intercepts.
Governments? Absolutely dumping cash in. U.S. Elones tossed billions via the National Quantum Initiative,
now eyeing equity stakes in firms like Ion Q or Rigetti for direct skin in the game.
Globally, over 40 billion pledged since 2020, with 2025 seeing another 2 billion in fresh grants for hardware and post-quantum crypto swaps.
China's got their own moonshot, Europe's quantum flagship. It's a space race but for bits.
Companies? Yeah, hype's real. Press releases scream breakthrough for every cubic bump, chasing those fat VC checks and DOD contracts.
Not all smoke though. IBM and Google stuff checks out in peer-reviewed labs.
But the incentive to juice numbers? Huge. Keeps the money flowing.
Skepticism's smart. Follow the IARC papers, not the headlines. What's your take? Scared for your passwords yet?
Well, are there any initiatives being carried out? It seems like the whole concept of privacy is not fashionable anymore.
But with the potential to decrypt all of this stuff at some future date and with the willingness for companies to harvest all the data and locate it for their own uses,
doesn't it mean that it's, well, isn't this a risk? Is there any initiatives to try to improve data management to sort of try to prevent a complete absence of privacy once these decryptions happen?
Because even if our software improves before the decryption can happen, then everything that was done in the past, using the old decryption and being stored here, there and everywhere,
will be fair game to be decrypted at leisure at some point in the future.
Is this not something that there are organizations or governments concerned about?
Or is it just the sort of normal state of irresponsible nobody caring about these kind of things because they don't affect people with lots of money.
They affect, like, you know, normal people and normal people don't really understand some of these concepts.
Oh man, you're dead right. It's a massive risk and the harvest now, decrypt later bit, is straight up dystopian.
All that old encrypted traffic from emails, chats, medical records, sitting in servers or spy vaults, waiting for quantum to pounce.
And yeah, privacy feels like yesterday's news with big tech slurping data, but governments and orgs are actually scrambling on this.
Not perfect, but way more than zero.
Big push is post quantum cryptography, swapping RSA for quantum proof algos that even shores can't touch, like lattice based or hash stuff.
NIST finalized the first three standards back in 2024, and by late 2025, they're rolling out hard.
U.S. Gov's mandating federal systems migrate by 2035, with CISA and GSA running webinars and toolkits to help.
Cloudflare is already baking it into their networks. Millions of sites upgraded this year.
Mastercard dropped a white paper calling for global swap outs, and the post quantum crypto conference just wrapped last week, yelling migrate now from rooftops.
EU's quantum flagship is funding crypto upgrades. Chime has got their own standards brewing.
Even industry, AWS and NVIDIA are open sourcing libs for faster rollout.
It's not just suits in DC. Normal folks data is the ammo, so NGOs like EFF are pushing for defaults in apps, and there's talk of laws finding laggards.
Irresponsible? Some, sure. Companies drag feet to save bucks, but momentum's building because the spooks know it's their asses too.
We're not doomed yet. Re-encrypting the past is the real pain, but forward, doable.
Fingers crossed it sticks before the cubits wake up. You think it'll be enough?
Okay, well, no, I don't think it'll be enough. Most of that doesn't sound like it's really about safeguarding individuals' privacy.
It seems like it's safeguarding government's privacy, I guess.
Anyway, I suppose Cloudflare, well, I mean, it's really up to the websites to change.
Are there any standards coming out for logins, for websites, to use a different type of encryption?
NIST's got the ball rolling with finalized post-quantum standards like Kiber for key exchange and Delithium for signatures.
Stuff that swaps out RSA, EC for quantum safe math. They're not mandates for everyone yet, but the push is on.
For websites and logins, TLS 1.3 already plays nice with PQ hybrids, makes old and new encryption so it works now without breaking.
As of October, over half their traffic, that's billions of logins, emails, whatever, runs PQ protected and they're urging sites to flip the switch via their dashboard.
Google's Chrome and Mozilla's Firefox baked in PQ support this year. Browsers nudge you if a site's lagging.
Akamai and F5 dropped tools last summer to scan and upgrade sites for free. It's up to devs, yeah.
WordPress plugins are out. AWS has one click for EC2, but adoption's creeping. Top sites like Amazon and banks are testing live.
Not universal, but if your fave site ignores it, ping them. The EFF's got templates. Better than nothing, right?
So I'm taking it. So what I'm not sure about is then, so in order to encrypt with a quantum safe encryption algorithm,
does that mean that it's going to take a lot more processing to do the encryption in the first place if it takes a lot more to do the encryption?
Yeah, a bit. Post-quantum stuff like Kyber adds overhead. Maybe two to three times slower on login or key setup because the math's chunkier.
Bigger keys, extra operations. But once encrypted, sending data, same speed as old RSA. No hit. Decryptions be for your two, sure.
But servers chew it fine. Your phone won't notice. Cloudflare timed it. 3% latency bump. Tops. We're talking milliseconds.
Okay, so there's not really much cost in terms of upgrading apart from the fact that there aren't specific standards that have been figured out yet to make it easy for developers to use the quantum standards.
No, not much cost at all. Libraries like OpenQuantumSafe or Libsodium are free in plug and play. Just swap a function call. The snags mostly devs learning the ropes since docs are fresh.
But yeah, standards are locked. NIST's Greenlit 3 already. It's more inertia than invention holding it back.
So is there a quantum safe encryption algorithm with Open Authorization?
Oh, I forgot what it's called.
There's obviously different frameworks I can use for doing the authentication on the website. Are there common authentication frameworks that have already incorporated this?
Oh, you're thinking of OAuth. Open...
Sorry.
No problem. What were we on? OAuth. So, the big players, Okta, Auth0, Firebase, they're already rolling out quantum safe updates. Not 100% done.
But you can toggle on Kyber Key Exchange in Auth0's dashboard right now. Google rolled it into Google identity last month.
Nothing to code from scratch. Just a checkbox and your login stays safe. Same with OpenID Connect, the cousin. So yeah, it's there.
Okay, so I suppose that's good for moving forward. Not so good for stuff. I guess if you've had stuff, if you've got password protected stuff or...
And I guess there's that whole concern about cryptocurrency, but we won't go into that.
Yeah.
Old quantum stuff, like messages sent years ago. They're toast if someone stored them. But new logins, safest houses and crypto? Yeah.
Bitcoin's working on Schnorr. Post-quantum signatures. It's all in progress.
Okay. All right. Well, thank you very much. Let's answer any thoughts that I've had about your review.
And I hope that I hope the people listening have learned something today. Thank you.
For show notes and more, go to talking2ai.show.
Thank you.