[Music] every time humanity gets a hold of new laws of physics we're able to radically change what's possible we're starting to take advantage of some of the really really strange or unintuitive kinds of powers that exist in quantum mechanics things like entanglement or
superposition but in the new round in Quantum Computing we're trying to treat these quantum mechanical effects not as bugs but as features I am troubled sometimes by the knee-jerk view of many they kind of freeze knowledge as of now and then make all of these to me
outrageous claims what they're ignoring here of course is the new knowledge that will be generated and will be acted upon I think the real National Security invocation is not having the economic benefits of this technology happen we need as many smart people as
we can get working on it and we need to give them on-ramps into the technology that are not due a five six year PhD and with linear algebra in Python and some of these open source toolkits you can actually get pretty close to The Cutting Edge in some real engineering problems
you know it's not that quantum computers are saying here's all these limits to the computers you could imagine building it's actually saying I can do things that you don't understand what it's Computing like that's your challenge foreign Ty with yet another infinite Loops today
I am delighted to have the very first winner of the O'Shaughnessy Fellowship which we started this year Dr William Zhang who is the founder of the unitary fund a non-profit dedicated to an open source Quantum Computing ecosystem for the benefit of all people he has a
degree in physics from Yale is a Rhodes scholar as a PhD in Quantum algorithms from Oxford oh inventor of the quill Quantum instructional language has been picked by corpse in the science category in the 30 under 30. William I gotta tell you the first thing I thought of when I was
getting ready for this podcast was Kennedy's remark about all of the brilliant people he had in the white house at one given time poets scientists Etc and he said I don't think that there's been this kind of concentration of intellectual Firepower in the white
house since Jefferson dined alone welcome William well thank you Jim that's very complimentary but I I will say one of the things that's super exciting about working in Quantum Technologies and Quantum information is it's such a magnet for a whole bunch of smart people
I mean I I get to find myself in rooms where I'm the dumbest and it's just this is such a hard problem that it attracts so many interesting and really multi-disciplinary people uh anyway we'll talk we'll talk more about that but you know thank you so much for
having me on here and I'm super excited about the fellowship as well of course it delighted to have you on and as you know from our conversations I I have been uh interested in quantum mechanics and then Quantum Computing for a long long time we had some great discussions
uh about a particular problem that we're going to talk about a little bit later in the podcast but first can you can you help us out and and sort of explain in intelligent Layman's language what's going on with Quantum Computing um why is it such a massive potential
Innovation you have a great quote which is every time humanity gets a hold of new laws of physics we're able to radically change what's possible um and so maybe let's use that as a jumping off point to explain in in language that sort of key intelligent
Layman could could grok sure um I mean you could think of our understanding of physics as the operating system that underlies whatever technology we're building at the time so understanding thermodynamics in many ways is a part of what drives the Industrial Revolution understanding
electromagnetism leads to electronics and then information technology and we've known about you know quantum physics for almost 100 years actually I think um there's going to be an international year of quantum in 2025 as sort of an unofficial 100 years
but the starting point wasn't just one moment something we've known about Quantum Physics for a long time and in many ways Technologies we use today like lasers or transistors and regular chips they wouldn't work if we didn't understand quantum mechanics but the new thing that's happening is
we're starting to take advantage of some of the really really strange or unintuitive or different kinds of powers that exist in quantum mechanics things like entanglement or superposition and we can talk more about these subtleties and use them to basically amplify them into features
in technology that we build so every computer we work on today works with rules of physics it's the rules of electromagnetism and and when the chips get very very small they're actually trying to avoid quantum mechanical features but in the new round in Quantum
Computing we're trying to treat these quantum mechanical effects not as bugs but as features so so tell me a little bit uh let's let's go a little deeper there uh on on you know what's a qubit what what what is different from a Quantum Computing uh perspective to the old you know uh
Seymour Craig super computers I know they're very different but if you could explain uh the the differences so the one what's really fun about it is you really think about the basic model of computing a little bit differently so you know instead of having a bit you
have this thing called a Quantum bit that behaves somewhat differently um and from a mathematical perspective um you can think of a Quantum bit as a generalized version of a of a weighted coin um so there's we're going to move in three steps so you can start with a
normal bit it's just either zero or one and then you can think of a probabilistic bit so that's something like a coin which if it's an even coin then it's 50 50 zero fifty percent zero fifty percent one but you can change that could be 75 percent one point five percent zero
and you can model that as I guess we're on a podcast so I'm using my hands so I don't know we'll see it but it has a line no that's it that's actually okay because we also released this on YouTube so the people who watch it there will get the extra get the extra yep so um so
probabilistic bit so I don't know a bit is just like two two points and a probabilistic bit you can think of as existing on a line between those two points where zeros at the bottom and one's at the top and an even coin is halfway and then one that's weighted towards one
is maybe a little bit higher up 75 percent of the way towards one and a Quantum bit is uh you represent Quantum States using a generalized generalization of probability Theory these things called amplitudes but you don't model that on a line between zero and one you model it
on a sphere where the zero is the North Pole and the South Pole is the one or the other way around it physicist and computer scientists flip this which makes it a little bit annoying but in either case the polls are the zeros and ones and then the state of the quantum
bit is somewhere on the surface of that sphere instead of being somewhere on the line and one of the ways you can start to think about how this gives you additional power is if you thought about the the interval line where a Fair coin is just at the midpoint
for Quantum bits there's not just one midpoint there's actually the whole equator around the middle that all act when we measure them like their Fair coins but they're actually different states and so there's extra states that you can take advantage of even in a single qubit
and then mathematically this gets more complicated as as we as as we scale them up and so we can talk more about where that power comes from but fundamentally on a Computing you're thinking about a different Computing model then what's usually thought about within
what we call classical Computing and there's these two important things about that one is um there's actually some some of the original work on these things called Quantum Turing machines showed that everything that's computable by a normal computer can be computed by a quantum computer
and vice versa but that's only given an infinite amount of space and time resources or at least skills exponentially so a very very large number um but if you look at more practical kinds of questions where we restrict the space and time resources then there are
algorithms that are available quantum computers that can run much much faster even exponentially faster so it's taking something that would be you know billions of years on a very good classical computer to something that's trackable on the quantum computer and
that's because the fundamental models really different and and give us a sense of the scale of that difference you know the uh you know the old joke I don't know if you're a Douglas Adams fan uh but I am and uh he's got the computer that they ask what
is the meaning of light and it just keeps Computing Computing Computing Computing and because it does because time and space right and and then It ultimately gives the answer reporting too um you talk about the the the difference between what might take the fastest conventional computer
um uh the amount of time to solve a very intractable problem versus a quantum computer well one way to look at the comparison is to look at what it would take to use a regular computer to simulate what the quantum computer is doing and let's for now let's think about
simulated exactly and this is where that those kind of fundamental scaling difference between Quantum and classical starts to to rear its head so if you have n Quantum bits then it's going to take you 2 to the N classical bits to exactly simulate what
those Quantum bits are doing and so that scales quickly so something like 10 Quantum bits you could do on an old compact from 20 30 years ago um something that's more like 30 or 35 starts to take a pretty beefy aw something you could provision on one of
the cloud providers once you start to get to 60 or 80 or 100 if you're trying to exactly simulate it then the state of the quantum computer requires it says 100 Quantum bits requires two to the 100 complex numbers to write it down um and that's significantly bigger than
uh you know we're starting to get to astronomical Universe size scales in terms of numbers um so that's where the sort of fundamental one of the fundamental interests in the power of quantum Computing comes from now it's more subtle than that because in in this power example we're
usually not trying to exactly stimulate what a quantum computer is doing where we're just trying to solve some problem so sometimes we can just approximately simulate the quantum computer and still do it much more quickly on a classical computer even though we have 100 bits
um and so one of the interesting races that's going on right now as some of the first quantum computers are coming online that are in this regime uh you know 60 100 a couple hundred Quantum bits is to Benchmark them and to show that they can do things that classical
computers can't do even while we keep improving the classical algorithms so there's been sort of this back and forth that's been going on for for a couple years so so uh what are some of the things uh speculatively that the quantum computers that can that can accomplish that a
conventional classical computer can't yes I think that we're going through the four categories um the first is one of the original actually impetuses for building quantum computers which is to simulate other quantum physics systems um so this scaling of what it takes for
a regular computer to simulate a quantum computer doesn't just apply to simulating quantum computers it applies to simulating other Quantic quantum physics systems so molecules catalysts materials superconductors interesting states of matter we know how to write
down the theoretical in some cases in many cases actually we know how to write down the the rules for how these things should behave but it's really hard to do the numerics because of the scaling and quantum computers could at least you can take intuitively
um are pretty good at this and there are some there are some specific algorithms to do it in detail but but at a high level you're using a quantum computer as a quantum physics system to simulate some other one so the first one is in Quantum simulation and that might sound Niche
but it's it's not we can we can come we can come back to it but there's other ones in Monte Carlo simulations so this is important for risk and pricing in in finance or risk modeling for all sorts of different applications there's machine learning applications it says
optimization applications and in each of these cases there are there's work that's ongoing and has been done on proving separations between um as well as we can because it's hard work as well as we can between classical algorithms and Quantum algorithms um so I I can dive into we can we can
pick whichever direction of those you want to dive into next but there there's sort of a pretty big broad class yeah so the two that I'm most interested in and I would uh guess that my audience is as well would be first the ability to simulate Quantum systems themselves and
what we could learn from that uh in terms of our ability to peer into that that we don't currently have but then also the use case of uh risk simulating risk um and and what a Quantum Computing world might be able to do in transforming the idea of risk the idea
of investment Etc so why don't we cover those first those two verse yeah so the first one you know simulating quantum physics systems might sound like a very uh esoteric or even academic domain but it can have pretty big impact um so to give a couple examples
um so one that people in the field talk about a lot um is in catalysis so a lot of the big industrial chemical processes that we run as a species depend on some Catalyst to activate the reaction and so one really big one is called nitrogen fixation which is what we use to make fertilizer
and the Catalyst we monocallus we use for nitrogen fixation is uh was invented it's the Haber Bosch process it's from about 110 years ago the Catalyst operates at 400 degrees Celsius and 200 atmospheres of pressure so it's a pretty energy intensive
reaction and what it does it takes nitrogen out of the atmosphere puts it into ammonia and then we use it to make fertilize it and we do this so much that something like one to two percent of the world's annual energy budget goes into this whatever you do better and when we think
we can do better and one of the reasons we think we can do better is there's natural systems as bacteria that do this kind of reaction but obviously not at that kind of temperature and pressure and we want to understand how and in order to understand how we need
to simulate the quantum physics of the of the systems and there's other kinds of examples one of them is in trying to understand why photosynthesis is so efficient in in plants there's a there's parts of the electron transport chain that look like they might involve quantum mechanics to
really work as efficiently as they do and it's hard for us to use computers to simulate and study this you know for designing an airplane today we don't do it so much on wind tunnels we can mostly do it on computers because we know the law the physics and
the computer algorithms are good for materials which have quantum mechanical Behavior like trying to make high temperature superconductor um we know the rules but our computers get stuck and so if we can build quantum computers that are good enough to simulate these then we can
start to unlock I mean not just the ones I indicated but probably lots of exotic kinds of materials that we've never thought of but again one of the reasons why I'm so absolutely fascinated by this is the just that alone uh could lead to remarkable uh Innovation and uh our ability to uh
conquer things that have been really uh bedebling Us in the past uh and you know sort of the the way all of that reaches in as you're pointing out with uh the the energy usage if we could drop that considerably that as profound effects on you know every uh everything really the
world economy and all the way down uh talk a little bit now about the more pedestrian I guess some would think uh but also very important uh element of how it might be able to transform the idea of risk or analysis of risk uh why why is it uh potentially better than
what we have with a classical computer yeah so this is I've been looking at using Quantum compute so actually one thing's probably worth saying is quantum computers are not better at everything at the outset and we're talking about specific examples because there are
certain things that we know quantum computers is going to be better at sometimes a lot better but they're not just better at any generic problem like word processing and it's not unclear to me then that'll be better um but in calculator so in actually until recently
um I had been leading a uh an advanced Computing and Quantum Computing research group at um at Goldman Sachs and we've been doing a lot of work looking at uh risk and pricing applications of quantum computing and one of the reasons we looked there is there is a uh there's a type of
algorithm for Monte Carlo simulations that has a proven speed up for quantum computers over classical and this is what we call quadratic speed up and so typically Monte Carlo converges the number of samples you need to take to get to some error Epsilon
converges as one over Epsilon squared so if you want to get to an error of one in a thousand you need to take a million samples dropping some constant back but the quantum algorithm scales is one over Epsilon so if you want to get an error in one of a thousand it's a thousand samples
and if you want your error to be you know one in a million then instead of being a million squares sample is only million so it's a million actually speak up then now that's just in theory and the constant factors really matter and you need to worry about what do I mean
by a sample if the quantum computers samples are much slower than each class of computer sample then the speed up doesn't doesn't uh you know gets washed out um but the the fundamental Improvement is there and we've done a bunch of work in estimating you know what specs of a
quantum computer you need in order to make that to make that practical and then in terms of you know the business applications it's that weakened with the fixed compute budget just know our models more accurately in the same amount of time and and that kind of leads me to the uh
you know the speculative World of Science Fiction Etc where everyone is making uh elaborate claims about you know whoever comes up with the first quantum computer will rule the world because of all of the things that it can do uh what's the reality what what's the
reality uh when when uh serviceable quantum computers uh dependably come online what what does that world look like I mean I think our best analogy is to look at what happened when the first round of computers came online uh there so a few synthetic takeaways one is
it's not about building the quantum computer and then you have the quantum computer and you run things on it and it's it's and it gives you all these advantages um there'll be a bunch of different product lines from a bunch of different places with increasingly better quantum
computers some of which will specialize for different applications but they're only going to be barely good when they first start getting applied um they're they're not even barely good now I'd say that the quantum computers we have now are are basically good
enough to say they're quantum computers um but not too much else um so I think there's going to be an on-ramp and you know economically what will happen is the folks that figure how to apply them first will reap a lot of benefits from that but eventually like other Technologies
um it'll start to spread and and then we'll we'll all benefit from um from the general Improvement now there is one asterisk here which is maybe worth mentioning which is there's a particular application of quantum Computing that gets talked about a lot
which is in code breaking um and you know this is there's an algorithm for um prime factorization that's from 94 1994 and that spurred a lot of initial interest um and there that there could be even more first mover advantages the concern is you know we have a long way to go you
need you need a pretty good quantum computer in order to run um an algorithm to to break a a key in 2048-bit RSA or something like that um it's over a while out but the concern is that people are today storing encrypted information and then waiting until the quantum
computer they have a quantum computer and I don't know what it'll be you know 10 15 20 years you can you can take your pick um but then once they're they're able to once you have good enough quantity you can just look through that encrypted data and and everything you know relatively
speaking uh looks looks transparent from the past and for someone if anybody wants to keep something secret for I think a lot of US Government entities for example is 30 years they aim for and then that's a concern that they have now so that's a little bit more it's the
first mover uh advantage yeah and and like if we if we go down that path a little bit that that could lead to like a variety of ongoing crises right or or am I overestimating the importance of our ability to keep certain information encrypted yeah it could be a
really big problem um but I I this is from the podcast and I'm just speaking for me so I get to really say my view which is um I think it'll get um I think it'll get fixed um like I I think when we look back 100 years from now um at the invention of quantum Computing
and the arrival of the quantum technology industry we'll see a whole bunch of enormous economic benefits and this really big Y2K bug upgrade that we all had to do for crypto that was painful and maybe some people made some mistakes and and didn't do it but but like I think the
100 Year view is that we'll go through that and one of the reasons I I have that view is that um there are alternatives there are approaches that nist has been working on a a standard for a while now and they're making some recommendations for what we call Post quantum
Quantum secure cryptographic schemes that run on regular computers you know one way to fix this is actually to use quantum networks and Quantum Technologies which we can talk about too but even if you stick with regular Hardware there's reasons to be optimistic
the execution you know getting these things deployed it's like upgrading from Windows 95 which I think some people still run um you know there's risk in that kind of thing but net net on net um I think some people make a lot of money fix providing those Solutions and
there's reasons to be worried about if you really want something from now to be secure for 50 years you should be a little worried but that that's going to get washed out and um in in sort of the vast advantages that you get from using Quantum Computing and other kinds of
applications let's make it yeah yeah that's my view as well um and uh I I in fact that is my view exactly that uh we will come up with fixes um and they're not yet readily apparent but so are a lot of the advantages that we will enjoy when we do have uh truly
quantum computers online a variety of other uh aspects of Science of business Etc so um I I I am I I am troubled sometimes by the knee-jerk view of many uh you see it right now in the debate or around AI for example um where where they kind of freeze knowledge as of now and and and then
make all of these are into me outrageous claims about you know oh and you know we're all going to be turned into paper clips or or what what they're what they're ignoring here of course is the new knowledge that will be generated and will be acted upon right and I see
something very similar happening with encryption uh uh through the the quantum Computing era what's what's your view is that just I kind of view it as a as a bug in human OS where where we we failed to think about the idea that there's a lot of options which are not currently
on the menu because they haven't been in bed yet uh but we do have a very long history uh as you let in to our discussion Innovation uh upon Innovation upon and the more we learn right the better the tools that we make and and that's an ongoing process and am I am I
too Cavalier in that attitude I I it's hard to say about it in general um but there's a concrete downsides are very people overweight them and you can see a concrete downside of oh I'm gonna someone's gonna hack into my bank account um upsides that might be very big like
the advances that come from custom designed Quantum meta materials um that you know it's an ups and upside so we're weighted against it because we like downsides more and it's also a little it can seem a little Vigor and so people have Sometimes some trouble
balancing those but you know in in our space I think we're in a pretty good spot um you know the cyber security Community has been worried about this for 20 years um excuse me we're coming up on 30 years really uh it feels getting a little bit older um and taking it very seriously from the
beginning um and so the I think my concern now is that there might as the field grows and sort of gets a little bit more popularized outside of the experts who've been tracking it um there might be an over rotation when I think the real National Security
implication is not having the economic benefits of this technology happen yeah I I agree um that leads me to uh the old idea of open versus closed Source uh you you specifically are looking to create an open source uh ecosystem uh talk a little bit about why you find that to be
the more beneficial of the two paths yeah I well so maybe it's sort of worth saying so I started um the unitary fund about five years ago uh when I was at a Quantum Computing historical regatti and we when we when we were starting at Righetti we were building some of the
world's first quantum computers and one of the things we had to solve was a credibility issue like if people didn't really believe what what you were building it's like that a real thing and one approach to solving it was by saying well look we'll open source the
programming toolkits and we'll put the computers online and and just like let people see a little more open kimono what these things are um and when we did that uh a couple interesting things happened um one is that uh we got contributions from Usual Suspects um you know Harvard Caltech and phds
whatever um but we also got contributions from hobbyist software engineers in Eastern Europe and students in India and high school students in in Canada and and they were building I said they were finding you know not necessarily new Quantum algorithms but you got API bugs
you know and they when you make the software available for people to work with then they can learn more quickly than if you make them March through the math in a textbook you know I think learning Quantum Computing the what I'd recommend more than a textbook is to go take a look at
some of the open source platforms you know things like his kit uh which IBM's done a bunch of great work in building up they've got there's actually textbooks if you want a textbook to go with your programming Library you can have that too but these things are a lot
more Interactive um and that actually and it's becoming a long story but um I I myself I didn't have any of these tools uh and so when they started to come up and it started to become obvious that this was potentially a more accessible space than people would think
um I eventually taught a a class at Stanford um based on Quantum programming for undergraduates who hadn't taken about senior undergraduates and master students who weren't from a physics background you needed to the prereqs were linear algebra and python
and with linear algebra and Python and some of these open source toolkits you can actually get pretty close to The Cutting Edge in some real engineering problems in in Quantum computing and so those experiences have led me to I think of them as evidence of the
broader philosophy that Quantum Computing is a hard problem you know really hard and so we need as many smart people as we can get working on it and we need to give them on-ramps into the technology that are not do a five six year PhD and I as someone who did the PHD route
and and then writes PHD quality code sometimes I I also know that that's not even the optimal path for a lot of the engineering kinds of problems that that we need to solve and so we need ways for really talented engineers in control electronics and technicians and cryo and
all sorts of things we need ways for them to understand what's going on in the space um and my view is open source Technologies is one really great way to do that um so that is the first one is just kind of growing the uh the first reason to do open things is
just to get more more people involved which help Advance things quicker and the second is that this is a really hard technology and we need to show progress and progress is basically where we are a progress is not just watching Revenue go uh to the moon so to speak
um it's going to be a little more complicated than that um we need to show that we're making progress on basic technical Milestones which we're still I mean I was just at a there's a Quantum Economic Development Consortium which is an industry group in the U.S
um and I was just at their plenary meeting uh yesterday talking about standards and benchmarks and we don't agree um but what I push for is let's try and make what we are benchmarking at least be transparent about what we are benchmarking and and make that stuff
available and that way investors and end users and people who are seeing the space can see the progress we we are making even though quantum computers aren't solving all the problems that we wish they could solve yet um so so growing the workforce getting more people involved and also being
transparent about the milestonian progress for a very long League technology are two reasons why I think being open is is incredibly important yeah and uh I'm I'm very in favor of Open Source as well um and you know historically Linux runs the web for the most part uh open source
technology that many corporations uh were trying to demonize uh when it first came along are you seeing are you seeing any of that um from uh more proprietary groups are are they trying to fight against the open source or are they are they playing
nicer than maybe some of the AI uh groups that are duking it out right now yeah I mean these things go in waves um you know AI was kind of academic and open and more recently the people have started to close out so it started close got more open got he's getting more
closed in in Quantum we're pretty open um or this is because we got in early with making that kind of a norm and second is that people recognize the things that I was talking about you have a lot of Hardware vendors For Better or For Worse feeling like worried that they're building a hammer
looking for a nail and uh and they really want to make their programming toolkits as accessible as possible you know the core IP around how do they make the hardware you know that that proprietary stuff compete on that all the time actually even there you'd be
maybe surprised about some of the transparency but in many ways it's because we're like a rising tide helps all boats kind of thing right now where everyone knows we have a long way to go um and trying to grab I'm too much of I mean it's hard to grab too much of the
of the like let's say user base right now because nobody has a quality computer that's super performant right and and no but my guess is that people as it gets closer will uh attempt to grab more places on the value chain uh from from that um and hopefully the work uh of the open
source Community uh will uh continue even as uh those those uh stakeholders try to try to plant their flag on a particular part of the of the value chain um right I want to talk a little bit about uh the the uh experiment that actually we we really kind of bonded on when you joined
me over here uh and and it's called and it's called wigner's friend uh could you uh outline first for our listeners and viewers what is wigner's friend and why is it so deviously complicated so maybe just to frame a little bit or people who are listening what we're talking about
um we've talked a fair amount about the kind of economic practical applications of quantum computers [Music] um but there are also scientific applications and you can think of a quantum computer as a very well controlled very large Quantum system that you can use to study
questions that we have about auto mechanics and we still have questions um so there's uh in some ways there's this this hundred year long history of of quantum mechanics consists of someone saying well it's a lot of things but one of the ongoing themes is someone saying
you know quantum mechanics predicts this really weird thing about the world like that you know particle and wave at the same time and then people come up with some sort of thought experiment like oh there's a double slit or I'm going to put a cat in a box and or
something um and then we argue a bit philosophically and in theory about the interpretation of that thought experiment and then a couple decades later our technology gets good enough to run it and not you know do a double slit experiment where we don't just imagine
that there's an observer looking at one of the slits we actually have a device that acts as an observer and we and we see um the uh the decoherence happen so and then you know invariably what happens when we run this experiment with the better technology is it turns so far
uh in ways that are maybe disappointing it so far turns out quantum mechanics has been correct um and maybe the most striking recent example of this was the Nobel Prize in physics from this past year which was given for um confirmatory experiments of these things called uh Bell
inequalities which are um so in the in the 60s and if you want to read more about this there's a book called How the hippie physics um in the 60s what's so funny is I just yeah opened that book anyway continue please um so so in the 60s uh this guy John Bell
um sort of it took a bit of a look at uh a question about quantum mechanics um and the and the question is uh is something that had vexed a lot of people for a while but it's um it's sort of a more formal version of do do parts of the world have properties
that are fixed separate from how you look at them um and you can ask this question one of the things that I want to point out about this question is um it's not necessarily a question about quantum mechanics it's a question about any theory of um of physics it's just about just about
how reality might behave um it turns out quantum mechanics says the answer is no but you could have other theories that say the answer is yes um and so the a term for this statement like and the formal statement is you know is really described by local hidden
variable theories so any any kind of physical theory that has local hidden variables um plus some other assumptions um you can call this a question about um experimental metaphysics so it metaphysics it's not one physical Theory it's about a bunch of them um and it's experimental because uh that
question can be turned into a series of experiments that you run and you do some statistics on them and you see whether or not an inequality is violated and it will tell you the answer and if you use the the predictions of quantum mechanics then you predict that
you'll get a violation of the inequality that you that this bound that you should get if your world is described by local hidden variables anyway this experimental metaphysics term I learned um this is a theoretical physicist named Eric couplecanti um who's written a bunch about this
season he's in Australia so long and short of it is uh these the predictions that Quant mechanics and reality violate local hidden variable theories is made in the 60s people started to run experiments um initially when they were running experiments nobody cared
because they some people said these experimental medicine physics questions are not physics but lo and behold in you know a few decades later and in 2015 some really really serious um large-scale versions of these experiments we call them blue pill free
closing a lot of the you know basically dotting the eyes and crossing the t's were on um and uh and then a couple books on the on the Nobel Prize for for those experiments last year so to come back to your original question about wigner's friend um William's friend is another one of
these and as we get better at building Quantum Technologies we should go back and revisit which of these strange theoretical implications that we see in quantum mechanics we might now be able to make experimentally accessible so weakness friend is one of these also
from the 60s and from Eugene wigner and the the very very basic way it was originally formulated is um uh is about what happens when we put quantum mechanical systems in superposition so in in quantum physics um like a Quantum bit that I was talking about it's not either zero or one it can
be so zero in one of the poles of that sphere that I was talking about it's can be I to zero one separately but it can be anywhere on on the surface of the sphere and those are these mixtures of zero over one at the same time um in kind of a generalized way that
like a flipping coin is a mixture now the difference between those two is that a flipping coin is a mixture just because you don't know something about its state like if you really tracked the way the coin was flipping you would you would know that it's going to be zero or
a woman hit the table a Quantum bit is different there's no internal thing to track it's at the end of the day saw the mixture and okay so we're just friends uh Jim you're uh you're gonna be a beginner and I'm gonna be the friend okay so you're you're gonna put me
um you're gonna put me in a box and you're gonna take a uh Photon that's gonna be in one of these superpositions of being zero and one at the same time and so that could be like a photons when they go through the air they spin clockwise or counterclockwise let's say
um and and this box that I'm in uh leaks no information out like it's really very sealed and so what you do is is you send the photon into the box to me and then I'm going to measure this photon and from my perspective when I measure the photon it's not going to be in this mixture
anymore it's either going to have been spinning count clockwise or spinning counterclockwise so I'll do the write down on my piece of paper I'll measure it and I'll write down it was counterclockwise I'll write down a zero or I'll see it going clockwise and I'll
write down a one so that's from my perspective what happens now what happens from your perspective outside the box well what's gone into the box is this Photon that's um some superposition of zero and one and then you've seen me make a measurement well you don't see me you
know I do because nothing leaks out but you infer that I did and so that means that inside of the box from your perspective there is actually a superposition of two different branches or worlds one world where I will measured the photon as clockwise
and wrote down one or counterclockwise and wrote down zero I think that's how I did um and the the thought experiment the sort of original suggestion by wigner is to ask who is right like from my perspective there's no superposition anymore it's either zero or one but from you just
sitting outside this box it's still in a certain position of both and that seems like something that shouldn't be relative I mean those are different different states Auto mechanically so how can that be relative so in the last couple of years um people have picked up this uh It Was
Written in like relatively philosophical language um originally uh had picked this up and they have uh some some some excuse me some different physicists um there's some folks at Renato Runners group chazza Bruckner in Vienna and Howard Wiseman Eric cavoconti lriful
um have and and Veronica Workman different collaborators of theirs have been working on formalizing this thought experiment into inequalities like these Bell inequalities but they're called local friendliness inequalities um and basically one of the so one of the um
there's a there's a bunch of ways of actually formulating them um but I can give you sort of the headline of of maybe one of the simpler ones so there's a paper by hadara and cavalcanti called them something like a possibilistic version of wigner's friend
experiments I think some something like that title and they make so the setup is you make two assumptions about these are now these are meta physical assumptions right kind of like little blend variables the first assumption is that if an event happens
it happens for all Observers Loosely I mean they do it more formally I'll just give you the like the looser version if the event happens it happens for all observing and then the second assumption um is something called local agency which means that I can make independent
variables there can be the independent variables that are uncorrelated with other regions of space time and and at its cooler this second one of local agency if you violate that you start to violate things that let us do science because you sort of need to have
reliable independent variables otherwise a response to any sign to any experimental paper could be you uncovered nothing about the physical laws of the universe you got that outcome because there was a solar flare on Elvis centari last Tuesday or something I don't know there's like some
like hidden correlation all over the place you could try and Rule it out slowly but um so anyway there's these two assumptions seem reasonable the verse is called absolute observed events second one is called local agency together we call those local friendliness
um and if you set up a version of this weakness for an experiment not with one friend in a box now with two friends and two boxes and a little bit of detail um then you can show that predictions of quantum mechanics say that those assumptions should contradict each other
and you have to drop one of them and they both seem pretty reasonable um so you know well maybe maybe I'll pause there um but like when I started reading about this I was like oh this is you know yet another one of these really intense questions that quantum mechanics raises
for us about how the world might behave and Goff we should probably test it I I agree and uh to even simplify it more it it seems that you you gotta give up something and you've either got to give up objective reality number one or you've got to give up free agency right one of those two
uh if the if we get to conduct these experiments is is going to go right yeah I mean it looks like it I mean yes and and it's important to mention this like this the objective reality we're talking about here is not it's not a objective reality like we disagree
about some subjective thing um you know in language um it's really something about physics and it's not like in relativity like in relativity we will disagree about what time something happened because one of us you know might be in a gravity well compared to the other um
this is where we we just we disagree about whether or not an event occurred at all um and um and Eric cavakanti has these amazing this amazing phrase for for what these little different realities are and he calls them wigner bubbles um and so the idea is that one of the
questions is do can you create these these wigner bubbles so these little it's sort of loose language but these little mini universes where events occur for The Observers in them but but not for people outside and the terminology is so great because it also lets you
have a sense of why this experiment is so hard to do these are very fragile bubbles um if any information about what's going on inside of the bubble leaks out to observers outside of it um then they kind of Pop uh and and uh and the this this violation doesn't
doesn't occur anymore um and so that's that's why at least naively the why but that's a way to see naively why these experiments might be tough to do and and where are we in terms of you know our ability to actually start doing these with some verifiable information as a result of the experiment
um so there's well the original the thought experiment as we've described it and discussed on this broadcast a person in the Box um yeah and you need to have a bunch of control over uh over a person and and people are big and and hot and heavy and Complicated by a lot of information
theoretic measures uh which make them tough to turn into quantum mechanical States um you know fundamentally we we should be quantum mechanical according to textbook quantum mechanics every everything is it's just that when they get big and complicated and heavy and
hot those quantum mechanical rules average out so mostly the things that we control that are as Quantum systems are small and light and cold um so what we've been talking with a few folks about are imagining a a series of experiments um where at one end
um you know you do the whole darn thing with the person or two people actually um in these boxes uh and and I have no idea how to build that and I I if anyone doesn't like that's sort of like a very the technology to do that is is is difficult to imagine but at the other end
um you run the experiment but you replace the friends in the Box um just with single photons so in this case you know you sent it in a photon in the box that I was in there doing a measurement but replace me with just another Photon and the two photons interact or something like that oh
another another photonic cube in some that experiment actually was run um sort of two two groups that run a little bit of this experiment the Nora tischler's group um in uh in Australia um ran it and uh and saw the violation uh so there's kind of this um
Spectrum where the Observer is a single Photon and the at one end and at the other end the Observer is a person and you can start you can sort of March up from the smaller to the bigger and at each stage you want to ask yourself what you have what your experiment has actually shown
you um so let's like look at this this first example with a single photon one reaction is to say you know pat yourself on the back and say job done we have shown that reality is not objective we violated the inequality um and uh and then we should probably go
for a drink and think about the implications or maybe even two drinks yeah yeah maybe maybe I'll call it one because I want to keep it like I said there's a lot so um but another response is to say look um you put photons in us to replace friends and and photons aren't observers
You could argue that um an observer kind of what this what's one of the fun things about is bringing friends experiments is they have these inequalities that they get to about these basic assumptions but they also start to butt up against some of the
open questions people have about quantum mechanics in general like these things called the measurement problem and this kind of too often when quantum mechanics is presented in a textbook form there's two sets of rules there's a set of rules which is the
starting equation broadly that describes how everything behaves and then there's this other rule for measurement of observers it's like when observers make a measurement and practically speaking we know how to work with these in the lab usually um but there's no definition of an
observer that is provided we just say like well when I look at it or when I use a device that to me looks like a measuring instrument then I say that's doing the measurement rule instead of the other rule um and so one response to the awareness friend experiment with with photons is
to say uh you know those don't count as observers so the the reality that that Observer had isn't isn't real it doesn't matter it didn't happen you know it wasn't Observer so it doesn't have a reality so there's there's no interesting metaphysical applications
um and so someone says that then you come back and say okay well but what for you is an observer and then and then they might say okay it needs to be two photons probably not they'll probably say it needs to be something bigger than that but you know
maybe they'll say something about some um maybe it needs to have some amount of math that they care about and there's some threshold of mass above which they believe the soup Position will collapse because remember how information can't leak so imagine if if the friend inside the box
is heavy and if the Met if you measure zero or one determines whether or not you're on the left hand side or the right hand side of the box if that's true then outside the box I can detect how the gravitational field shifts because of what the Observer measured so the information leaks out
um and so we might say that there's some fundamental collapse that happens due to due to gravity when something gets hotter so they'll say okay I'll believe you when you run the experiment and with something that's heavy enough and so maybe we work towards that and we
and we run it with something that's heavy enough and then one of two things will happen we'll either get the violation in which case they will have to accept some drop local agency or drop objective reality or we will confirm this collapse model that sometimes
something gets heavier you know it it isn't but you know it's it's not acting in superposition anymore that we we have some definition of an observer that we can agree on and it doesn't have to be something that's heavily there's other things you could argue about
you know information theoretically when it looks like um branches occur in quantum mechanics um and so the the program we've been thinking about is to start with this this this great experiment that in order and collaborators have done with one photon and start to March up to bigger and
bigger experiments and sort of rule out um you either find out that something is an observer and there's a collapse which would be a big deal or uh um get more and more evidence that object reality is violated I think is really one way to think about it uh speculate for me
um on you know these are two very big outcomes right I mean if if you have to uh basically agree that uh objective reality we have to give up uh in the very technical sense that uh it got violated through this experiment and a war uh we have to give up the idea of agency
and I guess I don't know I don't want to extend the metaphor too much but free will in other words it would essentially say that the universe is entirely deterministic am I right in that it's good it's sort of subtle points but people who uh um often people who drop that local
agency assumption are called super determinists um so these things are related but it it is subtle and and I don't know what free will is but but I do know a local agency is we can Define that but it is also still weird to drop that yeah so so so the speculation I'd like you to
give me is you know under under each of these let's assume that we have to drop the first one and and we have to give up uh in quotes objective reality or conversely we have to drop the second one uh and uh take the Super deterministic point of view
how does that change um you know the trajectory of of where we are going uh with the science gosh um yeah I mean I I don't even we haven't even spent too much time thinking about that because it's it's definitely in speculative territory and I should probably sit down with like uh
if anybody knows Greg Egan he's the science fiction author I'd love to sit down and talk to him about this he just read some amazing part sci-fi um uh look if we succeed at creating something that looks like a wigner bubble I don't know I mean I I
guess the answer is we're we're reaching the limits of uh of my imagination I mean you know it's here's one thing that comes to mind basically um which is you know that our Arthur C Clarke quote that any sufficiently advanced technology is indistinguishable over Magic
um and if we went back 400 500 years and we showed people some of the things we can do with electromagnetism today uh like talk to a piece of glass and have it talk back um then they would just look like magic and it you know we wouldn't fault them for not having enough imagination to
like connect the dots between like lightning and basic batteries and that um and I think what were getting at with Quantum Computing and Quantum Technologies more broadly because there's things in sensing and networking too we haven't talked about
too much yet but that are really exciting um are also just hard to imagine where we're going to get to in 400 years um you know if we're if we're really able to create little realities like uh you know movies are going to get even better like I can't imagine the technology to
even do with one person today but that doesn't mean that like a thousand years from now not gonna be something you know that's ad supported we'll figure a way out to monetize this is what you're saying yeah and and I actually agree oh with that one uh let's
chat a bit about the um what you just mentioned you know um which I'm utterly and always fascinated by which is the the movement of understanding uh ourselves our Sciences Etc and and the Arthur C Clarke quote is very apt right because if you even when we even made it current
right and and you and we went back to just say when I was 10 in 1970 and I brought this bad boy with me I'm holding up an iPhone for our listeners um people would like they would I think be able to understand the concept that this was a communication device because
you know after All Star Trek was uh on the air in the late 60s right and so maybe they'd make that leap and say is that a communicator uh and and and get it but the reason I use that example is like for many people 1970 is relatively you know not that long ago when when we extended to like 500 years
right of course 500 years ago uh we didn't have any of this right but I think that even if you went back to a relatively recent period with some of our existing technology I mean if you really wanted to blow mines you'd go back with gpt4 um and and and show them that
um is there a process and again this is a very speculative question for you and you can simply say I'm not going to speculate on that but is it is there as as one of the things that I a thesis that I have is that we have become and are getting better at abstraction
and as we get better at abstraction we are able to contemplate things like wigner's friend in the very beginning I I don't know if you're aware of the IQ researcher uh Flynn who's got an effect named after him um so he started looking at normed IQs from I think 1900 and then from Recently
he died so whenever he did that before he died and what he found was a consistent and and persistent increase in IQ now they always renorm it to 100 on standard Benet right uh but if you didn't read Norman if you just looked at the raw scores you would see that IQ had
increased relatively significantly from 1900 on and he was fascinated by that and one of his theories was that in the world of 1900 at naive real was still pretty much dominant in you know even educated late people um now it wasn't in scientists and it
wasn't in a certain group or groups of people but in an educated Layman naive realism was still pretty much what they looked at the world through that lens and and his theory theory was that as we learned more quantum mechanics and physics being a big part of the uh
increase of people's ability to understand abstraction um he said that that allowed for our ability to continually push that barrier forward right and because we could now abstract we could construct an argument uh whereas we you know the world of
thing that's so to speak of 1900 uh was giving way to this world which is is far more abstracted do you think that there is an upper limit to what we humans and uh like achieve under this kind of uh regime or or or much like we saw with Flynn are we just going to continue to be able to
push the boundaries further you know one could say something about like you know big theoretical limits on like if you have enough information density then you like fall into a black hole kind of kind of thing but but I'm not sure that that direction I want to go
with what you're talking about um because I think I don't know if I would agree on the night you would realism by the way I mean there's there's a you know there's a lot of I don't think any like a lot of religious and spiritual cultural things are definitely not ain't enough true big
guy true but but I do think what has started to happen in many ways technology is a integration uh [Music] mind and body um you know we're we're taking what we our desires often and putting them into the world and in a way that that connects um what we we are internally with
um with what happens externally and one example of that is people talk about what's going to happen when we have mind uploading or um when we have extended consciousnesses but we have extended consciousnesses today if you get rid of everything I fit
my phone and my ability to have telepresence and all this I mean I'm not I'm here and with you and will be with other people in this recorded podcast my memory is definitely stored in many many different bits in many many different data centers that I would be a
different person with without them um so as technology advances we are integrating ourselves into the world through applied physics and chemistry and biology um and Quant I mean Quantum is One Direction of that of that happening with very very weird and unintuitive versions
and I think one of the things that's most special to me about Quantum Technologies is this this integration of Mind and Body often seems like it might be one way I think people might even think it's well in the way it's we we have things we want to imagine we imagine inside of
ourselves we want to build them in the world to take care of us and take care of the people that we love um and like only if like physics didn't get in their way and mean that it was heavy and energy expensive and all this kind of stuff um but quantum mechanics is the opposite
you know it's not that quantum computers are saying here's all these limits to the computers you could imagine building it's actually saying I can do things that you don't understand what it's Computing like that's your challenge like think of this as a
Computing system and these rules that you didn't imagine exist and figure out how to use them um this analogy doesn't just apply to this kind of two-way of nature reflecting back on us that doesn't just apply to Quantum Mechanics I mean a lot of the reason they think we have trouble
with uh with climate and Natures we're not listening um to a lot of other kind of complicated systems that have rules that we don't understand uh ecologically or biologically but but Quantum is a very very striking version of it so anyway I yeah maybe there's some there are some limits but
um but we're gonna we're gonna change reality and reality is going to change Us in uh in at an accelerating pace I love that answer uh because you highlight the fact uh that many people often don't even think about and and that was in your response too you know
we already have that extension right there are bits of you uh in various systems that we have that we would have never had right so it I think for at least I hope for our listeners and our viewers they're going to be able to think I never I never thought about it
that way yeah that's that's quite true right and and uh then a little part of me is also in them or a little part of someone else that me that gave me that thought that is them you know it's the usual yeah yeah but but but wait but when you when you put it the way you did
uh it it makes it more practice at least to me it makes it uh uh more uh graspable to a practical uh uh understanding of yeah that's already going on right now and uh we're just gonna see an extension of that you know uh I'm a fan of David Deutsch uh who wrote the beginning of infinity and he
he made an assertion that I want to get your reaction to which is um anything that does not violate the laws of physics is true what what what what's your reaction to that um you know it's an interesting question to ask um I think I mean my immediate reaction is
um we don't know what the laws of physics are that was my immediate reaction too and I'm so you know we we have some rules of thumb that apply in different kinds of circumstances pretty well practically now um but they're very different from what we thought they were 100 years ago
and I think it would be naive to think that we have the answer now uh and so something comes after quantum mechanics and it's probably one thing I will probably bet on is it's not going to be more intuitive um it's probably going to be more weird
and so we would we would change ourselves more to understand it um so that that that might be the case I mean you know another reason that's hard is what it is to have a law of physics is also um a historically contingent thing um maybe I need to ask David what his
definition of law of physics is there but um you know one one of the interesting actually to be to be fair to Dr Deutsch he he did I I I gave the the last little line of his statement he he did preface it by saying the laws of physics are always changing and hopefully we are
adding our to our understanding of them but that's where you uh sort of the uh whatever doesn't violate the then understood laws of physics is true uh I see yes so then truth you know there's this fun book which we won't have to talk about by Applause for named Ian
hacking who unfortunately recently passed away and the title is historical ontology and I think that's kind of what we're talking about here yeah I I love that I'm gonna pick that book up uh so thank you for the recommendation so what's next for you like uh you know in the
applications you did an excellent job um and you know we're an easy choice for us but one of the things you did a really good job on was in our application we asked what what are the implications of if you're what you're using this grad for fail and and you
gave a really great answer in fact probably one of the best that we saw a lot of the a lot of the uh the ones that we got um didn't didn't really dig in like you did uh on because what I loved about it was your attitude of how much you can learn from failure um and and and so uh
what's what do you hope what's next over the next uh the balance of your fellowship and and then beyond what what are you going to be hoped for milestones well in the context of trying to figure out how to run something about Winger's friend um
there's a bunch of exciting family I guess I mean so really what we're trying to do next is make well first off make sure that the basic Theory all makes sense in a couple of the papers that have been put out over the last couple years talk to some more people kind of get get a
little bit broader circulation um and secondly come up with fundable Milestone experiments um and uh understanding what a good Milestone is starts to get to theoretical questions about what makes a an arguable observer um and and that that's a like a fascinating question about a couple
ideas um there's lots of people who are working on on different kinds of um things related to this in Quantum information and then the second bit is fundable so like what can we actually build that is a definition of Observer that makes sense I mean like One
Direction you could go is it's a biological system or it has Consciousness so we'll put something else that has Consciousness in there and and I I also don't know how to build I don't know how to put things in Consciousness in the superposition um if anybody knows who's a listener
knows what is the smallest arguably conscious biological system smallest and physical scale I would I'd love to get their take on on that um but even still uh that's tough so um so the the real Milestone that I'm targeting next um with in terms of weakness friend
is picking out some specific well-defined milestones that can have a budget against them um to go and build things that are bigger than a single Photon but a lot smaller than a person but still be but still be meaningful um and yeah if we fail it's because we broke the theory which would also be
fascinating um or we fail because uh you know it's like really not fundable and we just don't have enough technology as a species yet to do it but I one of the one of the things I actually haven't mentioned this so far so I should but one of the things that I'm thinking
about is putting quantum computers in as observers and having them run programs right I mean quantum computers are big Quantum systems that you can control um where big can have a bunch of different definitions um so that that's a direction that that
we're pretty excited about piggybacking on a lot of the industrial investment that's gone into into Quantum Technologies and we'll continue to very cool well well this has been absolutely fascinating as you know I uh this is like catnip for me um and and I'm delighted uh to see the
work that you're doing and uh all of your colleagues as well as your foundation I think uh bringing this all together exposing it to the most cognitively diverse group of people is the way to gain footholds that you wouldn't otherwise so I think that your
strategy is bang on um and I'm delighted that we can be a small part of it um at the end of our podcast uh this this will be a really interesting one uh we we always ask people okay so we're gonna have a thought experiment and we are going to make wave a magic wand and
make you the emperor of the world for one day you can't kill anyone you can't put anyone in the Faraday cage uh head you can't uh put them in education camp but you can we're going to give you a magic microphone that you can speak two things into and the next morning whenever the next
morning happens to be for the 8 billion plus uh people on the planet currently they're gonna wake up and the two things that you incepted them with they're gonna think were their own ideas and they're going to say you know what I'm going to act on both of these starting
today what what two things are you going to incept in the world's population so it's like a big Communicator rather than Emperor power it's just the yes the Philly pulpit oh man um I know where the first one is everything about the second one I don't know maybe it's a big clue but
um we could talk about it more I think the first one is um what are quantum mechanics uh I really mean that um I think it's like a I don't know so a lot of people in our field you know go through a bit of a it's a sort of a breakthrough moment when you
realize that the world is like that and and there's a there's a humbleness that you have to take on to realize that it's not only ways that is different but very different than what you thought but even at the very very core fundamental level sorry number one
is learn quantum mechanics um the second one is probably spend more time with your family laughs I actually loved those in yeah in connection with each other uh I I think that the the world would be a much more interesting place if both of those got accepted into our into our population
well listen this has been a tremendous amount of fun for me thank you so much we will include in the show notes all the books mentioned and the various uh websites uh that where people can find you um and thanks for being with us and back and thank you for being our first fellow
this you know there's a bunch of stuff that can get funded around normal paths easily uh you know incremental academic papers or the next generative AI startup um but there's stuff that falls through the cracks um and often those are the the really
exciting and interesting things um so I'm like we're just friend Foundation experiments or some of these open source Quantum public goods so thank you for that support and then I guess the last thing I'll just say is if anybody's looking to get involved in Quantum Technologies
um unitary.fund is the website and we do a bunch of Open Source stuff we just concluded a big hackathon um might be easier than you think if you're technical prereqs are basically Python and linear algebra for a whole bunch of things and so I hope to see some of you there