sam: https://twitter.com/sflamini5
paul: https://twitter.com/PaulRBerg
max: https://twitter.com/maxdesalle
shafu: https://twitter.com/shafu0x

The beginning of infinity: https://en.wikipedia.org/wiki/The_Beginning_of_Infinity
The fabric of reality: https://en.wikipedia.org/wiki/The_Fabric_of_Reality

I'm your host Shafu and Behind the Bytecode is a podcast about more philosophical kind of topics.

What I would call layer 0, so things like social consensus, crypto philosophy, basically all the things that make the by-code possible in the first place.

Today's episode is about Karl Popper and David Deutsch, mainly about Deutsch's Thoughts as outlined by him in his 2 books, The Beginning of Infinity and The Fabric of Reality.

And what we're going to try today is connect his thoughts with Crypt, and how they relate with each other.

For that, I have a really great group of guys on today.

We are not only really passionate about Crypto, But what's really exciting to me is we're also really passionate about David Deutsch.

So I had a really hard time thinking about how I'm going to structure this, because Obviously, David Deutsch's books are super dense, a lot of ideas and all of them are really important.

Sam, if you could summarize the main important ideas for you personally out of David Deutsch?

So it's quite difficult to summarize David Deutsch's work in just like a couple of minutes even.

I'd say for people listening, the 2 books, just to highlight them, that he's written are The Beginning of Infinity and The Fabric of Reality.

So in those books, he lays out, in The Fabric of Reality, he lays out a kind of really fascinating worldview around what he calls the 4 strands, right, where he goes through the theory of computation, quantum mechanics, epistemology, and the theory of evolution is like defined by Richard Dawkins, so neo-Darwinism.

And then in the beginning of Infinity, at least my take on the book is that he goes far deeper into epistemology than he does in the fabric of reality.

And for me Personally, all the most interesting ideas, at least for me, this kind of connects to Karl Popper, are the epistemology bits, right?

So there are topics we're gonna discuss today that have been very impactful to me, like this concept of every single problem being soluble, as long as that solution does not defy the laws of physics, right?

That might be 1 thing we discuss, but the big ones are just around how to actually view science in a practical way, right?

So all the things you grew up learning about how the scientific method works, all these things where maybe you're in university, you're talking about p-values and you get into Bayesianism and statistics, Deutsch kind of really refutes a lot of that thinking and presents science and the scientific method in a way that I think is far more practical for people in their lives.

And I think it's very—all of the things that you discuss are very applicable to what we're doing in crypto because we're trying to build new things.

Every new protocol is a conjecture about a new system, right?

And there are all kinds of things that we have in crypto, in DeFi in particular, that help us criticize these systems, whether it's hacks, audits, other things going wrong.

So for me, his books and his thinking have given me a really powerful worldview, I guess, to explain what's actually happening in the world around us and be more useful in our industry.

Paul, Max, Jeff, what would you add to my summarization?

I'd say that those are great summary and a great starting point.

My background with Deutsch and Pulper is that I started with the beginning of Infinity 3 years ago.

Then I continued with the fabric of reality.

And I agree with you that, By the way, should we use the shorthands BOI and FORR, because they will probably be used a lot on this podcast.

So I'm gonna say, I read BOI in 2020, and then I continued with FOR in like the next year.

And I agree that the fabric of reality was more rooted in physics.

He focused more on, you know, the many universes aspects of quantum theory, which were not, and they continue to not be taken seriously.

Whereas, BOI is more focused on the place of humans in reality, right?

Like why we are important and why knowledge is key to understanding the world.

And, you know, he gives that great example with the champagne cork in the, you know, the scientists' labs.

You cannot explain the behavior of that champagne unless you can, you know, explain astrophysics.

And, you know, over time, it took me a while to understand, to grapple with all those ideas.

But over time, the way it, like this worldview has helped me is just help me correct my errors more quickly.

As you guys are, you know, developers and like, you know, working in crypto, we make, we build new products.

And, you know, making new things will necessarily involve a lot of errors at the beginning.

And For me, at least, you know, Deutsch's worldview has just helped me first understand that I will always make errors.

And the point is just to make them more quickly and correct them.

And yeah, it was super practical philosophy.

And you know, that's why I immediately said yes to this podcast is because I just love learning about this and sharing these ideas.

And I think in the current context that we find ourselves with chat GPT and AI, these ideas are even more topical than ever because the world is mired in misconceptions about the future of knowledge.

think Deutsch's work is tremendously optimistic, and it really shows how optimism is true in a sense and how pessimism is just inherently wrong.

And I think, I mean, Deutsch's work is so, covers so many topics that you can't really summarize the books, you have to really just read them.

But aside from all the topics which were mentioned here, which are indeed incredibly interesting, I think what I found the most interesting was, I think it's at the end of the beginning of infinity, it's a chapter around beauty and how there is actual objective beauty.

You know, if you take a piece of Mozart and you change just a single note, it will just sound wrong, right?

And which is just proof that there is actual objective beauty, that we can create more beautiful things.

And I think these types of things which correct misconceptions, as Paul said, you will find plenty of them in his books.

And it's about addressing all these biases, all these misconceptions we have about the world, like the spaceship earth metaphor for example, which are wrong, which usually a lot of people take for granted in a common sense, and which are just wrong.

Yeah, so to me personally, for me, the main takeaway is, as Paul said, error correction, right?

And to make that more specific, we all moved, this is very specific, we all moved from hard hat to foundry, right?

It's really hard to put into words, But a great summary would be because you can fail quicker, right?

And this allows this iteration to be quicker.

And so you can improve and move into the right direction.

The other thing is about error correction, where I really have a hard time, and in smart contracts especially, is immutability, right?

Because on 1 hand, you know, crypto values, 21 million, all of this stuff is about something that doesn't change.

And me as a small contract developer on a stable coin right now, we really, we are really trying hard to build something that isn't going to change, right?

But on the other side, something immutable, you cannot improve that anymore, right?

So I wondered how, how you guys are thinking about something like that.

Speaking of Foundry, just to quickly comment on that.

I mean, it's The idea is that you can run your tests more quickly, like 100x more quickly.

And what do you write tests for to fix the bugs in the production code?

So it's literally the best example for crypto devs.

like test runs, the best example for like, you know, crypto devs, you know, like 100

techs, like more, you know, quicker, like tests, like test runs, you correct your production code more quickly.

Speaking of stablecoins and like immutability in general, well, I'm not a stablecoin expert.

But what I do understand is, I mean, there is a certain tension with stable assets in crypto due to the reasons that you said.

And like the way I think about stable coins is that there's always some honest responsibility behind that.

I mean, even if you forget about crypto for a bit, the stability of like fiat is not for free.

And I mean, by proxy, you're paying for that, if you're paying taxes, of course.

And therefore, there is no such thing as like true stability in the long term.

It's just that somebody is, you know, taking a risk for you in the short and the medium term.

And I mean, it does create society, right?

Like, it's good that we have certain assets that provide some sort of, like, peace of mind to people.

I just wanted to flag that, you know, if you're going to go really deep philosophically, there's nothing stable.

It's just like an economic term we use to build society.

Maybe that would be helpful to set the stage, like maybe explain why we're talking about error correction so much and what we mean.

So in the fabric of reality, I think it's the fabric of reality, Deutsch basically lays out, I think it's the chapter on problem solving, where he basically lays out how knowledge grows.

And it's somewhat different than I think what most people are conditioned to believe if they just like read the headlines about new scientific studies and are trying to understand them in the context that like the mainstream media would try to explain them.

And the way that Deutsch lays out how knowledge grows is it basically always starts with a conjecture.

And some explanation about why x is a certain way that is hard to vary while still actually explaining what it purports to explain.

And we can talk more about explanations in hard to vary if we need to.

And then criticism is probably going to poke holes in whatever it is, and it's probably going to poke holes in whatever the initial theory was, the initial conjecture was before you even run any kind of experiment in the first place.

We're coming up with ideas, And most of the ideas we have, before we even implement them or try to implement them in the world, we shoot them down on our own head, via this process in our own minds, right?

When it comes to what you just mentioned about immutability, there actually is a tension here.

Maybe not like on an industry wide scale, if you really zoom out, right?

If you really zoom out and think very long term, if someone wants to build an immutable protocol and they feel super strongly about deploying this immutable stablecoin protocol that's going to have some kind of algorithmic process to do this.

Well, like, you know, it could fail, it could get hacked, or the whole system could fail.

And the entire industry could decide to create a new conjecture and build a different system and make that immutable, right?

And we just progress kind of like evolution progresses relatively slowly, biological evolution that is, relatively slowly, as each project kind of dies or runs out of money, and we push this forward.

So that would eventually probably happen.

I think the tension, though, is with all software, it's all fallible, right?

We all know that we write software, and there's probably something wrong with it, right?

So to say, to have the confidence or the gumption to say, you know, I'm gonna deploy this 1 protocol and it's gonna be good for forever because of X, Y, and Z reason, that's not really epistemologically sound or deutsch.

So this is something that we all do need to wrestle with.

And honestly, in your guys' case, I'm curious, Paul, how do you think about this?

It's kind of 1 of the more lindy projects, right?

You've had a lot of time, a lot of exposure to the market.

I mean, how do you see this tension between immutability and the fact that we're all fallible and write bad software?

I think Deutsch's point is that all knowledge starts with a problem.

A conjecture comes only after you have a problem and then you propose that conjecture to solve the problem.

Don't quote me on this, it's just like my folly ball.

In terms of like immutability and like contracts and like Xavier, yeah, for context for like listeners, we first shipped Xavier v1 in

and we are glad that we have never been hacked.

You know, it was a simple protocol, like a simple single contract protocol, but still We are fortunate enough that funds were never lost.

Now, in the meantime, we're starting to work on a new version, and we will also be— I mean, we are also exposed to the same problems, like how do we

we innovate while at the same time, taking care of immutability.

We don't like having upgradeable contracts for sure.

And I guess, if you think about the long term, like the real long term again, what we are building now with my contract is just, you know, temporary software that will solve a problem in the short term, and then something else will come along in 5, 10 years, or even shorter than that.

And then we'll have to re-implement our systems using that new paradigm.

The benefit of immutability is that it gives you that short-term benefit of like, there's nothing we can do about it.

And it solves decentralization problems and just scales up the software.

But I don't think for 1 second, I think I tweeted about this like a few months ago, I don't think even the EVM will be around in 10 years.

That's like a bold statement for somebody who, you know, my life's work is attached to the EVM.

But I think my take on this is that, you know, we're just having fun.

We have to build them and like ship them to see if they will solve problems.

It's just that I'm trying like not to cling too hard to, you know, the fact that, yeah, it's mutable, therefore it will forever be successful.

I mean, that will forever have users because of that and because the underlying attack will somehow magically remain forever untouched, unlike anything else in the universe.

Yeah, sort of the long rambling answer, but I guess that's my take.

Yeah, and I think a lot of what contract building is about, is about creating these simple contracts, removing as much as possible from them as you can.

You know, whatever can be done outside of the contracts, outside of the protocol, should be done outside of the protocol, precisely because of, you know, immutability and because it's just hard to fix things when it's already on the blockchain.

And that's also why, for example, we've been seeing this rise of programming languages like Rust, for example, where the whole focus is precisely on error correction directly from the start, right?

So when you start building in Rust directly from the get-go, when you compile, the compiler will be very, very strict and will raise a lot of bugs and errors it's able to find in your code.

You know, you could see languages like Rust as kind of like a Popperian in a sense, because they focus so much on error correction.

And I wanted to come back to something which Sam said earlier, which is that in crypto, the markets, and actually the market in general, but especially in crypto, is a giant conjecture refutation machine, basically.

So anyone is able to conjecture something even more easily than in the normal economy, because you can just set up a DAO in like a few clicks, launch your ID, and then the market will criticize it.

If you have users, if your token goes up or whatever, you know, you may be on something.

If not, if you get hacked, if no 1 is using your products, well, there you have it.

And so you can see capitalism and specifically crypto because crypto is really like free markets all the way as a sort of like gigantic contractual refrigeration machine for progress, which is, I think, really awesome.

No, I was just going to say, I think you're spot on with that.

And if you look at crypto and compare it to something like traditional finance, the speed at which this process can happen in crypto is so much faster and While defy looks tiny and it is tiny compared to traditional finance now the speed at which these systems are going to be able to evolve because they've been set up in this very open way.

There's really been no intentional setup.

They've just been allowed to exist outside of this system where anyone, anywhere in the world can come up with a conjecture, right?

They can try to solve a problem that they had, whether they're an anon that lives in South America, or they live in the epicenter of technology in 1 of the major cities in the Northern Hemisphere, right?

So I think it's really worth highlighting that because it's a big feature of, I think, what we're working on.

Yeah, we are sprinting through financial history.

And sometimes it feels, you know, it doesn't, it isn't perfect, but it shouldn't be.

So sometimes it feels like we are reinventing the wheel.

Sometimes it feels like we are reinventing the stock company, right, where you see governance and stuff and you're basically saying, hey, wait a minute, this is a stock company without equity rights, right, which you often get.

But yeah, we have error correction on our side and that's why we are all so bullish on this, right?

Because we believe basically that that's the only process that we know so far to make progress.

I don't think it's a coincidence that streaming is being developed in crypto.

You know, I don't have any data to prove this, but I'm sure as hell that 10 years ago, 15 years ago, there was some guy in a bank who said, let's build buy-to-second payments.

And some guy in a suit above him told him, fuck off.

I'm not going to do it because the infrastructure is like too hard to build or, you know, people want this or whatever.

And here you are, you have, you know, 2 companies on this, just on this call.

multiple, there's like an explosion like streaming companies, right?

Like, like nowadays, especially for like fasting tokens.

And we can do this because we didn't have to ask for permission, we just built it from our laptops at home.

And that's exciting because we are putting forth a new money Lego that didn't exist before and we can rapidly innovate it and so on and so forth.

I just want to quickly go back to 1 of Max's points about Rust and Popper.

I think I had a tweet about this, but I said that Rust is what Popper would have developed if you turn to software engineering.

I wrote Rust code for a bit, I published this tool on GitHub, and, you know, I'm not an expert by any means, but you only have to spend a little bit of time in Rust to realize that it takes a hell of a lot of time just to make it compile.

And the reason for that is that it literally throws errors at you.

And it's like having a mini-popper next to you saying, correct it, correct it.

And, you know, it's a really nice paradigm.

And I'm glad, by the way, to see new programming languages like Sway, which combine the best of both worlds, where you can have, you know, like the solid, like, ironically, the solid build of like Rust with Solidity, which has the Contracts API.

Yeah, I was going to bring up Sway as well, because it's another good example of people taking old ideas, right, and trying to build new useful things in our space, and to being able to do it relatively quickly, right?

I mean, we're already talking about building the systems we've been building in new programming languages and on new virtual machines, like fuels working on a lot of different things in this regard as well.

And in traditional finance, they're still buying mainframes and things run on COBOL servers.

Yeah, I wanted to come back to your analogy, Sam, about the market.

Like, it's so crazy that the stock market is so inefficient when you think about it.

It's only open for certain hours every day.

And then you have crypto, which is just this completely open system which anyone can access anywhere in the world whenever you want.

And it's just a much better error correction mechanism.

Like, you know, if you have like an arbitrage opportunity, for example, which is an inefficiency in the economy, which is an error, it will be fixed much more rapidly, simply because anyone can fix it, not just people who have a war in America or US citizens or whatever.

So yeah, crypto is the ultimate error correction mechanism for finance in that regard, I think.

There's also like a question of, or like a benefit of providing cheap global public goods, because you have an evolutionary process that filters out the hacked protocols, right?

Like, you know, those that get hacked and continue to get hacked over time will simply continue to lose users.

I mean, I don't want to name names, but everybody knows certain protocols that just kept being hacked over time, especially for like lending and that kind of like financial, super financial use cases.

And the benefit of the public, of like the layman, is that you have the publicly certifiable proof that this system has functioned for like 24 7 for like 5 years.

You simply cannot have that kind of guarantee with a bank, as recent events have shown us, even the greatest and, you know, Silicon Valley bank, right, like failed, right?

So it's like a different kind of evolutionary process that happens in crypto, that you just simply cannot replicate with Teradfai

Yeah Let's talk about the EVM guys because you Paul just said something really interesting because you said you probably think and again right impossible to predict the growth of future knowledge so take this with a grain of salt but basically you said in 10 years you don't think there will be the EVM or basically it's maybe it's something totally else or it just changed so much that we wouldn't call it EVM anymore.

And I actually disagree here a little bit because of universality, right, which is also a big theme in David Deutsch's books, where I think that the EVM is Turing-complete, right, it's universal, or Turing-complete at least in the theoretical, not really in the theoretical kind of sense because you don't have infinite storage, but let's say, let's say for the sake of argument it is, that through universality and network effects that it's good enough, right, that it's, it's the same story and again impossible to predict the future through the past but it's similar to the JavaScript story, right?

Where we had the thing, it was universal, it was good enough, we optimized the shit out of the EVM, right?

And we built on it and through something like layer 2 or layer 3 or whatever comes ahead, you have the ability basically to escape the EVM on higher levels but on the bottom it's still the EVM because of network effects because of Lindi because of security tools and so on and so forth so What do you think about that?

It's just that, you know, it's an educated path that I made publicly and I continue to make based on my hands-on experience building on EVM.

I think we had an exchange about this on Twitter when I tweeted this.

Now, I think I could be wrong, but I think you said that, yeah, I mean, JavaScript, you know, is ubiquitous today, even if there's like various better alternatives.

And let me just restate what I said on Twitter, which I still hold those views.

Firstly, I think the fact that JavaScript is ubiquitous today is a mistake.

There are so many people lost money, time, and so forth because somebody didn't use the correct types in the back end, right?

And like with TypeScript, at least, they're like better alternatives, but with like TypeScript, you at least like get that, you know, typing system on top.

The other thing is that I think JavaScript was originally developed in an environment where technology was like less flexible and less adaptable.

You know, JavaScript was basically developed at the same time with the hardware that we're using to record this podcast.

And people were just buying them for the first time and they were just using them for the first time and so on.

So there was like limited bandwidth for newcomers to, and like developers to say, I'm going to use this and then I'm going to switch.

It was like that, you know, Cambrian exposure of like, let's build the first ever internet companies.

Now we are in an era where we've passed that time.

Like we have the internet wires, you know, and like we have the internet set up properly.

And I think we are much better suited as like a society to say, you know, we can swap out software, like we can replace software.

Yes, there will be some path dependence, like in certain cases, like maybe ERC20 will stay forever, I don't know, within the EVM.

But I don't see any reason why entire new platforms cannot be developed in parallel and eventually garner more users than EVM.

To expand on that from a hands-on developer perspective, you can throw in as many layer 2s and modular blockchain, quasi-modular blockchain architectures that will boost the EVM till I don't know millions of transactions per second, that will not solve the problem of account abstraction, which on Ethereum is a pain in the neck because, for example, you have a token, you have to approve it, and then you have to transfer it.

But you can't do those 2 separate transactions like with 1 Ethereum, like with 1 user signature.

And I have hands-on experience with this because I built this proxy contract called PRB proxy, which, if you're familiar with DS proxy from Acre, It's super similar, but I just developed it using the latest Solidity version

I wish I didn't have to put in the work to think about it, design it, test it with Foundry and so on and so forth because it's a liability.

I wish the core protocol provided me with tools to do that.

And, you know, like again, that's 1 issue which will never be patched by multiple millions of transactions per second, it's something that is about the qualitative aspect of the virtual machine.

I would also say that back to the topic of error correction, over time these things kind of calcify and become much harder to change, especially when they have more users.

See all shafu transcripts on Youtube

Behind The Bytecode #0 - Crypto's Beginning of Infinity