Les Amis d'Ethereum France #3, avec DeFi France — La Sécurité pour les devs de dApps

Okay, thank you everyone, we can get started. So, delighted to welcome you tonight. My name is Latence Exr, I'm the Vice President of Détériore France. I'm also part of a project called Six Mots, which is hosting you tonight since you're in the offices of 6 Mo, but also Jelly Faille. We're delighted to be doing the third edition of "Détiore France."

Its aim is to bring together the best of the French ecosystem and theron community in a broader sense. We did this with 6 Mo Djelfa and Morpho Mangrove last month, and this time it's with Ten Faille France. We 're super happy to be doing this with them; it's a large French community. And I might introduce myself. It's a slightly different format this time since there will be a panel and also a remote presentation, so it's a test.

We'll see how it goes. I think it's going to be very good, but it's a test. And it's a superb edit; I think we're improving a lot with this edition. In publishing, and I'm quite proud of it, let me quickly introduce you to our organization. We've been in France, so we're a non-profit association since 2015.

We bring together the terrarium community. When I said " around the French-speaking community" in the broadest sense, we have a website called François. We publish articles, technical articles, and we've translated technical articles. And above all, we organize community events around terrariums in France. We used to organize monthly meetups with friends in the field in France, and we've been doing other meetups in Onex for some time now, as well as workshops.

We've also been organizing a major conference since 2017, which we're very proud of. It's great! We're organizing another one this year, so we're really looking forward to it. What else? You can find us on Twitter.

We have a YouTube channel where we 'll publish these methods, which are filmed, I should point out. And a Discord server that we're opening, which I invite you to join, which is slowly but surely taking shape. A quick word for our sponsors: I find it much classier than... Sponsors, but it comes down to the same thing. We learn from Vala, which finances open- source projects.

It's really been around for a long time, they're not well- known enough in my opinion. They give money to cool projects around terrariums, so they finance our half-time work and then gave us the money for this, so it's really super cool. We thank Riom ES Para Swaps who sponsor us for this goal, maybe for others as well, and his pout, Ryjkov, who does our recording. So really, thank you to them. A quick addition, and we'll see a little intro to Difference anyway.

By the way, you can start coming here. We're going to have a remote presentation from Adrien Courbevoie of OPEP Line, so it's going to be really cool. It was a panel with Mounir from Parra Swaps, Charles from Ledger, and Adrien from 6 MO. So, without further ado, I'm going to hand it over to Brice from Difference. Thank you.

So yes, just a quick point for those who don't know Ten Flaws France. We are a community. A slightly more recent one, which was in France, a bit smaller, it's a little less structured, but that's also what makes it fun. So it dates back to 2019, it started as a Telegram group, and the idea is really to exchange information on the A10 vulnerabilities between enthusiasts initially, then over the course of three years, many of these enthusiasts have become professionals in one way or another, whether formally or not. There was also a migration to Discord.

I'll let you discover all that, but anyway, search for "Ten Vulnerabilities France" and you'll find the same thing. We have a YouTube channel, we have a whole bunch of people doing really interesting content production initiatives, so if decentralized finance intrigues you, don't hesitate to check out the French threads. Thank you very much. So now for the technical switch, we'll hand it over to Adrien. Well, hello everyone, sorry about the lack of in-person interaction, I would have liked to see you all and participate in the panel afterwards, but here she is, Lyon, her family, but It's still a bit far, so we'll do this remotely.

So, let me introduce myself quickly. My name is Adrien Crew Bois Couloir, I'm a smart contract developer at Open Zeppelin (Open the Pianist something like that). You may have heard of them for auditing, but we do much more than just auditing. Let me talk about it a bit, and today I'm giving you a quick presentation on digital asset security. It's a very, very broad issue.

I'm going to approach it from my perspective as a smart contract developer, so naturally it's biased. I'm mainly interested in vulnerabilities and, more specifically, the difference in practice. If you're a fanatic or just a regular reader of our blog, don't hold it against me. I think all the advice I give you can be applied to all contexts, whether it's Bitkom or others. So, a quick overview of this European company, Zeppelin.

You probably know us, but we're actually an international company interested in securing this... What we call the open economy, so what our city of ADIE is noting is that it goes beyond all those who finance these parties on ski trips. The economy is... we are... we are...

we are... we are followed by numerous projects that trust us. As for audit services, you may also know us for Paulo Smart Contracts. If you have developed in hardware, you have probably already used our contracts to make NFCs, to make water... their CVs or other things.

We really have a suite of products, and the idea is to support developers in securing applications, the integration of the Hague Diff, whether upstream with contract writing and auditing, but also downstream with applications and live via Defender. So, my vision of security on the Blood Chain is really the Gruyère or Emmental model, depending on your favorite cheese. And the idea is a model we've seen quite often: it consists of saying that all security methods have flaws, or areas where you didn't know about them, and so... Basing security on a single system or solution is doomed to failure. Security processes must be accumulated throughout the entire development cycle, from application use to the very act of adding them.

It's only by incorporating all these imperfect solutions that we can asymptotically approach a perfect solution where there are no loopholes. That's why today I'm going to try to talk a little about my vision of asset security from two perspectives: first, the user perspective, which might interest some of the participants in 2009, and second, the developer perspective, which is a complementary and also very interesting point of view. So I've tried to identify different points and highlight how things can and might not work, and why we need to be careful. And then, at the end, there will be recommendations. Obviously, the first point concerns wallets, Brother extensions, and so on.

I don't know if any of you... It's used, in particular, to ensure that our toilets are secure, and they have vulnerabilities, as this tweet shows. Someone saw a pretty nasty hack where a hacker managed to change the application, put a mask in their browser, and get them to sign transactions they didn't want, or extract the private information. These things happen in these cities; we'll talk about it a little later. To prevent this from happening to you, there aren't many solutions.

One simple one is to not store your private key in the browser, even if it's an application with security. So Ledger is a good solution for this. It's not just gamers; there are others, but of course, in your arsenal as players, you have a Ledger somewhere. It greatly improves your security. Incidentally, Ledger or not, when you have to sign a transaction, be careful what you sign.

It's sometimes a bit tedious, but you have to check before signing. The second thing is... This is the next step, this is the interface. It says that if your Ledger is properly configured, if you're asked to sign something malicious and you do, you're apparently getting scammed. There was a fake game of chance today.

There's a fake mobile app that was a hoax, a story. So you connect your toilet to the rooms, you think you're making trades on a story, AP, but in fact you're transferring your funds to someone else. You can get scammed very quickly, so it's quite dangerous. This is true for mobile apps, it's also true for apps on platforms, on the internet, on stops or anything else. There was also a story like that about someone who had, I'll repeat, an attack that came from the middle to replace, I hope, my crypto with a clone of my crypto.

But when you enter your private key, well, of course, it's Lalique. And sometimes there are solutions, which are simply to download the interface, when possible, onto your role, onto your disk. I think Michael Tôt did it for those who know about this. It's going to be It's possible with a possible exit, but often, in fact, it's all on the client side. There's the choice of platform, which is always very important.

That is to say, even if you secure your system well, if behind all fronts you entrust it to a poorly designed project, whether centralized or decentralized, you can say goodbye. We have a good, well- known example from my blog: Kone's centralized solutions. My job was just in Arak, or in Europe, where we'll never really know. I think everyone has their own opinion on this. Beat Connect: you had to be a bit gullible to believe all the profitability claims they made, but even without that, applications disappeared.

We think it's secure and safe, and we're never truly safe. They're safe if it's well done, but now the user, a priori, isn't necessarily safe. What's in the code, and if there isn't a way for the user to exit, we could also think that these are risks that will also apply, moreover, as Developer, if you had ether on Solana with Bridge and by Wall this week, you must have been sweating bullets, and maybe ten different applications were built on this value of the Bridge state. It's safe, so there you have it. Chain blocks have inherent risks, and that's true for all the blues.

They take that to be true, for XD, it could be true to lower it, it could be their wife for polygon. So, a few moments of recommendations for users. In fact, there's not much to do. What you need to do is try to be smart enough with your veil and not let that linger. That's what 's prized anywhere.

Use solutions, whether it's smart contracts with recovery mechanisms like silver, arts, Douar Ouled, or Ledger. You can even combine the two. That's what I do. I have silver behind a Guardian because the players... it works very well.

So, avoid over-trusting the interfaces. If you can download them, that's well done. Be careful with Bridge. Bridges have let themselves... The news is the new centralized platform, and of course, always maintain a critical perspective when a new application with vulnerabilities comes out promising 15% per month.

Well, that's still a lot, you have to be careful. Certainly, as soon as the calculations and feedback show more than 100 % per year, it means they're calling it a scam. It can happen, but the risk is always there. The second phase is the class, which contains Ribwar. Now, we might have looked at the application developers' side.

So, just walk around the room, there are some who are application developers, or who want to become one, or who are just curious about how applications are developed. Well, the same as before, the choice of the belote chain is important. That is to say, before anything else, when you're going to deploy your application, ask yourself where it wants to exist. If you want to avoid the controller due to significant storage costs, that's perfectly understandable. If you're going on PC, it will be a problem.

If you're going on Solana using player lending, it will be problematic. Malls posed a problem, and even if the chain block is, in quotes, viable and capable of handling particular problems, think about it separately. In what context is it maintained? The skis followed, there was a little story a few months ago about the lack of customer support and the hatred of the chain by Carbine, who underinvest, and in the eyes of some users. So, there you have it.

And tell yourself that when you know it's a block, there's no end. Again, I might make enemies by saying this, but there's no magic solution. That is to say, if a blog post tells you it will necessarily be faster and cheaper, it's because there's a trade-off somewhere, perhaps related to security. And often, well, if it were enough to reconfigure the terms, why would that make it work by magic? But it's already been done.

When we reconfigure, we centralize. Potentially, other problems appear: dependencies, our firmness. It's a very, very, very broad class of problems that can arise. Again, dependency It could be dependence on a bridge, as in the case of Devoir Maul; it could be dependence on a price oracle. There are a lot of AQ/CQ recently that are linked to this, and interoperability is always something to consider regarding how your application can be used, especially with regard to everything that's hot.

That is to say, when we have an idea of ​​the events that have been taking place for a few years, for the last two years, Faye disease has exploded, and AGS and Define are also exploding, and more than half the time it's the combination of a price rule that gets attacked with flash mobs that allow exploitation with a built-up volume. Sometimes even the flash mob is used to hack their prices. So when you have independence on an external system, think carefully about what they bring you, but also the risks they present. Can there be doses that can happen? Finally, to talk about what he's saying, just from the back, what can happen is, what about you, the users?

are at risk if it doesn't work, especially if you trust another contract that you use. The independence of our system, if the Alba sandstone, be careful because the CICAS needs you and an exit strategy and avoid reproducing known bugs. Our beaches could have gotten you all out, friends and girlfriends, done in terms of contracts that you can reuse rather than writing them yourself, but instead of that, I released statistics on the origin of the most classic bugs. So there's the desire, % CD, logic error in a code, that won't help you design it well. You have to do a lot of testing, you have to audit it, but there are things like unauthorized access, the 1.

36 stop, men's vagueness on integers, all that, these are things. Normally, all the basic building blocks to do it well exist, so rather than trying to reinvent them, shooting yourself in the foot, try to... well, there are plenty of things you can make mistakes, avoid making mistakes on things. The queen, 36, is quite easy to study today. There For years, we've had static analysis tools that allow us to test your code, so I encourage all developers to try and understand these tools a little.

There are audits, and they're important. It's not just because I've suggested a lot of them that I'm saying this; it's really important. The idea is to have as many brilliant minds as possible thinking about your code. Maybe knowing what's running it will help, maybe not. In any case, an audit is expensive, it's tedious to plan, and the good audit firms are all taken, so then we try to find one that's not too bad and not too expensive.

But really, it's important. Or, failing that, a bug bounty program. I don't value the community of other hackers and developers enough. Make your code public before it's published on the blog. It's not just that it's been frozen.

Create release candidates, and you'll have to deal with Banshee. There have been quite a few times when, if there had been a real bug bounty program, maybe the The bug would have been stopped before it was exploited, you never know, it's never a guarantee, but it's a possibility. There are plenty of white hats doing fantastic work in this ecosystem, and make sure they can contact you. That is to say, if you have a smartphone, Tracts on my glasses that has parity checked, rescan, put a dense contact email address on it. Really, these are simple things, but they make all the difference.

If, for example, we at Open Zeppelin detect a vulnerability in one of our contracts, we go beyond mechanisms like that to try to detect all the contracts that live on the blockchain blocks that use our contracts. Moreover, we'll end up with a dump with dozens, hundreds, maybe thousands of hexadecimal addresses. It's sometimes complicated to know who's behind it, who 's responsible, who to contact to warn them so that the EU can take action before the bug becomes widespread. If there's a contact email, it's always much simpler. And the last thing is monitoring, that is to say...

Once the contracts are deployed, they shouldn't cross paths. The pushes crossed, saying it's over, that could have been a problem. In the prayer parties, the wallet part was a private key from the NOC, which had said it was a key not used on the nets, only used on other networks. In fact, they managed, through monitoring, to realize quite quickly that it had Licques and Kanak was underway, and they didn't manage to save the tokens on the secondary network. However, they managed to close the bridge in time so that the agreement couldn't transfer the funds back to the controller, where it would have been absolutely uncontrollable once they switched to the swaps and everything.

So, it was possible to block the fund transfers on my nets, to avoid the oil and is safe on the different child chains that depended on the bridges, and I know exactly that I had no way out at the end, but it had the possibility to redeploy a contract on the other chains. This patch is more secure than relays, it 's about giving the channel back to its users. I don't know exactly how, it's all pretty good, it's over, but it was going to end well, in the faith I had looked at, and in Nantes I haven't heard much more about these problems, so it's all over. So there you go, if you monitor quickly, the larger the attack, it's encrypted. There's a way of two meetings and an attack again, it's difficult but it can happen, so monitor your contracts, set up alerts, that's something that few decline, it can help you do.

You see, defining it yesterday, a first fight called Defender or its decentralized version that we are developing called Forte, to define alerts, like if there is a transaction with the new system that uses a lot of gas, that transfers a lot of tokens, it could be suspicious, so you can trigger notifications on our selection, Discord, possibly automatically trigger maintenance operations on the contract. So there you go, it's true, I'm going to repeat myself, but don't reinvent the wheel, really, there is There's a lot to do in this ecosystem other than rewriting their CV, or 6,721, the same if you have issues with 2.2 potential, the rice 1,36, there are plenty of building blocks that allow you to help with the risks, the mitigations, so look at the tests, do a lot of unit tests, but that's obvious, but again, what to do is check the coverage of your tests, and unlike 4 or 5 years ago, today we're starting to have fairly powerful and efficient static analysis tools, I'm thinking of these leaders that I personally use, there are others, so use them and they'll tell you if there are risks of recurrence at 360, someone, you can easily mitigate them, beware of your dependencies, the oracles and against pleasant and the shackles, these things make you reread our code, so the audit is a fairly direct type of thing, having someone reread your contract, now it's not the be-all and end-all, I think it's just one of the slices of cheese from your solution another slice it's really there to move on to the frac noted our set up bonuses bug bounties publish your code in advance I think Samsung had a boycott I think it's a million dollars in bonuses of sushi for detecting a bug in my jumps before an update to life it may be expensive but if miso had been published with the bug it's perhaps tens or hundreds of thousands of dollars that would have vanished into thin air that we didn't want that to happen to you an approach that I use quite regularly and I found very interesting is put yourself in the head of someone who is going to break contracts participate in contracts to go to competitions of two there that followed there are already quite a few game production there is a real idea me fiction of sade and try to break programs because by doing things like Dany Fayoux our smart contract that you will learn what the attack patterns are frequent and maybe behind that will allow you to find something wrong with your man in your code to know the plastics necessarily is perhaps a completely new attack that you haven't seen but at least the classics you the elytra and as I said make sure you are reachable someone from your if you are a serious project when we find your contract on the fields we must easily find your security address which we do n't know if your project is called I do n't know Kleenex Mons pass a brand sorry but if you prayed is called Kleenex you must have the security act in exploit as appear in your rate in the sport contract and prepare a response plan it is an idea to prepare the scenarios mode what happens if such a contract can be what happens if such a thing happens some shine try to think about what you could put in place both from the point of view of communication to users but also of mitigation do you have the means to solve the problems will be turned in the Stress is truly terrible, so if you can anticipate it, even just a little, it will help you a lot. I'd like to present some tools I use. Of course, this isn't an exhaustive list of all the tools to use; there are plenty I don't use that are very good, but this is my approach.

So I think the slides will be available if you want to look at them later. For the application development part, and the fairly basic and simple monitoring tools, there's Defender (which can be used to file a complaint), but there's also this powerful graph/folder if you want to set up monitoring mechanisms. I use a combination of two: MetaMask Argent and Ledger. For example, on the applications, knowing that on Tamas, I only decode or use tests, everything that's serious is to stop leaving two things unattended. It's Argent, it's Ledger, but Ozil is a very, very good option too.

And if you ever see suspicious activity on your blog, channel, or toolbar, because it's very, very powerful for trying to understand that they pass and then if you want to go even further there is the THT X for a madman which allows you to really follow this so it has very very very advanced tools for developers who know how Vienna works but it's okay it's what we use to try to understand BEAC when we try to react so there are two or three links that complement each other useful for the more technical among you here is the classic slide feel free to ask them questions on the forum feel free to test our Defender products in particular but also just use our contracts and I thank you I think that was quite quick if you have any questions feel free to ask otherwise by the UA the panel a little earlier if we have to face them and how we can manage the problems at 0 from the second many more dependent are not called for example it is there are the Gueydan casinos in the sauces as well it's a good question what they are his girlfriends Zeppelin does everything to prevent them but it happens it happens unfortunately we found one during My summer vacation was quite a hassle having to deal with this on the highway. There are two things to consider: firstly, we try everything to prevent this from happening, and we do this by implementing all the security measures I discussed, including Open Zeppelin. A mini- vulnerability bounty for Nagorno-Karabakh would be found if a bug is reported. We do everything we can to address it, propose solutions, and try to support and help the affected projects. That's why we need an email address in the contracts and Interscan.

If we find a bug in a contract, you 'll need to contact me. If there are projects that are particularly sensitive to the fact that the heroes can't reach their dependencies, another good solution for you is to join the robotics team. That is to say, if there's a bounty on Lesparre... Contracted Terium, obviously, however, in the long run, Opel Zapping and in addition to the thousands of euros or dollars Copenhagen Line offers, there are tens of millions more from the people who use us, to create an even bigger jackpot for the one who will report the zero day in a friendly way rather than exploiting it, so in a way we also encourage you to participate in security by putting in a nice jackpot and then, and then there you go, and then we specify if it happens, well, it happens, I do everything to help you, then you can also, you were on your side, games, there are insurance mechanisms and they are starting to appear on the block, so if you really have a big concern about that, also look at the insurances that you can subscribe to, thank you very much, I have a question, I would like the point of view of users, when I quickly look at the contracts, to use now all the dependencies, there is a proxy in front which means that we can, basically, the contracts that I am rereading, and how can you arbitrate between making protocols that are scalable, and ultimately, there isn't a version that has been audited, but the contract hasn't been replaced in the meantime. Again, I think that's a very good question, and that's where I think the monitoring tool comes in.

That is to say, a contract that, at the whim of the user, is very dangerous for the user, and therefore the user must be wary of that. Now, if the application wants to do this properly, it can also do so by implementing timeblock and governance mechanisms. I come from Zig Zag Shenzhen; I had implemented Habré-Dable contracts, but I made sure that a "brave man" (a term used in the original text) was forced to go through a timeblock, and the table meant that if the developer tried to create an "ogre end," it required a latency period of 3 to 4 days, during which users still had the potential means to exit the application. So that's where monitoring comes in. That is to say, As a user, or as a community, you already want to ensure that there are these kinds of safeguards in place, and that if a malicious person is ever put in place, you have the means to detect them.

Then there are many things that can be done. There are also people who have thought about how to make a great person so that they are n't visible to everyone, sometimes they're fixing a bug, but that they can be certified by an auditing firm. So there are many solutions. As such, you have to think about it, you have to be creative, you have to protect your users. And it's true that users shouldn't have to deal with contracts that can be made by people who don't know without delay.

It seems to me that it can be done. However, there you go, complete your objects. It lives in the automotive blog state via governance mechanisms with votes. That would be the solution. Thank you very much, and nothing for this great presentation that we finally had, that you agreed to make accessible to everyone.

So thank you again, it was great because I found And we'll move on to the next part, so I'd like to invite Mounir, Charles, and Adrien to join us for the panel. I'll leave you to it now; I'll leave the presentation open if you want to follow along with us. Adrien, feel free. So, before we begin, there are three prizes to be won for the three best tweets. The hashtag is #parasol.

[Laughter] At the bottom of the page, we'll be joined by the crypto stars in France, Adrien and Charles. I'll let you introduce yourselves. So, let's get started. Charles, Guillaume, and I... I'm 6-2 Ledger.

I'm a veteran with four years of experience. My background is cryptography and security. I've always worked in that field, mostly in secure hardware. I've spent ten years trying to develop hardware-based security solutions and also trying to break them. But in the end, it's a bit like...

it's a bit like two sides of the same coin. I joined [the organization] four years ago, and at the time, I created what we call the Dungeon. You've heard of it; it's our security research team. in internal security and then had the whole team, many big studios, well, after that, well, I'm Adrien, a bit like the "h" in my first name, and for a long time I've been doing smart contacts on ROMs and I'm trying to start a business in this industry where security is omnipresent, so I'm not a security expert per se, but I know traders well, sometimes it's about building something that makes sense, which creates thefts. I started a project in 2007 on the A10 vulnerability called by Apple.

I spent a year there, so I was able to do things like updates to a protocol that has several billion on it, and here I am at 2 of 6 MB and we make zeros, we don't find attestations, we want to lighten the load when we touch cryptography, so if we look at Rect 12 and A10 Fayyad, we see that there are more than 3 billion dollars that have been stolen following acts in the said vulnerability, even more, I'll go into The SSI flaw that it emerges from, lessons can we learn and what are the potential solutions to perhaps improve this in the future? Yeah, but even for a small context, it's quite... I have the impression that they always arrive in slightly different places, but always at a fairly young stage of maturity. We were able to talk about my dock, this beginning, Bassett Bitcoin, it's new, we don't know how to manage private keys, these centers and things that are going to get hacked, prayed for the DAX, we discovered smart contracts, we did whatever we wanted with the SLE, created the sense of towers, 2 of Wolf, these remnants, until something new. If we try things that I find we can differentiate all the axes and not just ten flaws or Bluemotion, its differences and new things, it's a little different, it's a bit of a stroke of innovation, so for me the solution is to clearly notice that experimental things, things a little less experimental in education, especially with whom we, how to use them in FIT, no, we should have only on The computer downstairs, that's something a bit experimental, like these new technologies.

That would already help a lot, since it's ugly. It's not much, 3 billion compared to everything we've done, so the innovation is worthwhile, but we could greatly reduce the impact of certain things. For me, it's a bit like that, it's always new things, it seems like an incredible idea. The first time it works on a small scale, then we put all the millions into it, and it can be... well, it might have...

disclaimer... and finally, raising awareness among legislators that there is a risk when these shutdowns are tested. Clearly, the environment... something ultra-ultra- new. Ten years ago, it wasn't...

I remember when I studied cryptography... we could... by the winds... we don't have all that... for the record...

in the teeth... bank cards... they said... always suck... rsa...

is tripled... it's a little bit like that... and now it's been about ten years since we started using increasingly complicated protocols in cryptography, and we're starting to do zero-knowledge... the protocols... blocks...

chains are quite complex... Okay, there's also the MPC, which is popping up in a lot of places because when I learned photography, what we learned afterwards was that things had to stand the test of time to see if these sectors, if we do n't stand the test of time, we put them to the test of production, and in the drug test, there's the bug bounty program right there, you know, Dogg, it's a lot of money that can disappear overnight. Indeed, his nose is into something very new, very innovative with protocols that are n't tested at all. It's just normal, actually, that the actions or the time... then there are lots of quite different hacks.

We can still differentiate between... I have centralized exchanges where it's more about normal security, key management, user management, but it's still pretty risky, always worried because the gain is enormous compared to... compared to the application, there are quite few applications in the whole world where you access a database and are directly at a potentially... billions, and then there's nothing that blocks are immutable, so it's quite new, like the type of application. So clearly the stakes, the security stakes are gigantic.

And there too, well, in the talk about nothing, we were talking about auditing, which is also a way to better secure contracts, but is it enough? Because there are so many protocols that have been hacked, that have been audited by big auditing companies, and is auditing sufficient, or is it one step among others? And what should actually be done to maximize or minimize attack vectors? Obviously, it's not enough, that's for sure. Well, it's necessary anyway.

To do security, you have to do the maximum, the maximum. God, who analyzes things and often says to be open source, is really important because your text makes things auditable by a large number of people, but that doesn't mean that the relevant people will audit things. What's important is that relevant people spend time auditing the implementations. Then... There are good practices, static analysis, there are plenty of things that can be implemented and are starting to be implemented, but we're still at the beginning, so we're learning a little bit the hard way.

I want to go there myself, I'll stop there. The press, by the way, but perhaps the most interesting thing is this notion of curing it there, and each layer is interesting, it says so without saying, and it has to be used in something. It's not a savior, damn it, it's part of a global strategy that makes sense, and what is the objective of my code? What's happening? I'm getting used to it, okay, I'll say it for this part with important French, it's one tool among others, and so if we differentiate the different types of AC, we'll go through the acts of consensus protocols, what we saw with the King blocks as cyclic tools were present earlier, the EAC of Bridge, we saw it recently with your banners, what are the other types, or how did you see How do you actually distinguish them?

Are there common patterns between each sector attack, each exploit, after all the acts related to cryptography, where there are catalysts around it, but still a little different? There are also complicated things in the sense that, going back to the producers, we start to be able to chain them together, and the stability after a real application thing, where we were talking earlier about Datta Khel, better Flash Lawn, which will attack an Oracle Extra, all that when you start composing the contracts with each other and the applications with each other, the stability of everything is still quite difficult to predict. The impression is that there's something interesting, trying to model these things in advance. I'm talking a bit about unity in general, I don't really know how we could do that, but there's also the stability of the protocol once things interact with each other, your interesting thing too. If there are n't two audits where there are meta-audits that will look at all the lewd blocks, it's because they'll see if, when we put everything together, it still works.

Let's leave that aside. Okay, maybe in the industry, all the Metz leaders think there's an interesting social aspect emerging around governance. Adrien said that one way to secure the skill gap in the protocol is to put it two meters behind the governance structures, and the social tax, which is known outside of blue-chip companies, is known to be a demon between its breasts, the weakest links. That's interesting; there are real attacks, also social attacks on governance, things like that. There's that to take into account too.

It's not just the code, whether it does what I want, but what are the parties behind it that can collude? That's all Proust's research, and what could be seen were things like that. That's interesting; they're attacking on the economic side, on the design of the protocol, which... is not easily auditable from a technical or static point of view, very well, and on the Ledger side, already known for being the most secure hardware device, how is your approach to securing these Diffa protocols and to maximize this security and avoid or minimize the risks of exploits? At the German dawn, we have a somewhat in-depth security approach at Ledger since the design of the dungeon, it's part of that, it's internal auditing, or we also do external audits at Stara, but overall the security model is quite simple, I'll go into it, it's long, but keys in an enclave secure the keys, they never leave the enclave, so even when you make transactions and when you don't generate addresses, the RSA keys, the interest of your enclave, which we call a secure element, is therefore a circuit that does this.

Then when we could make signatures, they are made inside the secure element, and the 3G property that was really important is when you start to, when you make a The transaction is a fairly simple thing, a sign, a bit of the Aude, which is always the same extra, like a transaction, and on Monday, as soon as I start interacting with Bismarck, contracts happen, things get a bit more complicated. If you sign the water in a completely blind way, who was n't able to understand the father above, otherwise, well, potentially you can give, you can send every time a contract, or an attacker is different, not happy, it seems that this is really a third property and which is for us the most difficult to implement. We feel where when you interact on Spa, on the right, contracts of parasols, if Justin of period by a parasol, it's an exam, in fact, potentially sending all the lands a contract, you see it, so to avoid that, we are obliged to, inside the sector, be able to park the pilot and make his human readable hockey, swapping and 2 and r against 8000 us of tea or something like that, and what 's complicated is making this parking lot make it an aluminum plant Generic bowls, well, it's not really possible, actually it's a really complex problem where you need to read before clicking, like that, which we'll do of course, through ASO, which has gone from using, indeed, you have several screens that say to wait a month, I put in the most secure option where I look at each driver, each value of the thing, but it happened to me once where I have 60 and 70, something I don't have yet, the binary is exact, but let's talk about it, people always tend to choose the easiest and least secure option, and Ledger is clearly very secure and offers a multi-SIC, it's clearly more secure, but people always tend to use the simplest one where I have it, it asks me to click, and what about the compromises to be made between the two, or those having convenience, and we also have security, clearly this way, the solution must be that which solves this, it is necessary to solve the trade-off between head shakers and be usable because if A follow-up to the solution is ultra-secure, but nobody uses it. Well, that's cool, and it's the same sector, but that doesn't solve any problems. So, we need to find where that cursor is, and indeed, it's not simple at all.

For me, these are tools; it's also the role of developers to take that into account. We have tools, the school, that are really good; that's exactly what's happening. They're a little less usable; it's the trade-off, and there are Aude jet skis, Adam Benhamma, this is false. We need to make third-party states aware that what's happening, we put Hamas, can get hacked, but that it's not a big deal because you can park it. There are methods where you have the admins on your Ledger.

We can think of intelligent Venes, the controller, the vein of that one, and not in PL at D. A point is the most super admin, and it's in your Ledger, and so at any time in Strakka mode, you will be in control of your account. But on the other hand, you can give the use of your name to an address. Meta masks and so day by day you sign 102 points off the head and if Thomas ever gets hacked late - to recover your name you don't lose your application either I think that's a great tool poet from Sète no I thought that too it's war between the two will really have to be that he has who will and speaking of that what do you think of the principle or philosophy contribute the water that reigns in the crypt so that's always it's a debate that we are most in sync I like I like this question I find it good you want to say anyway that it's wrong to steal money and on the choices of contracting but doesn't belong we agree with that she was alive that somewhere the thieves are either condemned to one day I read earlier that the guys who stole the Bitfenix four or five years ago they just got caught it was good a parenthesis but on the other hand when you Tell me, when you say something different than Cody Juno, you want to say no, we want to add two Laval players, etc. The problem is, I see two, I don't know where the line is.

You see, when we attacked, it was like doing a Flash Clown to unbalance an Oracle. How do you distinguish between a market arbitrage and an attack? Here we're already talking about something a bit complicated, and here we're just talking about attacking, something quite simple. You see, we can imagine it's much more complicated. The second problem I see is how you have the strength to applaud Collins, and what do the waters say?

No, but how do you manage to catch people and prevent them from doing it again? It's a real difficulty. I've chased, criticized, for me it's completely [ __ ]. No, but the blocks come, it's a social consensus, it's a way for humans from Agde to accept a common truth, but they're still humans behind it, and in fact, we saw it with the mess we said. Well, wait, in fact, we It's the very beginning of summer, Ramon does n't understand what we're doing, we're live from the top, we're going to rewrite history, but that doesn't mean that our Eto'o isn't trying to do things, the truth is there and we can't have corrupt systems, but bad ones, we just saw that too, it's a bit the same, the Solana hack, he got Bella August, so that's a human thing, we have an interest in rewriting history a bit because of the social consensus of the web, we will do it and the software is just a tool to express this consensus and blockchains are a better tool than anything political, from my point of view, it's a great transparent tool and everyone sees the laws and bills in an interesting way compared to a political party, but in the end it's humans who choose, in another case, for now it's the intention that should count, not necessarily the invisible constitution, in fact it's not just systems, it's imposters, there are many Social issues aren't just about computer systems; they're about human systems, and so we mustn't forget that there's this layer of implicit constitution that says what we have the right to do, if we don't, but it remains in the collective unconscious, and it's interesting.

I see, for me, there's finally something that comes to mind now to perhaps reinforce these rules. Perhaps these are the key solutions, Ross, where we have a judge who has a chain of commands who might be able to write the constitution. In fact, now we write something, but it's no longer code, it's more intention. An event occurs, and how will I interpret this event to know if I should be uncompromising regarding the limits of authority, or if I can afford to be on the other hand, when I could have said yes, indeed, I am known as, following, saying no, what they say, it doesn't work, what just happened, it didn't, and we go back to what happened, in blue, missing chains, breaking trust, also saying that The reel, ultimately, isn't so mutable, the chosen one, unless you're always studying everything. In any case, it means you can go back, which makes it notable, so for me, imitable CV, that's not true.

For now, there's no precedent on quickly being stupid, which shows there are inhabited myths, but at any moment it's beautiful and the blue blood that can do it means there are humans, and even if it's 90.60, all humans can recruit that, and that will always be the case for now, as long as it's not authors. In any case, we can say it's fertilizer compared to the current financial system where there's better confidence, quickly, maybe if in the questions by the crisis, for me it's more of a curiosity, let's say, because it's something that seems quite loring trout, what they tell me in English in terms of security, but I do n't see it done often, it's all that will be automatic analysis of what a guy does in his veil is before he does it so you have the simplest case for everyone to follow us well it's one of the most classic mistakes in encrypted form in the same months I was made like a bowl rise once I'll tell you the story if you want with 3 a bullet which is to send the crazy ones note that it contracts and so there in 99.9% of situations it's the save fun but in fact that's very predictable because it only takes two checks and the token that the Czech voice guarantor and his big tracts address is in fact it's a regularity shon and the wallet it can tell him be careful you're going to lose kills say that you're really doing it but in fact there is to my knowledge maybe there's one that he does I saw that because they do but like a meta mass by with the don't do that and I tell myself a priori a few hours of code an idea not too hard to implement and a lot of funds saved why isn't it there you never know a few hours of code It's good because if he says he's in a hurry and has lovers, but no, it's actually a great idea. Well, to be honest, we discussed it just today, Nicolas, and no, it's something we're working on.

Indeed, there are quite a few things we can remove, like what you just mentioned. That's one of them. Then there's also something Rami does that's interesting too: before you interact, he tells you she wants to be your new balance. After you sign the transaction, you see, there are also things you can add. There are plenty of things we can add.

Clearly, we need the Falcons Bill to do these things. Clearly, so many things to do, it's hard to say, but it's quite interesting and important to do. I'm thinking of thank you for the discussions. So, Magic, a question about the pickup: you said you can read Lopia Code, is that what tools are on the internet or elsewhere? which allow us as users a little bit of technique to learn to read Lopy Claude to avoid losing money and gold, it calms down, but well, after that, yes, but just if, for example, we have something like, used elsewhere in a transaction, when we see it, we don't necessarily see the function, what if the code is n't games on the terrace, Cannes, so on Ledger you manage to have tools, for my part, it's therefore the HB code, if we kill it, it's not even on Ledger, we 're on a toilet, etc.

, which doesn't suffer this possibility, what there are ways for, how did you learn to do this without it often being worth the same, you have to be a little aware of what you're trying to do and see the thing, of course, the summer in Rome is that you have the value, so there you already know, next we'll see the soils, not easy, we see it easily on Ledger but also on, put on a mask, but if there are many treatments, you ca n't, that's how I check, I just tell myself what I'm doing here, I'm doing To drop something, there's a variable. If I have a cold attack, it does all that. I'm thinking maybe it's weird, but I don't know if these divisions... you can't say precisely. Metascan is super strong on that because they read my name, the signatures, they translate it for you on lots of things, and it's really an exceptional tool.

It's not just that if it's on Earth, the scan knows that I've already sent the transaction in some way, no, after its request, so to the audio, what can it do? Otherwise, there are tools like Tender, a place where you can play the transaction in advance, well, you can simulate the transaction and see what it does. Otherwise, after that, it's a bit technical, beaten, a little... how to say... be against, assured, mix, make the break...

religious, but they hold relays, it's good, they have a strong email, it's incredible. You can play on email 900 and more, connected, your age... you can connect me, Juston RPC, I put it on them, and you go to normal AV and you play... had with its real stuff, but it's on a fork, so that's because What we're saying is that there are still efforts to be made on the Conte-Viant side, because we're not there yet for the trade-off. Thank you so much for making the Enter the Dungeon series because I think it's fantastic, keep it up!

And then you talked about the influence of the social system, of governance systems that influence protocols, and you're currently thinking about how to build governance systems that prevent manipulation via flashes, based on the protocol. So basically, systems described as political, that are resistant to the significant influx of monetary value, in quotes, to influence the protocol. And so I wanted to know if you, if you envy, if we also did, quickly get to talking about geopolitics, for example, or about these political systems that you're currently thinking about and designing, especially you who are building the policy, yeah, it's me, what years ago, no, but it's difficult to know, for me, the rulers... What we see, or have seen, with the emergence of the ADIE flaw in August, is that people vote to change something that is immutable unless this governance changes. It's the first attempt to create constitutions and enforceable laws, and so for now, it's still at a political level because it's not dense enough yet.

There are queens so easy that the flash, the onba, yes, we're going to protect ourselves against flash lawns. There's the apprentice to paladin, it's a protocol that allows you to borrow from someone without a vote. So there you lend, put, we have seven with it, it gives the right of power, but you lend directly all the right of power. So we're already trying that, it's easy, we're still in the phase of how to protect ourselves against attacks, quite easy like that, but one day, it's certain that if it has evolved, everything will be political things, and we've already seen it in a plan between heteros, but in classic terms, or bitcoin is and Rome, in the end, then the code, but well, I don't know if that It answers the question, but these are parties, and more or less interesting. In any case, it's interesting; you have to follow it.

It's more interesting than Zemmour for me, not just Zemmour and the others. And it's true that the strong yards pose major security problems because in an instant, the quantity of ascetics on the blogosphere doesn't bother you. That's what makes people instantly rich by doing a horde fork of Shane blocks, and we're rewriting history a bit, the previous one, you know, end of the classic. Historically, this often doubles things; the sum of the two of our chains rarely makes it. In the end, it sometimes makes it more than one of the chains.

You don't necessarily double it, like, and then the value of things right now is the consensus, once again a consensus of the law of the market, let's say. But yeah, you don't necessarily create more value directly, like me. It's me about politics; it's just about the job of a developer, solid tools. I meet quite a few people who really want to do solid work, and there they are, they've done the Petshop tutorials, there's Sekou Ly, and so they find themselves managing a bit of responsibility, like they'd set up a banking protocol. And how do you see the Coda profession in ten years?

This afternoon, for example, three web developers, we saw on OpenSi, who was getting signatures. I'm very curious to know in ten years if it will be a specialist profession or if many web developers will become solid developers with 3G web tools. I'm delighted to have your opinion on how this profession will be defined in a few years. I can vary, but for now there's a huge incentive for those initiated, everyone from the developers, to be fair. And so the two qualified developers are a lot like him and all the other EA developers.

For now, clearly we're going to see more and more of them. As for the future, I don't know what you think. My hope is for the Society developer segments. He's certainly in ultra-high demand, but there's a competitor, Carrier. It's the developer with zero knowledge, and for him, we're still going to double the price, I think.

In the market for things like that, and overall, I think that today, with Wave 3 mastered, we give users the honor of a Sadat chip, and in fact, we gave them access to Ledger. And we just learned about Baylet on Ledger doors; it gives access to cryptography to users, and so, thanks to access to cryptography, they will be able to enter protocols like Belote. Chains take many others six months; it's a protocol for doing proofs, things like that, and so it's Lyon who is a developer who understands how to use cryptography for applications, the first level... 'It's solid and then it can handle the air of other languages ​​too, like Wade and other languages. Maybe the languages of the past, from now on, do n't have to use them on the MUC, blocked in the future.

It's difficult to predict. When I was making an e-commerce site, I spent six or seven years with Stripe or PayPal, and in fact, I didn't even have the capacity as developers to create problems of the scale that today... I wonder to what extent tomorrow a guy will be able to do blockchain without taking responsibility because there, the developer who does it, I can't contract for a long time, by the leaders, sleep badly at night, it tips over. Yeah, then it's maybe also a question of touring, there are quite a few things to improve too. [Music] Because we went on the somewhat human side of security, it's cool, I like that we didn't forget that because sometimes it's a bit of a security vice, its features, with many technical measures, we forget that It's humans who lose their machines, and that's often where the holes in the cheese are.

So, in relation to that, it's a bit like what I call the paradox of security in a bull market. The master simply doesn't remember, it's different, it dies, and if you want, back then you could launch a new farm, and in fact, it was almost a game of how you could scare people the most. That way, the farm is very competitive, and only a few percent of the yields are good. So, there's no audit, others won't pay back enough. A cup, this top 4, that was what really had offices with things, and it worked at one point.

In fact, I think behind that, you see the trend that basically, the more optimistic the market is, let's say it like that, so bullish, and the less people care about security. It's actually quite paradoxical, people are very attentive, let's say, even worse. That's the feeling I have. So, in a ball, that's where there's little air, that's where we see that all The dumbest attacks also happen, the famous macaques who lose their 6000d, things like that. So basically, how do you resolve this tension, lacquered by default, spontaneously, when everything is rosy, people are crazy about security.

It's a bit like the insurance agent who was the guy who annoys people when you talk to him because they don't even want to listen. Ideally, he would like to be in a world where security doesn't exist. When I get the impression that's how people are, what interests me a lot because it's really my approach is quite great. It's about trying to understand the traders in which you are and having humility, a conscience. I don't know how to grow up, but we want to know if what you're getting into is right, and it's a bit like that.

As Fast is completely right, it's never the first thing that will crack, it's the first one to fall apart. There are people who have become rich, Susan Riesch, secondly, everyone will go all in, and there's nothing that's free. There aren't two, there aren't There's no such thing as a freebie, ever, so finally this is the soldier hack, it's perfect, it's both his chicken and Bree playing, there you go, applauded, finally there we go, we're going all in, hence the market fund, it's because there are competitors before the spa and to prepare to go there as a lateral justice on that didn't actually have everything, yes, but it's only a representation of the lands, it can be done, it only takes noise, surface hackathon collateral, your points, finally you have to be aware when you use it as TF1, when it's as a lateral of the ether on Solana, you have to know that you have plenty of links, you have plenty of potential hack sources and they could do it, it's going to get hacked, there the post is a bridge that does n't say it's him but it could be the lands, I think, I don't know if that answers the question, it's to be, yeah, that's it and then no, and as you said at the end there are often humans, there are humans who develop but also humans who often used, that's where the Finding the weakest link in this whole thing was a real challenge to educate and try to get past the emotional aspect and explain things properly because paradise is still quite new, well, we're kind of into it, it's been interesting to us for a while, but for someone who doesn't know anything about it, we try to explain the Nextera blocker, well, what is it? Explain the idea that having crypto means having a private key that allows you to prove that you are indeed the honor of Coyle, who are on the Shane McJutra blog, it's complicated, how do you explain it to someone who isn't familiar with it, who might seem like a techie, and so on. So there's a real challenge in education, making things as simple as possible, and indeed security is also about this part, because it's another question, it's a bit of a follow-up to everything that's been said, we have quite a few protocols that are done only once, the code, often the first area of ​​attack is the code, then humans, we see now with the multiplication of The governance of politics, which is interfering a little with blockchain, is what you see happening later than the first, the first security job itself.

It will be about securing against human error, saying, "Finally, there you go," because the goal is adoption. And later, we'll see—well, for now we don't really see our parents, or older people on the other chain—but the goal is that later on, there will be more and more of them. And how do you see the shift later from securing the code to securing the human? After that, I don't really see a change. It's a bit like what I was just saying.

It's clearly making security solutions, wallets, interactions, and protocols more foolproof. Clearly, it's already a big step that allows us to improve security and will also improve adoption for everyone. It's about education. For now, it's for tech. Seville and the CLAE on the fact that it's not for everyone, yeah, well, addressed to what penetration is called, to be one day, and it and all the tools will be better, but I have a question about the NRT because, well, it happened this year, really, to everyone, I'm everyone, I even have someone in my coworking space who told me, "I bought my NFTs," I said, "You didn't put it there, I'm not leaving it on the site, that's fine," and he had no awareness of what they were doing, actually.

And what I think will be interesting is being able to see your student, then towards Saclay Ledger, do you all think of something around NSTs that could really help people, everyone? I think we had passed the concrete without talking about NSTs, and it's a failure, we'll have a lot of friends, however, that's all, what does everyone answer, yes, well, we that Tedjar, we recently released NST support on a field, Jean, regarding the request, what happened today is not experimental, so you can see everything in the long term, you can see all NF Terium and you can send and receive them, obviously very securely with the display trace on the device. Then there's also a real question: how do you organize them physically, etc.? There are interesting things to think about around that; it's a subject we're also working on, but well, can't we say much more? Not only does it revisit the previous question, but how do you educate people, how do you share with everyone the ideas we have about arcs, is it a private key, things like that?

And I agree, this year is good, a lot of things happen, but something very interesting about clarity is that it's the first time in nine years that there are people, a bit like me, and these lives, who buy something because it's beautiful, because they themselves are worth helping, I don't know what, but When they go on multiple platforms and see their NFT, or when they can go on Twitter and check their profile picture with the 9 tokens, it's actually the first time in your life you go on competing platforms, like when you go on Harry Potter, Open6, or Foundation—these three competing platforms—and you see your NFC on all three platforms. You've never seen that before. On Facebook, you have Facebook friends, or on Twitter, your Twitter friends—everything fails. It's a good thing, that's a really interesting thing. It also forces people to learn because they need to make that effort so they can control their NFC.

But I'm holding an economic value. It's true that's kind of what you mean. It's one of the first things you see is the interoperability of Motion, which makes something like Blottiere exist beyond where we use the objects they say about Motion and the guests. You see, you can, and Remy, imagine seeing how you I was saying about a rival on OpenCL in 1 in a video game is just because it exists beyond the world, well, beyond the world of video games and the open world, sufficient success in the chat, and that's something to focus on, interoperability is really a big value, and I hope they'll complain because, in fact, there, they feel it, they feel that it's up to them, they'll wait until it's them now, so for ten years we've been saying yes, onions, dates, pretty bad social media, Facebook on a thing, but they do n't understand that it means giving your, your place, and really, yes, me, noting the victories everywhere I use it, and things, it's a great educator of what it means to have your data because we've never been in the know, it's actually, thank you. [Applause] [Music]