And for our next talk, I want to bring on Andre from Tenderly to discuss all things developer tools. So, please give him a big round of applause. Thank you. So I know this this whole event is about keeping things non-technical. So we'll do our best to not actually go too much in depth for the the developer side but it is important to understand how things have changed especially in the last two years um about just Ethereum tooling and and the rest of the developer ecosystem.
So my kind of first thing that I want to ask you is kind of set some context for for the audience which is tell us about what Tenderly is and what do you all do and um how we got started.
Yeah definitely. So uh we've been in the space for the for the past eight years. So we've been doing developer tooling and developer infrastructure for now we can proudly say the biggest projects in the space. We sort of grew up with them. we saw um all the problems that they had and you know we hopefully try to solve as many uh as many as we found.
So at this point what we do is we basically cover uh most of the development infrastructure needs when it comes to a bit more advanced a bit more sophisticated development setups. Uh something that we feel is uh desperately needed as we see a slight ramp up in these hiccups and um and missteps when it comes to people losing money and stuff. We had one today or yesterday.
There's there's a lot of incidents that u we unfortunately still keep seeing. Um Tenny obviously started by supporting Ethereum debugging and that was kind of the primary primary chain but obviously in the last few years we've seen so many other L2s, L1's just EVM compatible chains come up. So I'd love to kind of understand from your perspective how have develop how has development changed um now in 2026 and how has multi-chain world sort of affected how people code?
Yeah, that's a that's a good question. uh we've sort of come full circle uh on the topic of of multi-chain because when we were starting uh way back in in 2018 uh there was a a huge wave of these private blockchains that are going to be used by all the banks which we saw didn't really happen as fast as we as we wanted it to. Uh but uh yeah for us uh the way that we see it is um long gone are the days where someone is deploying or choosing which network they want to deploy their DAP and this is sort of their home. This is their base camp and they don't really look uh beyond that anymore. Uh these days anyone who is successful actually has a solid road map of networks uh that they're expanding to.
They want to cover uh you know as many developers, as many end users as possible and try to spread their surface area across uh multiple uh multiple networks and that brings with itself a whole new set of development issues. uh you know we we've for the longest time we've been able to provide a pretty clean pretty encapsulated uh development experience on just a single network because you have all the state uh you have all the uh all the information in a single place and this is sort of what you work with both us as a tooling company uh and also uh the developers they have sort of less moving pieces to worry about but these days I mean on top of managing uh multiple mainetss on top of managing uh the state across all of these you have so many other infrastructure components u that make things a lot more complicated for you know fast and secure development and this is something that we that we saw is is one of the bigger issues where some hiccups
what are some examples of these components
yeah like bridges are the most obvious one uh one of the one of the ways in which we looked at development infrastructure was giving you the ability to uh fork mainet and then run simulations on top of a single network. We saw that that doesn't really cut it these days for more complex uh applications because you actually need two versions of two different mainets and you have uh bridges between them and you need to test the logic if things leave and land on different chains uh accurately. uh the moment we the moment we pushed that out as a feature that was one of the more interesting um more interesting um surprising uh levels of excitement that we've seen. So that on top of indexing, on top of uh you know staying alerted and up to date on all of these different chains and what state are they in chains maybe go down more a bit more often than they should except for the main one which is which is what's important. But uh yeah, in all of these little hiccups, lags, uh inconsistencies, it really makes it difficult to provide a clean consistent experience that realistically we need to be building uh for the next wave of of end users.
That makes a lot of sense. Um what is the current best version of a developer stack look like for for somebody coding and starting building smart contracts for the first time? And I'll kind of like in a way give a an intentional seven years ago perspective which is you had truffle and you had your MetaMask wallet and that's it. Yeah. Uh yeah, we've gone a long way in terms of what we uh can do and what we need to do if we want to sort of stay as safe and as protected as possible.
So yeah, as you mentioned back in the day, uh you know, you write some solidity, uh you use use truffle, maybe you use the ganache local node, uh you deploy it on the chain and and all is good. Um maybe you have a front end, maybe you don't even have a front end there, but effectively you're just interacting with uh with the blockchain. uh these days uh if we're talking uh like a production grade serious organization like uh likes of a or you know maker these types of teams you really need to think about first what's your surface area of development so the UIs have gotten a lot more complex no longer you can you can no longer rely just on smart contract state so you need to have some indexing solutions you need to think about okay do we introduce account abstraction as a as prime first citiz first class concept in my application we have these types of services am I am I bridging how many networks I'm deploying to uh so a lot more elements to consider so I would say on the general tooling side sort of aside from tenderly definitely we see a lot more foundry as a as a
seems to be the standard these days
framework uh framework of choice um and then sort of whatever is the is the flavor for for each of these uh building blocks. And then sort of how we thought about supporting this entire ecosystem is giving you basically a pretty clean uh continuous integration, continuous development uh environment uh for you to make sure that there are no regressions as you go from version one to version two. Um deploying upgrades of protocols become is a lot more touchy than just going out with the first first version. So making sure that uh you know all of your bases are covered, all of the edge case scenarios u are factored in is something that's fairly important. Um and then the other thing uh that's that's very uh useful for people to move faster is have these types of staging environments where you integrate your you know backend components your UI uh your u your you know data services uh whatever else you rely on for running the protocol successfully.
Um for those also uh having a dedicated environments that are a bit more sophisticated than a test net uh is something that we've seen people switch to a lot more uh frequently um these days. That makes a lot of sense and I I feel like um just to add on top of it, every time we do a hackathon, we learn something about hey this is what people are using now and a new library ends up being part of the core toolkit. So, um, I'm excited to see what I learn tomorrow. And, uh, it's usually always just fun because you keep on seeing great improvements. And I feel like we're going to get to this in a few minutes, but skills are going to be a big part of this from this weekend.
And people just want to delegate a lot of that Ethereum context into uh, your uh, codeex and cloud code. So, uh, I can't wait to see what that looks like. Um, one of the interesting things about Tenerly to me is that you have a good mix of obviously developers who are doing this thing for fun and just getting started, big DeFi projects, but also just traditional companies in finance and fintech. And the thing I want to understand is do you kind of notice any different patterns between those groups? Like do they care about something more than the other?
Are there any sort of trends on how they're using the platform versus how you want them to use? What are kind of things that stand out? Yeah, this is something that we're always trying to keep a keep a close u ear to because obviously we want to be building in the direction where the the industry is going and I think those three categories definitely uh have good u representations in terms of what type of usage they have on chain. So uh the people who are new to the space or the new projects uh we see them doing a lot more uh smart contract integration versus smart contract development. I think it's
they just want to talk to contracts and see things but not build it.
So yeah, they're they're interacting and they're building on top of existing protocols. They're effectively the maybe it's not exactly the first wave but one of the first waves of people successfully utilizing these DeFi Legos and then making a successful outcome out of that. Whether that's you know a DAP of its own or it is um you know some some of these services that are bridging the gap between web two and web three
kind of like a dashboard type behavior where it's just read only from the blockchain and then show make sense of the information on the client side.
Yes. Because I mean I think the industry rec is recognizing that you know yeah technically we have all the data right there for for the taking but effectively understanding it transforming it and doing something smart with it is a job on its own. So I think this is definitely a trend that we that we see people switching more from developing new you know hardcore protocols providing true innovation um in the in the space and are leaning more in terms of okay how do I piece all of this together so that you know UX is a bit better so that the final experience is a bit smoother and I think this is a great place for the the new people to to come into then sort of the pros the the OGs they are working on the next versions of their thing so a launch launched their V4 um during this week in Khan. So uh these guys are heavily heavily into pushing the frontier which we need those uh we need those two. So you know kudos to them.
Uh they are sort of holding the fort for the for the old guard. Um and then the the more institutional the more traditional players. I think the main difference between them and and everyone else is uh risk because I think we as an industry are a lot more uh ready to take on unnecessary amounts of risk where when the term cowboys
yeah maybe maybe um yeah but I mean when you're an existing institution the main thing uh that you're afraid of is you know compromising the thing you have built so far which I think is the main bottleneck for them to like truly on forward onto the thinking and onto the sort of mindset uh that has been pushing the space for for a while. Since you kind of mentioned risk, um obviously there's a lot of that strongly tied to just the DeFi as a category. And the thing I'd love to understand is like how do you think about simplifying or reducing risk for your customers because you're effectively giving them the tools to understand what could go wrong with their codebase and um and how much of that goes into the product thinking and how do you think about that as a company? Yeah, I think that's that's been one of the things the primary things that has been on our mind recently. So as we saw this uh you know split of uh where the energy where the new energy is going into um we saw that yeah effectively it's all about managing risk at the end of the day.
I mean whether you have a tolerance for it or not you still don't want your protocol to be exploited. Uh so how we how we thought about this operational risk is I mean there is there is so much more work and so much more ongoing active um action that needs to happen once you have deployed your protocol like you build you build the app and they come okay now they're here you have to keep the protocol running smoothly you have to manage your treasury you have to you know optimize uh optimize uh for for the yield there are a lot actions that you need to do with the blockchain now that uh there is a lot of a lot of stake on the line literally. Um so what we try to do is we try to expand from okay we help you with developing uh smart contracts we can also do a lot in uh helping you operate them more safely too. So our simulation technology is actually apart from it being very overengineered, it's very useful to be slotted into different contexts that you might be exposed to uh operating a successful a successful project. Uh so I mean people might know the uhnosis safe um UI and sort of the the way to simulate transactions before sending them uh through the through the whole song and dance.
uh this sort of
it kind of helps catch at least most of the cases.
Yeah.
Yeah.
And you you would simulate you would simulate the the output before uh before you know blind signing again blind signing being one of the downfalls of the of the hack yesterday. Um, so there's a lot of these little tools and little um improvements to the workflow that uh that should happen if you want to really bulletproof your uh operational hygiene and uh make sure that yeah you are making all the moves on chain intentionally uh and not leaving anything really to the up to chance.
I mean to add more complexity to the problem people have been coding a lot with agents now. So how does that change the game for building smart contracts and also people using tenderly like is that do you consider that as a good thing? Is that not good yet but it can be and right now it's like adding more damage. How do we think about money and autogenerated code?
Yeah I mean whether it's good or bad it's inevitable. So this is one of the things that is obviously improving the workflow of a lot of people you know and you can use anything responsibly or irresponsibly. So I I rather think about the responsible use case. So from our perspective, we try to see okay, how can we enable these um agentic workflows to give them the extra confidence, give them the extra safety net of just not fully blasting uh fully blasting all of your secrets, states, assets uh into into oblivion. So one of the interesting ways that we thought about it is, you know, code is only half of the picture.
The other half is sort of the data. What's the liquidity? What's the actual state on chain? What is the actual mainet uh representation of your protocol or the protocol that you're integrating with in this context? Um, and uh, yeah, I think for us, uh, our role there, as as far as I see it, is to provide the sandbox for the AI to sort of figure things out on their own in a risk-free parallel universe while still relying on all the, you know, protocols that you need to work with, with all the data that you need to to work with.
Is that a way of saying like the the fork test net that you have lets you kind of just create a new playground?
Yeah, exactly. So like you can get you can get the best of all worlds uh test net environment by just forking the forking the mainet state and even letting the AI you know tweak it and override state and set some balances and do some checks. I mean AI can do a lot of things well like don't get me wrong. So why not give it the full power of uh give it the full power of um
let it go crazy. Yeah.
Yeah. And then at the end of the day you still uh have executed this you know series one 10 a million of transactions in an isolated environment that if needed you can go in and you can double check and make sure that the end result really is something that you know you negotiated with the with the AI instead of you know the the slipping a couple of things underneath the covers u for you to just sign away
and that that only gets trickier every time because there's so many ways to uh kind of prompt inject the way out of this whole thing.
Do do you recommend something to your customers on how they should think about AI and smart contracts or um do you kind of see any common mistakes that everybody's making? Just you've kind of touched on that but expand a little bit more.
Yeah, I mean we we released an MCP as as everyone else did. Uh
so this is MCP for talking to onchain or
Yes. So effectively it's you know two functions that I think are fundamental to the development workflow whether that's you know human-driven or AIdriven. Uh so one is to to be able to debug an a problem a transaction and uncover sort of a lot of data from there. I think what's interesting this is this is one area where AI does it better than a human. So our debugger behind the scenes has maybe for each transaction 50 megabytes of data that we can dig up and sort of reverse engineer and make into something that's readable.
But I cannot give you 50 megabytes of data unless like you have a free weekend to to debug just a single transaction. But the AI for example doesn't really have a problem with that. You know it can work a bit more over time and and use all of the context that we can provide to achieve a better result. And then the other one it's the sort of sandbox use case. Uh so
just check this is still with the MCP.
Yes.
Amazing.
Um and then the other one the the other use case is just you know give the AI the ability to test out the the workflows or the you know trading strategies or whatever uh it wants uh in a parallel universe. And then at the end you get a simple link. you can open it, explore it, look at whatever uh whatever it did before you sort of decide, okay, this is what I want to put into my production. This is what I want to put into my uh you know regular useful. This is like why would this not simplify?
That was the idea.
Yeah. Yeah. Occasionally products make sense. Yes. Uh this is awesome.
Um, are you getting time to build stuff for fun these days or are you fully in a way just in the management phase of running tenderly? Do you get to build stuff for fun? I I now maybe have an excuse to to do more coding. I think it's I mean I think it's important for a development uh for a development company uh to understand its users and you know AI being this new addition to the toolkit and like a really transformational tool in your arsenal. I think it is it's important to understand how you work with it.
What are the pain points? What what's the friction uh so that we can you know feed that back into the product and and make sure that we stay you know at the top of of providing development tooling to you know AI enabled organizations of the future.
Nice. Um, and my last kind of question to you is kind of other than using Tenderly, what advice would you give to developers that are kind of just learning how to build smart contracts in this space and coming into this ecosystem this year to uh to be a good developer in in the EVM world? I think especially um I think especially in this sort of AI first world I think an important thing to learn and uh and and understand more about is sort of what's the big picture of the full system and what is the full security surface area that you need to think about it I mean I can start rambling so many different aspects of a of like a bulletproof DAP that you would need to think about but yeah so I think it all it is all in the spirit of making the project more secure, making it more reliable because ultimately we need to have that locked in as fast as we can if we want to be, you know, taken seriously by the by the rest of the world.
Amazing. Thank you so much for this amazing conversation and a lot of good advice. Please give him a big round of applause.
Automatic transcript — names and jargon may be misspelled.