Trust Isn't Given. It's Earned.

In this episode of Saffron Files, Justin Roberti and Dingo discuss the fallout from the Kelp DAO exploit, the importance of rebuilding trust through stronger security, and explore the risks and possibilities of AI-powered programmable money.

Transcript

Justin Roberti (00:00)
Hello, everyone. This is Justin Roberti. Welcome again to Saffron Files. I have with me my co-host, Dingo.

Dingo, welcome. How are you doing this evening?

Dingo:
Hey, hey, hey. It’s great to be here. I’m doing pretty well myself, man.

Justin:
Good. It’s good to see you.

It’s quite an interesting time in finance all over the world, and in the world of digital finance as well.

Of course, everyone has been talking about Kelp DAO, so let’s go ahead and start there.

I’m sure it’s something you were talking about earlier today, Dingo, but I think a lot of people were.

It was hit for approximately $290 million.

Are you ready to turn yourself in? Have you abused Kelp DAO enough?

Dingo:
Yeah. It has been absolutely wild.

We’ve had a string of hacks and exploits hitting us almost nonstop, like reverberating punches over the past few weeks.

It has been a brutal time.

I think it really highlights the importance of security and making sure that we’re taking all the necessary precautions, not only as developers but also as DeFi users, to insulate ourselves as much as possible from potential attack vectors.

You look at crypto Twitter and see overwhelmingly negative sentiment, justifiably so.

It has been a very difficult few weeks for us.

Justin:
Right.

What’s interesting is that we’re still very vulnerable to these kinds of exploits, and hackers are only going to become smarter. That’s the only direction they’re ever going to move in.

What do you think this does to the way people feel about DeFi?

I can’t claim to be an exact historian of this, although I’ve paid a lot of attention.

But if you remember the period after FTX and Sam Bankman-Fried, I think people cooled toward DeFi in a lot of ways.

Scandals like that have major reverberations outside the Web3 community, but I think they also have major reverberations within our community.

Do you remember that?

I feel like there was a cooling period toward DeFi in general, and that it started heating up again last year and this year.

Dingo:
Yeah, and I think that’s part and parcel of the renewed interest we’ve seen in DeFi.

People can tell which way the wind is blowing.

But as much as that attracts people who are well-intentioned and want to build or participate in the space, it also attracts a lot of malicious actors.

That’s something we have to remain very cognizant of.

You address that by being thorough and by improving security.

I don’t want to gloss over what happened here.

At the end of the day, this was almost $300 million. That money came from real people.

And it isn’t only about the people who were directly affected.

Even if you participate in DeFi and weren’t personally impacted, you’re still affected by the reverberating effects.

When something like this happens at that scale, it doesn’t only affect one protocol, one layer or one group of people. It has a chilling effect across the entire space.

Make no mistake: this was a systemic failure.

There are real people behind those monitors and screens who bore that cost, and that’s something we can’t lose sight of.

This is another example of why we’re so focused on security.

It’s why we’ve gone through 10 audits, and why we don’t ship anything until we’re genuinely confident in it.

We’re acutely aware, as anyone building in this space should be, that the downside of getting something wrong or rushing it out prematurely is very real.

This is still new and fresh, and we’re still examining the reverberating effects.

We don’t have all the information yet, but I’m sure more will come out over the following days.

At the end of the day, DeFi has enormous promise.

But you can’t take trust for granted. It isn’t given out freely. It has to be earned.

Every time something like this happens, all of us bear some of the impact.

We all have to work that much harder to rebuild trust.

The way you rebuild and earn that trust is by doing the rather unglamorous work of making sure your contracts are as thoroughly tested and reviewed as they can possibly be before anyone else’s funds ever touch them.

I’ve talked about this a little on Telegram, but we’re currently testing internally on Arbitrum using our own funds.

That’s my USDC on the variable side right now, which I’ve been posting about in the Telegram.

I hate that this happened, but it did happen.

Now we have to look toward the future and determine what lessons we can absorb from it.

How can we take what happened and use it to become more thorough and more secure?

That’s my spiel.

Justin:
I completely agree.

We’ll have to see how that develops. We’re absolutely going to need that kind of work in order to grow.

I wanted to talk a little about prediction markets today on Saffron Files.

I’m coming out with an article this week, folks, which I’ll definitely post a link to.

My argument is essentially that prediction markets have what I’m calling a Heisenberg problem.

The Heisenberg uncertainty principle states that you cannot simultaneously measure both the position and momentum of an electron with perfect precision because measuring one affects the other.

The idea has also been expanded more broadly to represent uncertainty, which is why Walter White uses “Heisenberg” in Breaking Bad.

We’re holding up prediction markets as a way to measure the truth, and that seems worthwhile.

For instance, prediction markets correctly anticipated the outcome of the last presidential election when many official polls did not.

I can understand why people trust the democracy of that.

But it is also very easy to influence the outcome, sometimes without even intending to.

Then there are situations where people very much intend to influence it.

I heard a clip earlier today. I don’t want to say who I think it was in case I’m wrong, but it was someone fairly prominent.

He looked up the bets being made about a speech he was about to give during a very public exchange.

When he saw which words people were betting he would say—including words such as “Ethereum” and “Bitcoin”—he said, “Okay, we’re going to add these words to the speech.”

I think he was mostly doing it for laughs.

But there are other instances where people do it for money, and some where it isn’t funny at all.

There was a story involving a war correspondent who reported on whether an Iranian missile had struck Israeli soil on March 10.

That was essentially the subject of a $14 million bet on Polymarket.

He began receiving enthusiastic anonymous messages from people all over the internet.

At one point, a group claimed it had approximately $7 million riding on the outcome.

The messages escalated into threats on his life, demanding that he change his reporting.

Specifically, they wanted him to say that it wasn’t part of an Iranian missile that had hit the ground, but part of an Israeli interceptor missile, which wouldn’t have counted toward the bet.

But look at what Tony Soprano would do for $50,000.

The right person might do almost anything for $7 million.

You’re creating an incredibly powerful incentive to change the indicator itself.

It creates a problem in gambling that doesn’t really exist in many other forms of betting.

What do you think, Dingo?

It’s a compelling idea to not only bet on outcomes, but to feel like you’re participating in them.

I think people often bet on the outcome they want to see happen.

I love the idea that the zeitgeist can coalesce around ideas, and I love prediction markets as an indicator.

But with all those financial incentives, how do you stop participants—including someone giving a speech—from tilting the scales themselves, even just for the fun of it, let alone when they have a direct financial incentive?

Dingo:
It really makes you wonder what the regulatory landscape will look like for this in the future.

I don’t see this trend dying down anytime soon.

It’s also interesting to look at it from the crypto perspective.

It seems as though the degeneracy of some parts of the industry has infected the entire world.

I’m excited to read your article.

You bring up a lot of interesting points about how prediction markets may no longer simply be betting on outcomes.

Taken to their natural extreme, they may not even be measuring beliefs anymore.

They’re effectively incentivizing people to do certain things.

Money can be a powerful force for incentivizing a wide range of behavior.

If you have enough capital or influence behind something, you stop asking, “What is going to happen?” and start asking, “How can I make this happen?”

You can have something funny, such as a bet about whether someone will streak during the Super Bowl, and someone could try to make that happen.

But when you gamify and incentivize things to this degree, it can almost become a race to the bottom for humanity’s baser instincts.

It’s fascinating to observe from a distance.

I believe one of the news networks now has a partnership with one of the prediction-market platforms and uses its numbers to discuss whether Thomas Massie is favored to retain his seat.

At that point, prediction markets have fully penetrated the mainstream.

It’s fascinating to observe and think about the positive applications, while also acknowledging some of the more negative consequences.

It feels like a massive human experiment where, theoretically, you could incentivize almost anyone to do almost anything under the right set of circumstances.

Justin:
Yeah, for sure.

You could imagine incentives influencing public speakers.

Someone presenting at the Emmys, for example, could see a bet that they’re going to say a particular thing.

I can absolutely see that influencing people.

It’s kind of amusing for now, and I can understand why it has become so popular.

Crypto has embraced it, even though prediction markets feel more like a cousin of what we do in the crypto space.

They bring in so much activity and excitement.

It’s a use case that people are enthusiastic about.

Do you think prediction markets really belong within Web3?

I suppose they do because of that two-way interaction.

Dingo:
Yeah, they definitely have that democratized feeling to them.

It’s hard to separate that from much of the ethos of Web3.

Justin:
I agree.

I’m generally bullish on interesting use cases, folks, and I like prediction markets for that reason.

I think we’re going to get better at designing them.

If nothing else, you can see the value in prioritizing markets involving things that are beyond direct human control.

For example, you might bet on the first day the temperature exceeds 85 degrees.

I’m not saying every market should be like that.

But you can see why people might prefer a market where nobody can interfere with the result.

It will be interesting to see where it goes.

Its hold on pop culture is remarkable right now.

Dingo:
It reminds me of 2021.

We talk about DeFi summer, but it also reminds me of the memecoin craze, when every celebrity was participating and everyone was launching their own memecoin.

There’s a similar zeitgeist around prediction markets.

Justin:
Definitely.

Prediction markets have also received a level of respect that memecoins perhaps could have deserved, and sometimes still could.

To me, memecoins are simply another form of altcoin at this point.

But people have to show prediction markets some respect because they can be worthwhile indicators of public sentiment.

Dingo:
Yeah.

I think someone is eventually going to figure out how to tie those things together in a more robust way than what we’ve seen before.

There’s a massive incentive to do that, and I’m very curious to see how it plays out.

Justin:
Definitely. We’ll keep watching that.

The other thing happening now is that people are essentially creating smart contracts where both the money and the AI are built into the contract.

Imagine a programmable contract that effectively becomes programmable money.

It isn’t only making rules-based decisions. It can make dynamic decisions, with AI driving transactions in the background.

This is another advancement in what agentic AI will be able to do.

How long will it be before human traders can no longer compete without robotic assistance?

What do you think, Dingo?

Dingo:
I think many of the largest firms are already using AI to some degree.

Even before people used the term “AI” the way we do now, they had algorithmic trading systems.

We may already be there.

Imagine you discovered a super-secret, foolproof system tomorrow.

You wouldn’t announce it on Twitter or publicly share it.

It could be years before someone reveals that an AI had been making incredible trades for them and coming out ahead every time.

What interests me is what happens when it breaks bad, to return to the Breaking Bad reference.

What happens when an AI makes financial decisions and everything comes crumbling down?

Who assumes the blame?

You have the builders who wrote the model, the protocol that deployed it and the user who opted in.

But once it’s running—and potentially modifying itself on the fly—who is really still in the loop?

When there is a catastrophic failure, who assumes responsibility?

That is the interesting question to me.

I’m sure we’ll see it play out as this becomes more mainstream.

We talk about AI a lot, Justin, and I love it.

We live in exciting times because all of this is being figured out in real time.

But I think we have some very interesting headlines ahead of us once this technology starts breaking in unexpected ways.

Justin:
Exactly.

What do you do when an AI agent robs you and that agent is a second- or third-generation descendant of another AI agent?

In other words, it was created by a different AI.

Who do you blame?

Do you try to trace it back to the original maker, only to find that it was created by a hacked version of ChatGPT or something similar?

What’s interesting about programmable money and programmable contracts is that you could have an AI agent that isn’t only managing your portfolio.

It could manage your income and pay your bills.

When one investment starts losing money, it could move the funds automatically into a lower-risk asset.

That’s pretty cool.

But you’re placing a great deal of trust in that machine.

Dingo:
I’m generally a glass-half-full person.

There are things to be cautious or even apprehensive about, but I hope the innovation and exciting new possibilities on the horizon outweigh the potential negative consequences.

Justin:
Honestly, I think interacting with AI is important.

If you work in technology, I think it’s the responsible thing to do.

You’re in a much better position if you understand it than if you don’t.

Anyway, it’s good talking with you, Dingo.

Shall we do a Saffron Corner?

Dingo:
I like that.

We’re continuing to develop more educational materials right now.

As part of that educational work, we launched an ETH/USDC vault on Arbitrum.

I posted about it in the Telegram chat.

It has been really interesting watching it ebb and flow.

We started this two-month vault during the Strait of Hormuz drama.

This may age like milk because I don’t think we’re finished with that situation yet.

It has been fascinating to watch the projected P&L fluctuate so dramatically.

You have people quickly moving out of Ethereum and into stablecoins.

Then it looks like things are improving, so they move back into Ethereum from stablecoins.

That has been very good for the variable side of the vault.

There are a lot of lessons we’re extrapolating from that.

I think it’s going to become a valuable educational piece showing how volatility and uncertainty—even when caused by negative news—can still benefit participants in the vaults.

We’ll have more to share soon.

We still have more work to do in building those resources, but once we’re ready, we’re going to hit the ground running.

Justin:
Definitely.

We’re going to keep communicating throughout the entire process, folks.

Thank you, Dingo, for joining me again this week.

Dingo:
Thank you so much, Justin.

I really appreciate talking with you. It’s always a pleasure.

Justin:
All right, and thank you to everyone who listened.

It has been great talking with you, and we’ll be back next week with another show.

We’ll talk to you very soon.

Take care, everyone.

Peace.

Dingo:
All right. Take it easy, guys.