LobsterDAO 🦞

10 Jul 2025 01:43

So fuck Lloyd's, to my point lol

LobsterDAO 🦞

10 Jul 2025 01:39

That's why all the projects mentioned above can scale to only $5b at best, and in the $10s of millions at the realistic end

LobsterDAO 🦞

10 Jul 2025 01:37

ugh, can we do it without Lloyd's, though? Let's make a a Lloydefi against the suit-and-tie people

LobsterDAO 🦞

10 Jul 2025 01:34

I truly think it’s an education issue and I’m proposing with a few people that maybe we do a defi day in lloyds soon. Get some quants in and some actuaries in and do some work

LobsterDAO 🦞

10 Jul 2025 01:33

The craziest property of protocol insurance is you can’t just hack morpho and suddenly be like “crime seasons boyzzz” and then hit euler aave and kamino

It’s almost perfectly orthogonal and has really nice diversification

LobsterDAO 🦞

10 Jul 2025 01:31

Problem is most protocols make very little money

LobsterDAO 🦞

10 Jul 2025 01:30

4% of protocol fees?

LobsterDAO 🦞

10 Jul 2025 01:29

ALL HACKS EVER in a single year and you still make money with a 4% premium. That’s including all the proper bottom feeder protocols not even just top tier

LobsterDAO 🦞

10 Jul 2025 01:23

But you are saying that GMX could buy the 5% chance of getting hacked to offset the potential lose? In theory, I guess that could work, but there aren't liquid markets for that on any prediction market, especially not for the thousands of various DeFi projects that are out there holding $1m+

LobsterDAO 🦞

10 Jul 2025 01:17

If I put $1M of ETH into aave and borrow 400k USDC to go do stables stuff I'm paying 5% to stretch goal get 8% + eth yield 2%.

The prediction market for aave hack in 1 year would be like $0.95 no; so I buy yes at $0.05 and now my yield is treasury rates 😂

LobsterDAO 🦞

10 Jul 2025 01:13

Well, I just mean it has to scale

LobsterDAO 🦞

10 Jul 2025 01:12

I also don't think that we need to reinvent the wheel. We could take the tradition model of insurance/reinsurance, decentralize the underwriting, pay high yield for high risk, and then we all have to price in the risk from the top of the funnel

LobsterDAO 🦞

10 Jul 2025 01:09

I'd be very interested to learn more! Can't really understand how a slashing-based model would work in a case like GMX, but sounds cool

LobsterDAO 🦞

10 Jul 2025 00:59

I don't think we are talking about the same type of "insurance" here. Yes, stETH is securing the network/base layer, but anything that is built on that foundation isn't covered by Ethereum stakers. I am talking about protocols that build on Ethereum/any chain. They should go through audits regularly, yes, but GMX's L1 was attacked (right?), which was audited many times in the last 3-4 years.

What I am talking about is the "oh shit" insurance that every smart protocol pays into, with a $1m deductible or something, yet backed by Web3 money instead of trad$ like Lloyd's.

Tbh, yes, it still sucks that we have these types of DeFi hacks in 2025, and it looks amateurish for institutions (look at L2 outflows). But, it's really not on the same scale as a CEX hack, which is now in the billions. CEXs have insurances in centralized ways. DEXs should have insurance in decentralized ways, inshallah

LobsterDAO 🦞

10 Jul 2025 00:47

But why can't there be a skating/DeFi solution? If a $1b pool were somehow set up with hundreds of DEXs paying in monthly fees, we spread the risk and share the reward, just like every pool works

LobsterDAO 🦞

10 Jul 2025 01:42

The issue for Lloyds is that in DeFi we haven’t shown that we actually have demand for this product. The demand needs to be in the hundreds of millions to actually make reinsurers give a shit. These guys underwrite big digits so doing 50m here and there on customised policies is literally a waste of their time if they can’t scale it

I think secondly they have compliance issues frankly as well paying out to randos

LobsterDAO 🦞

10 Jul 2025 01:38

Yeah, exactly. Degens aren't interested in an easy 4% lol

LobsterDAO 🦞

10 Jul 2025 01:37

My main concern isn’t the premium tbh - I’m fairly confident we can get good premiums as they underwriters aren’t stupid. My concern is that we can’t get enough underwriters to give a shit to cover the demand

LobsterDAO 🦞

10 Jul 2025 01:33

Yes, this is exactly what I was getting at! Again, I have no project to shill here, I don't think anyone is doing it at scale. But the economics make sense, and to disrupt the insurance (and really don't sleep on the reinsurance) industry is the most Satoshi thing we have left to do as an industry

LobsterDAO 🦞

10 Jul 2025 01:31

no that’s like an absurd extrapolation 4% is all hacks ever in a single year. I’m putting a statistical limit on what anyone could ever pay

If someone pays >4% they just don’t wanna insure you is the TLDR

LobsterDAO 🦞

10 Jul 2025 01:31

So you’d ask users to pay a 4% fee on each deposit or withdraw?

LobsterDAO 🦞

10 Jul 2025 01:30

What is a 4% premium?

LobsterDAO 🦞

10 Jul 2025 01:29

If people price the risk the stats don’t agree tbh. Out of the major defi protocols the biggest hacks have been largely bridges and even then it’s like 600m

https://www.theblock.co/data/decentralized-finance/exploits

Bizarrely they include bybit as a “defi” hack which makes no sense IMO

If you insured all the TVL in DeFi and paid out every single hack of all time (5.4bn) in a single year you would still break even on an insurance premium of 4%

LobsterDAO 🦞

10 Jul 2025 01:22

Send a link for prediction market of aave hack

LobsterDAO 🦞

10 Jul 2025 01:15

Insurance is the most well understood market design in essentially all economics.

You can have direct peer to peer market. A pooled model. A government model.

Pooled models have always inevitably ended up government models with maybe the exception of car insurance but even then disasters tend to get subsidized heavily by gov when there's an act of god. Peer to peer never scales.

Hacks are essentially the defi equivalent of acts of God.

In realistic terms, prediction markets are the closest bet.

"Will GMX (any version or v1 or V2 w/e) get exploited (and admit it was an exploit) for >$10M before Jan 1 2026 UTC".

But you don't see those markets do much b/c baseline assumption is no and spreads are wide relative to yield and protocol revenue

LobsterDAO 🦞

10 Jul 2025 01:13

its defi, what makes you say that? the underwriting is decentralized

LobsterDAO 🦞

10 Jul 2025 01:10

coming back to this…

what if instead of spraying bullshit incentives we paid the premium

What if because of our business line we can get cheap premiums as well.

Would fully insured DeFi be attractive at no extra cost? Or does that somehow end up centralised and anti-defi ethos?

LobsterDAO 🦞

10 Jul 2025 01:03

The link to staked eth is that I heard Eigenlayer was exploring a slashing based model for protocol insurance. Maybe someone knows more

LobsterDAO 🦞

10 Jul 2025 00:49

I mean technically that’s what staked ETH is right? It’s insurance against protocol downtime such that everyone who holds ETH sees deflation from the slashing event which is effectively an insurance payout

LobsterDAO 🦞

10 Jul 2025 00:45

The model I think is probably best is where nexus sit as the excess insurance and cover the first $1-5m loss to cheapen premiums but I think you can’t really get the tower into scale unless you look at lloyds market