Rezo🛡₿RRR @rezosh

@KobeissiLetter it'll be interesting when DeFi finally offers different types of products that traditional investors are looking for. x.com/rezosh/status/…

Rezo🛡₿RRR @rezosh

a hour ago

FIRE for everyone in trading and liquid asset management! A Vitalik proposal I'd advise you NOT to scroll past: index-tracking assets and synthetics built on options, instead of debt and liquidations. Worth stopping for: ethresear.ch/t/building-ind… The mechanics are simple...

0 0 1 101 0

FIRE for everyone in trading and liquid asset management! A Vitalik proposal I'd advise you NOT to scroll past: index-tracking assets and synthetics built on options, instead of debt and liquidations. Worth stopping for: ethresear.ch/t/building-ind… The mechanics are simple... nothing new for anyone who's been around options, but the mechanics transform in the crypto world. - Split 1 ETH into two halves: P = the cash half, holds value up to the strike; N = all the upside above it. - Together they always equal exactly 1 ETH. Want stable exposure = hold P and roll into a new strike when price gets close, that's it... and now the chain that makes it all matter. Liquidation is impossible... and not because risk parameters are carefully tuned, but by construction: the two halves sum to exactly what was deposited at any price. How do we get around the Oracle? Exactly like this. Because if there are no liquidations, no real-time oracle is needed. You need a price once, at expiry, and real-time oracles are the weakest point of the entire on-chain stack: making them safe is nearly impossible, and whole protocols have burned on that attack vector. I wrote after the Polymarket episode: the best on-chain primitives are still being built. This is exactly that. A primitive that's fine with a slow, verifiable oracle. Perps took the liquidity and the attention, but they stand on the same fragile stack: debt, liquidation, fast oracle. A liquidation cascade in every stress event isn't a glitch. It is, unfortunately, a property of the foundation. Implementations are already on testnets, and Vitalik himself says "it's happening" and insists on formal verification before mainnet. After the Orchard bug in Zcash, that's no longer just caution, it's a lesson learned and a necessity. We ran formal-verification research once, and I really regret that back then, early in my crypto journey, I didn't have the budget, and we paused the research. Anyways... Either way, this still isn't about "ETH up tomorrow," it's about whether on-chain exposure becomes acceptable to the capital that's been sitting out... cos it didn't want to depend on fragile oracles or catch someone else's cascades. Options are a huge part of derivatives markets in TradFi. So if, as Vitalik writes, it's finally taking off... that's great for everyone. Serious money will come as soon as we remove from the foundation the thing that kept breaking.

vitalik.eth @VitalikButerin

7 hours ago

Looks like the options thing is happening already! See also: various people thinking through and building different versions of the idea in the thread: ethresear.ch/t/building-ind… Though I do strongly urge that if any of these get on mainnet quickly, we formally verify it first. I

240 133 1K 125K 313

@ryancarson how much did that question cost?

@bridgemindai OpenAI can afford to lose more money in the short-to-mid term cos they actually have more compute, not just runway. They can afford to handle heavy usage.

@coinbureau meanwhile Bitcoin still following the four-year cycle. x.com/rezosh/status/…

Rezo🛡₿RRR @rezosh

a day ago

We are still in the four-year cycle. People thought institutions, ETFs, corporate treasuries, and debt would break it or make it completely different this time. They said the old cycle is over. That’s what everyone says every cycle. The cycle isn’t an external pattern that got

0 0 7 2K 0

@gmoneyNFT idk if 12 is enough

Fable 5's pricing shows the real economics of frontier AI. A short window inside paid plans until June 22, then pure usage at $10/$50 per million tokens. It's a structural bet is that the marginal intelligence of a Mythos-class model only justifies its marginal cost on a narrow slice of high-stakes workflows. SWE-Bench at 80.3% is strong. The question isn't whether it's strong, but on which workflows that delta converts into money. And the answer won't come from Anthropic absorbing losses, it'll come from power users' spend. For the first time, the market gets an honest measurement of how big that slice is. Also, the "permanent underclass" takes feel premature. The solution isn't cheaper frontier access, it's routing. Open-source and local models already handle most production flows for a fraction of the cost... you call the expensive model only on the steps where it actually moves the result. That's how we run it, and that's how everyone will. Now the market reveals it: either the slice of high-ROI workflows is as big as the hype assumes, or most production work was always a job for specialized, cheaper systems.

@rohanpaul_ai easier to calculate returns on SaaS than AI. x.com/rezosh/status/…

Rezo🛡₿RRR @rezosh

a day ago

Fable 5's pricing shows the real economics of frontier AI. A short window inside paid plans until June 22, then pure usage at $10/$50 per million tokens. It's a structural bet is that the marginal intelligence of a Mythos-class model only justifies its marginal cost on a narrow

1 0 6 3K 2

@stacy_muur @tokenterminal @manifoldxyz @DeFiSaver it would be better if things like the Humanity incident didn't keep happening. x.com/rezosh/status/…

Rezo🛡₿RRR @rezosh

a day ago

The Humanity Protocol (in two parts) hack wasn't a sophisticated smart contract exploit, it was private keys on a laptop, and by then, the project had already spent weeks building exactly the kind of shady-shitty-dirty exit-liquidity structure that usually ends badly. Part One:

1 1 7 1K 0

@gmoneyNFT Fable 5 on a pay-per-use is interesting... either the slice of high-ROI workflows is as big as the hype assumes, or most production work was always a job for specialized, cheaper systems. x.com/rezosh/status/…

Rezo🛡₿RRR @rezosh

a day ago

Fable 5's pricing shows the real economics of frontier AI. A short window inside paid plans until June 22, then pure usage at $10/$50 per million tokens. It's a structural bet is that the marginal intelligence of a Mythos-class model only justifies its marginal cost on a narrow

1 0 6 3K 2

I think there are also more people (myself included) who care about model stability, not just intelligence. As these labs have proven, intelligence-as-a-service will only improve over time, but stability esp in actual workflows, is extremely important. Can't have distillations and nerfed performance.

@milesdeutscher @aiedge_ ultimate final indicator is ROI, does the expense really produce meaningful profitable output? x.com/rezosh/status/…

Rezo🛡₿RRR @rezosh

a day ago

Fable 5's pricing shows the real economics of frontier AI. A short window inside paid plans until June 22, then pure usage at $10/$50 per million tokens. It's a structural bet is that the marginal intelligence of a Mythos-class model only justifies its marginal cost on a narrow

1 0 6 3K 2

@CryptoMichNL plus there are still shady events like with what happened to Humanity Protocol. x.com/rezosh/status/…

Rezo🛡₿RRR @rezosh

a day ago

The Humanity Protocol (in two parts) hack wasn't a sophisticated smart contract exploit, it was private keys on a laptop, and by then, the project had already spent weeks building exactly the kind of shady-shitty-dirty exit-liquidity structure that usually ends badly. Part One:

1 1 7 1K 0

@unusual_whales saw somewhere that even welders who were gifted stock options will also become millionaires. great example of how early conviction (or sometimes luck) in early projects can become very lucrative. SpaceX struggled for the first few years.

@TheMoonCarl would be better if we didn't have things like the recent Humanity exploit. x.com/rezosh/status/…

Rezo🛡₿RRR @rezosh

a day ago

The Humanity Protocol (in two parts) hack wasn't a sophisticated smart contract exploit, it was private keys on a laptop, and by then, the project had already spent weeks building exactly the kind of shady-shitty-dirty exit-liquidity structure that usually ends badly. Part One:

1 1 7 1K 0

@cryptorover It's not just the market... exploits keep happening, and on top of that, shady liquidity structures. x.com/rezosh/status/…

Rezo🛡₿RRR @rezosh

a day ago

The Humanity Protocol (in two parts) hack wasn't a sophisticated smart contract exploit, it was private keys on a laptop, and by then, the project had already spent weeks building exactly the kind of shady-shitty-dirty exit-liquidity structure that usually ends badly. Part One:

1 1 7 1K 0

@CryptoMichNL despite new market structures, the mechanics still stay. we're still in a four-year cycle. x.com/rezosh/status/…

Rezo🛡₿RRR @rezosh

a day ago

We are still in the four-year cycle. People thought institutions, ETFs, corporate treasuries, and debt would break it or make it completely different this time. They said the old cycle is over. That’s what everyone says every cycle. The cycle isn’t an external pattern that got

0 0 7 2K 0

@Cointelegraph New layers but Bitcoin still follows its four-year cycle. x.com/rezosh/status/…

Rezo🛡₿RRR @rezosh

a day ago

We are still in the four-year cycle. People thought institutions, ETFs, corporate treasuries, and debt would break it or make it completely different this time. They said the old cycle is over. That’s what everyone says every cycle. The cycle isn’t an external pattern that got

0 0 7 2K 0

We are still in the four-year cycle. People thought institutions, ETFs, corporate treasuries, and debt would break it or make it completely different this time. They said the old cycle is over. That’s what everyone says every cycle. The cycle isn’t an external pattern that got layered on later. It comes from Bitcoin’s own rules, the fixed issuance schedule, and the holding incentives built into the protocol itself. New market structures don’t rewrite those mechanics. Yes, this cycle has new layers, but the four-year cycle itself is baked into Bitcoin... the issuance schedule and the holding behavior didn’t disappear just because bigger players showed up. It didn’t go away... we’re still in it.

@craigweiss Fable 5 on pay-per-usage makes it easier for companies to see whether or not it really produces high-ROI workflows or if specialized/cheaper systems can do the job. x.com/rezosh/status/…

Rezo🛡₿RRR @rezosh

a day ago

Fable 5's pricing shows the real economics of frontier AI. A short window inside paid plans until June 22, then pure usage at $10/$50 per million tokens. It's a structural bet is that the marginal intelligence of a Mythos-class model only justifies its marginal cost on a narrow

1 0 6 3K 2