Jack Ratkovich

UNREAL: 🇺🇸 Brian Armstrong is blocking the CLARITY Act to protect a 34% commission fee. Not for crypto freedom. Not for retail investors. For his own bottom line. Every time he says "this bill isn't good for crypto." He means "this bill isn't good for Coinbase revenue." People lie. Numbers don't. Remember that next time he plays the hero.

Why Morpho Isn't a Great Fit for RWAs. Morpho is designed around the deployment of immutable markets. A market is defined by its loan token, collateral token, price oracle, and risk parameters; the interest rate model, LTV, and liquidation discount (inferred from the LTV). This is great from a lender's perspective. When you deploy capital to a market, you know exactly what you're lending against and that the terms won't change. But the fundamental problem with this design is that it assumes risk is static, when risk is very much dynamic. What do I mean by this? As a lender, when you deploy capital to a market you have a holistic view of the current world state. You might deploy to a market offering a 91.5% LTV (which infers a 2.62% liquidation discount) where the collateral has plenty of on-chain liquidity available for liquidators. But what happens when that on-chain liquidity starts to disappear? The risk profile of the market has changed, but the parameters haven't. Morpho's solution is to deploy a new market with updated risk parameters that better capture the shift in the external environment. In practice, it's not that simple. If liquidity in the original market is currently being borrowed, borrowers are unlikely to voluntarily migrate their positions to a new market with a lower LTV and a higher liquidation discount. This becomes even more precarious with RWAs, which carry a fundamentally different risk profile from spot tokens. Liquidators go from taking on price risk to taking on duration risk. Let's look at a concrete example. Take the Anemoy Tokenized Apollo Diversified Credit Fund (ACRDX) and assume we deploy a market with an 86.5% LTV, that equates to roughly a 4.22% liquidation discount. ACRDX has quarterly liquidity, so to keep things simple, assume redemptions only occur on the 1st of every quarter. If a position is liquidated on day 1 of a new cycle, the 4.22% discount is probably sufficient to cover the duration risk and opportunity cost of waiting 90 days for the liquidator's redemption to settle. But if the position is liquidated on day 89, the same 4.22% is far more punitive as the liquidator only has to bear one day of duration risk for the same reward. This creates a perverse incentive. Liquidators are encouraged to wait as long as possible before seizing a position, since the longer they delay, the better their risk-adjusted return on the liquidation becomes. By design, this heightens the probability of bad debt accumulating in the market. Risk modelling in a lending market needs to be dynamic. It needs to respond to the specific characteristics of the collateral it's modelling, not treat every asset class as if it carries the same static risk profile. What's good for spot tokens isn't necessarily good for RWA's.

I've been thinking about this all morning and I can't shake it. Neuralink just put chips in 21 people's brains. These people are moving cursors, typing, playing games, all with their thoughts.. by Just thinking. AI agents are coding apps, booking flights, writing contracts, managing entire businesses. Everyone talks about these separately. Nobody's thinking about what happens when they merge. Because right now the only thing standing between you and an AI that does everything… is typing. That's it. You still gotta type the prompt. Neuralink removes that. You wake up. You think "schedule my week." Done. You think "build me that app." It's shipping by lunch. You think "find me a house in Lisbon under 300k." Three options before your coffee's ready. Just a thought and it's done. And before you say "that's sci-fi" both technologies already exist. Separately. The chip works. The agents work. The only thing missing is plugging one into the other. The smartphone made the internet portable. AI agents made expertise free. Neuralink makes the interface disappear. I don't think people realize how close this is.

The real concern from AI isn't that it will take your job, it's that it will take the power your job conveyed.