G

PM @coinbase

[G @gregdamx]

I miss when writing was valuable.

A thread or post that clearly took lots of time to research and make sure made sense was hugely beneficial for the reader and writer.

Now if you read something, it is generally -EV for you.

[G @gregdamx]

Conversations like the “prd is dead” are difficult to have because of very different optimizations and tradeoff appetites across the spectrum.

Similar to discussing trading/investing with different time frames and risk management.

[G @gregdamx]

In email notification, especially with links/CTAs, is antiquated and should be obsolete unless there is a regulatory requirement specifically for email.

Especially in age of agents connected to inboxes, the only trustworthy mechanism is in app notification.

[G @gregdamx]

@dampedspring Most people accessing the models are selecting a front end first which dictates the model. This will obviously change over time.

[G @gregdamx]

@apriori0x So you’re saying ultrasound money was correct

[G @gregdamx]

Excited to roll this one out.

Native support for wide variety of tokenized and traditional cash equivalents for collateral and lending in our cross margin model.

[Coinbase Institutional 🛡️ @CoinbaseInsto]

Get more out of your capital.

Cross-Margin lets institutions go long or short across spot & futures, netting risk in one account. 100+ assets as collateral.

10 new cash equivalents now eligible alongside USD and USDC: SGD, CAD, CHF, EUR, AUD, GBP, EURC, OUSD, PYUSD and USDT.

[G @gregdamx]

Product, Eng, and Design becoming the same role is largely overstated.

Lines are blurring to an extent, yes.

None are “cooked” (for now) despite the initial hyperbolic reaction to every new launch.

[Coinbase Markets 🛡️ @CoinbaseMarkets]

The CFTC has approved the launch of Coinbase Clearing LLC, our own USDC-native clearinghouse.

Built for 24/7 settlement with USDC collateral, Coinbase Clearing completes our full stack of regulated derivatives infrastructure.

[G @gregdamx]

@sjzhang_ Yea this also makes the assumption that switching costs miraculously disappear because of AI

[G @gregdamx]

Good ideas was always the bottleneck

[G @gregdamx]

@apriori0x @frachtisvc Huge congrats. You will crush this.

[apriori @apriori0x]

"If you're a builder and your goal is to be a trench dev, Robinhood chain is where you go.

The money is good. The casino is hot."

@chang_defi, on Credible Commitments

[G @gregdamx]

@apriori0x Banger

[apriori @apriori0x]

Ethereum cannot shorten slots today. Slot times are fixed at 12 seconds. Quick Slots (EIP-8198) fixes that.

The business case is clear: faster confirmations, fresher onchain prices, and faster finality. Pay the one time cost, reduce slot times to 10 seconds then ratchet lower.

With Glamsterdam scaling improvements coming online, maybe it's time to reorient the performance canon towards shorter slots.

In this episode of Credible Commitments @barnabemonnot makes the case for Quick Slots inclusion in Hegotá. Full episode links below.

[G @gregdamx]

@mignano Music is math but there’s a deeply personal, human connection to it that I don’t think AI can easily replace.

So Consumers are unlikely to directly pay for AI produced music.

I guess it monetizes through slop Spotify playlists and TikTok/YouTube at scale.

[apriori @apriori0x]

I do not expect Ethereum to be useful as sanctuary technology if the price of ETH were to eventually follow that of alt L1s.

1. Sanctuary technology requires deep liquidity. If you can't move size it's not that valuable, even if you expect billions of small users.
2. Sanctuary tech requires security. The sanctuary security budget is denominated in ETH. If ETH declines materially in value then the chain can be attacked easily.

L2s move congestion off main-net and increase trust assumptions (post 4844) whilst reducing the demand for L1 blockspace consumption both from L2s themselves (tx call_data) and users: developers and end users.

L2s can help with distribution so long as there is interop between main-net assets and the L2, i.e. ETH. However L2s are mostly distributing stablecoins, BTC, and native assets. ETH itself is only ~19% of the total value secured on L2s (l2beat), which is ~2.4% of total ETH supply. To put that in context, Coinbase and Binance hold in sum about ~ 21.5% of ETH supply. The retail onramp to ETH is through centralized exchanges who now also own the chains. They own the user from end to end. There is no incentive to prioritize ETH. Robinhood will be no different.

At best L2s are verifiable servers for businesses and at worst multi-sigs taking advantage of regulatory arbitrage for a quick buck. It's fine if the broader community wants to keep DA fees negligible. Robinhood and Base are reacting to incentives. It's significantly easier to lean into regulatory arbitrage than it is to build a fully decentralized Layer 1 both technically and from a GTM perspective.

But note that isn't going to accrue any value to ETH. You need size trading on L1. MEV burn + base Fee burn + tapered issuance burn *could* in aggregate outpace issuance in a future. This is true only if L1 orderflow grows by about an order of magnitude. If size trades on mainnet this is possible.

Size does not necessarily need sub-second latency, it needs best execution and censorship resistance with strong finality guarantees. Accelerating RWA issuance brings quality collateral, propAMMs make tight quotes possible, and deep credit markets facilitate capital efficient margin.

Sanctuary technology is not all that useful to the world if it's only for cypherpunks and idealists. Ethereum only becomes sanctuary technology if and only if Ethereum is the venue for the deepest spot markets in crypto and captures value from that orderflow.

[G @gregdamx]

For the first time in a very long time it feels like there are too many interesting coins to get exposure too

[G @gregdamx]

Crypto is the greatest MMORPG ever created

[G @gregdamx]

Ok new idea. Let’s pay gas fees with AI and implement a version of 1559 that buys nvda with the gas fees, puts in the treasury, and burns the AI.

Roll this out across the entire EVM