Introducing Sovereign SDK -- the first rollup framework that executes 99% of transactions under 10 ms, fast enough to bring price discovery fully onchain. Now even solo devs can build efficient, global financial apps. Production-ready today. Go build the next Hyperliquid.

I support Hydration 100%. Blockchains are great for DeFi. Bitcoin started it in 2008 with magic internet money. The double spend problem is inherently financial. Polkadot will benefit from helping apps like Hydration and the DeFi vertical as much as possible.

I just read this article by FlashBots. Cool analysis But I believe they overstate their central thesis: that all excess blockspace will be filled with MEV-induced spam, and this will in turn raise fees for regular users, pricing them out. It looks more to me like scaling does work, this is only a problem on chains with a ton of economic value, and the use-cases they describe as being priced out would easily trigger the scaling limits themselves. I'm just posting some thoughts I had here - it's not a takedown, and I don't even have skin in the game, but I think the mental model they present is unnecessarily alarmist & wanted to clarify for whoever's interested. The key thing is that the amount of demand for spam is a function of the MEV of the network: how much _money_ there is up for grabs with brute-force searches. Their thesis goes something like this: 1. MEV searchers will fill all unused blockspace to perform atomic searches for swaps. 2. These search transactions will bid the price per transaction up 3. This is the real limit on scaling blockchains: no matter how much space you add, the remainder will always be filled by junk, and this will make fees prohibitively expensive for end users. I am with them up to point 3. This is for a couple reasons. First: MEV searchers in the aggregate won't spend much more on on-chain searching than they manage to extract as a reward. Second: The total amount that is spent on on-chain searching is spread out over _all_ searching transactions. Therefore increasing the blockspace capacity of the network, while holding MEV constant, will still reduce transaction fees proportionately. So it doesn't look to me like the central thesis holds: if you increase the capacity of the network, all else equal, you still get correspondingly lower fees for users. And I believe that, at least in the current environment, that technological scaling can still outstrip the extractable value side of the equation. Note that fees in a blockspace market are effectively zero until the whole block is being used. Is there room for improvement? Absolutely. The kind of dumb searching Bert describes is definitely wasteful. Is this inefficient for users? Not if fees for the operations they want to perform are cheap enough. In an extreme example, if there's $1 million per second up for grabs in MEV (that is, if on-chain brute force MEV were a $31.5 Trillion industry annually), with 20,000 TPS total, the marginal demand per transaction would be pushed up to a floor price of $50/tx. But with clear scaling routes to much higher TPS, and with real MEV quite a few orders of magnitude lower, the real effect on a chain with 40k TPS and a "modest" $10B/yr ($317/s) amount of available MEV would be a transaction floor price of about $0.01. I believe that a chain with $10B/yr up for grabs on brute-force MEV would have to have a lot of demand overall. Especially because a lot of MEV isn't even captured by these kinds of dumb brute-force approaches. Picking the low-hanging fruit of brute-force MEV would already get you to a tiny fraction of a cent of user impact. I'll concede here that this probably would price out a lot of logic for things like micropayments and social networks. Though the implied blockspace demand for apps like this at scale would probably already outstrip 40k TPS and start driving the price up independently. In either case, I think the solution is to use a rollup or separate system that doesn't have the same amount of MEV at stake so there's less incentive for spam. So it's clear that MEV will have some influence on scaled chains, but this doesn't seem nearly as dire as presented. The interventions that FlashBots propose will of course help. Anything that reduces the "MEV available to dumb atomic searching" will.

Very excited to see teams like Bullet using our state database NOMT (Nearly Optimal Merkle Tree) to build great products across the entire blockchain ecosystem. This is one of the reasons we've been excited to partner with @cemozer_ at Sovereign - they are building a userbase that can use our tech to have an impact. NOMT is a neutral piece of technology. It will be used to make rollups or L1s on Ethereum, Polkadot, Solana, Cardano, or any other chain an order of magnitude more scalable than they are now just through good engineering - and without compromising on decentralization. Just the beginning :)

Thrum has signed a service contract with Parity Technologies to integrate our merkle trie database NOMT into Polkadot-SDK. Target date is EOY 2025.

Last month we landed a host of new optimizations for NOMT to save disk space and improve performance. Here's a benchmark with 1B accounts, with transfers to freshly created accounts. Previous TPS: 19k New TPS: 34k Previous space usage: >1TB Current space usage: 140GB

I have to say this is a whole new low and only shows desperation. You guys obviously can't compete on technology so you resort to poisoning our SEO? Masterful gambit honestly. But it won't save you.

We've just released NOMT v1.0-preview. The new state database we've been working on for a year is just about ready for production. On hardware that costs less than $1200 we have achieved >50,000TPS with 100M accounts and >25k TPS with 1B accounts. Full post in next tweet.