Intents

essentials team built out two pieces recently, declarative execution environment (essentialvm) & constraint driven language (pint)

pint contracts are composed of a persistent storage block & a series of predicates which constrain how state transitions happen. state access is made explicit in predicates, identifier (counter) may constrain existing state, identifier w single-quote suffix (counter') may constrain next state.

vm itself is a 64 bit stack machine w two execution stages, one for loading storage variables into a tx buffer, the other for constraining values in the tx buffer. isa is minimal, w a handful of logical, arithmetic, elliptic curve, sha, storage, and control flow ops

note that state is not explicitly assigned to a single value, state is arbitrarily provided by solvers, predicates only constrain the bounds of what solvers submit

obligatory counter attached:

do you think that they'll have success convincing Solidity developers to move to a new language and VM?

I'm bearish on any "intents" play that doesn't leverage the existing stack

i'm bearish solidity & evm

i think devs that want to stay there will be just fine, in the same way java devs are still just fine. but ppl wanting to leverage smth more flexible that's built with the knowledge we have about blockchain execution environments that we didn't then will absolutely try out new vm's and dsl's imo

Don’t you feel there’s too much critical mass around the evm? I get that it has serious flaws, but nothing competes well with a first mover advantage like this 

I see JavaScript / TS as the better analogy. Garbage origins, slow maturation, but massively pervasive and sticky given tooling lock-in

no, imo the reason moreso that there are no other serious blockchain vm arches competing rn

solanavm is a little more restrictive but probs the most mature, move vm is over restrictive and fragmented, algorand vm is clunky and is algorand, fuelvm is cool but founders had a meltdown and now it's not clear whether they're ever going to mainnet, stylus isn't live & the value prop being rust-to-wasm falls flat when it takes a ton of configuring and tinkering to get rust to generate a reasonable wasm output that works for resource constrained env's like blockchains, aztec vm isn't live and actively changing, essential just released their code this week and is also not live and is actively changing

evm has momentum, but everyone i know that has used evm for years doesn't like using it anymore. i try to avoid it altogether these days. eof is a step in the right direction, but the language ecosystem is also p lackluster

maybe "competing" is too 1-dimensional

aztec vm will be best for privacy since afaik there's basically no competitor here

essential vm i think has potential bc the primary use case for blockchains atm is idiosyncratic trading and its externalities and intents are a natural iteration for on-chain trading where changes are globally propagated more efficiently and lp's receive less toxic order flow, so a swap intent vm alone would be a clean abstraction, but generalizing to all state transition opens unique angles we haven't seen before; idk what this looks like in full yet, just like we had no idea gas tokens or data contracts would emerge from evm

but i don't think any of the parallelized or llvm compatible vm's meaningfully "compete" in any way bc they all suffer the same drawbacks. linear scaling is not sufficient for global scale & while 64 bit words run faster than 256, these are cryptography machines at the end of the day and most fields we're using still rely on >250 bits

I think the absence of a credible competitor is even more telling that the incumbent has cemented its position

It's kind of like the "it just hasn't been done right" argument. Is that the case, or is it just a tar pit, and we're stuck with something shitty that has good penetration?

Do you think there are any compelling competitors?