gakonst

I don't think crypto has a demand problem. Everyone would be using crypto if the tech was actually good.

The demand is basically infinite and it's next door.

No scale is enough, and horizontal scaling always outpaces vertical scaling at global scale.

Biggest Q is what will happen where. Can all on chain finance fit on a single chain?

Different answers depending on who you ask.

But on the off chance it's not all low latency / high demand for colo finance, and it's stables, mortgages etc -- will it all fit?

Guess we will find out!

The Flashbots MEV Auction as a scaling technology framing is great and something they have been saying for a long time.

It doesn't tell you spam is bad if you can build the capacity -- but it does make you think on your levers to scale.

It gives you one more lever to scale via offchain systems and get the market efficiency you want without doing diminishing returns systems optimizations (bad ROI on time eventually) or consensus concessions which may compromise your open source-ness (bc MEV validators abuse the ordering as much as possible to make money), or your network topology and geographic decentralization.

Similar to how scaling pre training gets harder and harder from here, and you look for other places to scale which are earlier in their sigmoid eg like post training RL.

The Flashbots MEV Auction as a scaling technology framing is great and something they have been saying for a long time.

It doesn't tell you spam is bad if you can build the capacity -- but it does make you think on your levers to scale.

It gives you one more lever to scale via offchain systems and get the market efficiency you want without doing diminishing returns systems optimizations which may compromise your network topology and geographic decentralization.

Similar to how scaling pre training gets harder and harder from here, and you look for other places to scale which are earlier in their sigmoid eg like post training RL.

Who has the best express / 1click checkout for stables?

Who has the best express / 1click checkout for stablecoins?

What's the best way to do on-chain art NFTs in Ethereum in 2025?

SVG or otherwise, I don't mind, whatever looks best.

The framing of what you're doing matters as much as what you're doing.

Example: Ethereum statelessness and consensus upgrades.

Defense: simplicity and decentralization.

Offense: gigagas/s and trillion dollar security for the long run.

Same tech, different takeaway.

The framing of what you're doing matters as much as what you're doing.

Example: Ethereum statelessness and consensus upgrades.

Defense: implicity and decentralization.

Offense: gigagas/s and trillion dollar security for the long run.

Same tech, different takeaway.

The framing of what you're doing matters as much as what you're doing.

Example: Ethereum statelessness and consensus.

Defense: implicity and decentralization.

Offense: gigagas/s and trillion dollar security for the long run.

Same tech, different takeaway.

IMO the Reth state transition on tip of a large chainstate (eg Ethereum, Base) is the best benchmark for real world zero knowledge proof performance.

We already have that demonstrated with multiple ZKVMs, today, powered by a stateless MPT Reth executor program such as Succinct's RSP on SP1.

Dont have Reth in your ZKVM? Reach out to me, @r_krasiuk or @kevaundray.

IMO the Reth state transition on tip of a large chainstate (eg Ethereum, Base) is the best benchmark for real world zero knowledge proof performance.

We already have that demonstrated with multiple ZKVMs, today, powered by a stateless MPT Reth executor program.

Dont have Reth in your ZKVM? Reach out to me, @r_krasiuk or @kevaundray.

IMO the Reth state transition on tip of a large chainstate (eg Ethereum, Base) is the best benchmark for real world zero knowledge performance.

We already have that demonstrated with multiple ZKVMs, today, powered by a stateless MPT Reth executor program.

Dont have Reth in your ZKVM? Reach out to me, @r_krasiuk or @kevaundray.

Ethereum's ZK roadmap can finally decouple gas limits from validation costs, and a lot more.

Enshrining multiple proof systems makes no sense though from a shipping speed perspective.

Keep it offchain, let the user decide whatever combination they want to use.

Ethereum's ZK roadmap can finally decouple gas limits from validation costs, and a lot more.

Enshrining multiple proof systems makes no sense though from a shipping speed perspective.

Keep it offchain, let the user decide whatever combination they want to use for redundancy.

I am excited about the future of crypto with stablecoins, but I am more excited by crypto as a humanitarian technology to assist with instability, global financial access and peace in this world.

What are the best resources on using PoS to finalize PoW chains (eg standard heaviest chain rule or GHOST-style) in a super bandwidth efficient way?

The Foundry roadmap for 2025 and beyond is incredible. We'll be able to store trillions in EVM smart contracts. Use Foundry to build for scale and for the long run.

From a Frontiers applicant: "Foundry literally changed my life and I use it for everything. I don’t even know typescript because before crypto I used Python for my previous industry. So it’s so nice to use Solidity for everything and I use it for every single project I can."

Passkeys will change the world.

Passwords will be a thing of the past.

Passkeys will be on every device and every passkey will have a stablecoin account.

Compatibility issues across browsers / OSes will be fixed top down.

We are building the best SDK for Passkeys.

With Porto we'll be able to soon offer to every application to "queue" their transactions and order them however they want.

This is very powerful for DeFi, obviously. We will expose that in Typescript via a great API a la how our paymaster-equivalent will work.

I have some ideas, but I am wondering:

If you're building say a DEX, what is your preferred way to program that you want transactions to be ordered in such and such way (e.g. cancels top of block)?

One trivial way is to define a map that says "if this address and this function selector is called, give this priority N, and if some other thing is called, assign priority N-1", and then you sort in buckets of descending priority and sort by fee within the bucket.

Is that a good enough API? Or can we do a lot better?

Maybe Solana frens working on ASS (this needs a rebrand btw) can advise.