samlaf

Sheel: So if I send $1,000 from the US to my cousin in India via stablecoins, what would she get in INR? Brian: She'd get ₹83,000 worth. And the best part — it's basically free. Sheel: Basically free, or free? Brian: Free. Stablecoins make remittance fees go to zero. Sheel: Cool. So how does she actually get rupees? She lives in a village. Brian: Easy — she just offramps the USDC into INR through a local exchange. Sheel: And the exchange does this for free? Brian: Well, there's a small spread. But it's tiny. Sheel: Right. And how does she get from "INR balance on a crypto exchange" to "rupees she can buy rice with"? Brian: She withdraws to her bank account via IMPS. Sheel: Which has a fee. Brian: A small one. Sheel: And the exchange charged her KYC-compliance overhead, and took an FX spread on the USDC→INR conversion, and someone has to pay for fraud monitoring, and customer support when the transfer fails… Brian: Yes, but the blockchain part is free. Sheel: The blockchain part. The 0.3 seconds where the token moves between two wallets. Brian: Correct. That's where the $60 billion in fees disappears. Sheel: The other 99.7% of the process — the part with humans, banks, regulators, FX desks, and last-mile cash-out — that's also free? Brian: That's, uh, a separate question. Sheel: But you said remittance fees go to zero. Brian: With stablecoins, yes. Sheel: Right. So when Wise charges 0.5% all-in for the same corridor, end to end, including the rupees landing in her bank account… Brian: That's because they're not using stablecoins. Sheel: And if they did use stablecoins on the backend, which some of them already do, the fee would be… Brian: Zero. Sheel: It's currently 0.5%. Brian: I'm not sure I understand. Could you repeat the question? Sheel: Thanks for your time.

@EliBenSasson Those are the same thing for everyone except people who live in zkp-only world :P ECDH being broken means privacy is broken.

The next szn after privacy will be post-quantum security.

@uche2v @EliBenSasson a16zcrypto.com/posts/article/…

@EliBenSasson It would be nice if you did a little summary on post quantum security, similar to what you did with privacy last year.

Hot take: the "modular blockchain" thesis might still be alive if it had used the more accurate "dissagregated blockchain" instead.

@keoneHD Ever heard of fintech?

Many finance companies hold virtual monopolies that have been unchallenged for decades. The reality is that it is quite difficult for newcomers to take the mantle. Crypto offers builders a unique angle of attack to disrupt - by tapping into a unique, wealthy, influential, coordinated group of passionate users. It is not easy, but the reward is also massive. The challenge should never be trivialized, but neither should be the opportunity.

@victorzh @OpenPasskey Thanks for answering. Didn’t know you guys build your own terminal. Will have to look at your docs to understand how it works.

For QR payment with @OpenPasskey terminal. AUDD/AUDM goes from consumer’s wallet to a one time reciving wallet then goes to merchant’s vault. Then merchant decides how they want to handle the AUDD/AUDM. For card physical/digital with @OpenPasskey terminal. Similar to QR. For card physical/digital with other acquirers’ terminal(as any terminal can work with Visa, can work with our card as well), it depends on how the acquirers deal with the merchant. They route the tx to smart contract, AUDD/AUDM moved from consumer’s wallet to merchant directly or to acquirer then AUD to merchant.

The card tap, explained. That card tap isn’t a Visa card. It is an @OpenPasskey card. We hold our own IIN under ISO/IEC 7812, the same registry Visa and Mastercard use to identify issuers. Our card runs the same EMV chip standard as your bank card. Any acquirer that adds our routing entry can accept it on the terminals they already have. No new hardware. The chip signs with a P-256 key that never leaves the secure element. Non-custodial, the customer holds their own keys. Settlement runs on @base flashblocks, so confirmation lands in about 300ms.

@victorzh @OpenPasskey How/when/where does the merchant specify what end currency they want? If the merchant wants AUD, of course the off-ramp is part of the payment settlement..

@samlafer @OpenPasskey settlement happens on base is true. Merchant can choose to redeem AUDD/AUDM to AUD or keep the AUDD/AUDM. From AUDD/AUDM to AUD is not part of the payment settlement.

@a16zcrypto Hmm I guess you need certs for the leader so I’m missing one round: Client -> nodes (sign) -a2a-> nodes (cert) -> leader ?

“a malicious E could simply claim it never received I's value and essentially censor I. The fix is for I to broadcast its value to all validators” In practice can’t you just have the clients drive this broadcast? Then you get CR for 0 rounds latency. Client -> local node -> leader Client -> bunch of nodes’ IL -> leader

@ittaia @ElodinStorm x.com/samlafer/statu…

Work with @ElodinStorm and Ling Ren during last summer: eprint.iacr.org/2025/2136 One of the most exciting opportunities in the a16z Crypto Research summer internship program is the chance to work at the intersection of theory and practice of blockchains!

"The historical winners in each generation pretty much follow Moore's-law-like reasoning: when storage was mechanical and slow, SCSI's transport-agnosticism plus rich command set won; when storage became flash and the bottleneck shifted to per-command overhead, NVMe won by being designed for that regime"

Just realized: NVMe<>SATA RISC<>CISC

Extended my original diagram with hyperproperties. CIA x {safety, liveness} x {property, hyperproperty}

@saxenism :(

@samlafer sorry, ser!! just came to KL, Malaysia today XD

Turns out Canggu, Bali was never overrated.... if you want to relax, eat good food, work and workout. Also, great weather. Btw, let's catch-up for padel if you're around.

@ThogardPvP Printing press revolution. I like @oxidecomputer's "writing intensive" culture. Think that's the real win. Even with people in the office, force them write stuff down and argue in written form.

"Maybe we should start calling system memory SAM – sequential access memory – since it’s much better at sequential accesses than random accesses." duvanenko.tech.blog/2020/03/07/mem…

I feel like everything always circles back to: "cards already solved this". In the 4-party model, the acquirer eats merchant fraud, forcing it to vet merchants. Zelle's acquirer is not an underwriter, so they don't police receivers. UK has legislation which forces acquiring bank to bear 50% of the reimbursement costs. Problem solved (I think..?).

Banks want to blame Facebook for their scam problem, arguing that scams originate upstream from where they sit and that “the payment itself is often the final step” and is difficult to stop. It’s true that it is very difficult to stop a customer from paying a scammer in the moment. However, the argument that this is a problem for Facebook to solve rather than the banks is bullshit. While Facebook is certainly negligent on scams and needs to clean up its platform, the payment itself is not the first and only opportunity banks have to stop the scam. You know why? THE SCAMMERS HAVE BANK ACCOUNTS! Take Zelle as the example. Zelle is a closed network. It requires that the sender and the receiver both have bank accounts at banks that offer Zelle. This means that the bad guys have Zelle-capable bank accounts. They were KYC’d by Zelle banks. They receive stolen funds, instantly, into their Zelle-capable bank accounts. There is a lot that the banks and Zelle could do to stop (or at least slow down) these receiving accounts, but the uncomfortable truth is that they don’t really have an incentive to do that. Scammers are customers. They generate fee income. They contribute deposits that the banks lend against. Until you make banks liable for fraud on authorized transactions resulting from scams, you won’t see what types of scam prevention miracles (including on the receiver side) the banks are truly capable of.

@AlexH_Johnson How are they going to make sure the Anthropic fiasco doesn't happen for other stocks?

This will enable insiders at private companies to annihilate retail bettors with absolutely no regulatory supervision. But, at least with it being on-chain, we'll be able to watch it play out!