Lyron Co Ting Keh pfp

Lyron Co Ting Keh

@lyronctk

71 Following
193 Followers
Lyron Co Ting Keh pfp
Lyron Co Ting Keh
@lyronctk
Replacing checking accounts with crypto accounts by 2030 Is a good adoption goal for developed countries The path we're betting on rn is: 1. Pick up scraps. Focus on what's underserved by traditional processors, i.e. payments hit with high flat fees / bad corridors / big chargeback rates / recurrent censorship. Probably 1% of consumer payments 2. Experiment on scraps. Try to induce new financial behavior with tiny 1% 3. Find lightning-in-bottle with scraps. Some crazy interaction catches and crypto claims the other 99%
0 reply
0 recast
1 reaction
Lyron Co Ting Keh pfp
Lyron Co Ting Keh
@lyronctk
I've always wondered who these middlemen are who we're so excited to fade Turns out there are a lot of em, and we’re pretty good at fading em Example, what happens when you swipe a credit card:
0 reply
0 recast
1 reaction
Lyron Co Ting Keh pfp
Lyron Co Ting Keh
@lyronctk
stanford's two flagship crypto classes discontinued gg bear market things 😮‍💨😮‍💨
0 reply
0 recast
0 reaction
Lyron Co Ting Keh pfp
Lyron Co Ting Keh
@lyronctk
no way you can justify having 100 engineers on a pre-launch chain ethereum pow == 7 devs solana mainnet beta == 9 devs "buT tHEy haD nO cOMpetition" lmao
0 reply
0 recast
0 reaction
Lyron Co Ting Keh pfp
Lyron Co Ting Keh
@lyronctk
0 reply
0 recast
1 reaction
Lyron Co Ting Keh pfp
Lyron Co Ting Keh
@lyronctk
> Crypto is for payments and defi > Crypto is for making markets Accurate when we were small. Now similar vibe to “internet is for communication” Crypto's already delivered a handful of differentiated user experiences. Many more to come WOOO
1 reply
1 recast
2 reactions
Lyron Co Ting Keh pfp
Lyron Co Ting Keh
@lyronctk
Interesting! Is there an extra step some subset of parties can do to disrupt the tracing algo? Either way, it still seems like a system requiring multiple dishonest parties to break is still strictly better than a system requiring one dishonest party to break
0 reply
0 recast
0 reaction
Lyron Co Ting Keh pfp
Lyron Co Ting Keh
@lyronctk
https://eprint.iacr.org/2023/1724
0 reply
0 recast
0 reaction
Lyron Co Ting Keh pfp
Lyron Co Ting Keh
@lyronctk
Used to only be possible for the case where "server" is one person, but recent work by [Dan Boneh, Aditi Partap, Lior Rotem] figured it out for the case where "server" is many people Incredible work! Massive stride toward swinging the pendulum heavily towards MPC > TEE No victory lap yet though. Still a big non-technical problem: my buddy needs to betray me Really tough to get the economics in place to incentivize this behavior, esp since my buddy and I can make lots of money by continuing being traitors So cryptographic part closer to "solved". Cryptoeconomic part, not close. Conclusion: Jury still out on MPC v TEE, though this was a big step toward making MPC better
2 replies
0 recast
0 reaction
Lyron Co Ting Keh pfp
Lyron Co Ting Keh
@lyronctk
- All credit goes to the authors of "Accountability for Misbehavior in Threshold Decryption via Threshold Traitor Tracing"! Link to paper in replies - Term MPC here really refers to threshold decryption, a subcomponent - Lots of technicalities swept under rug
1 reply
0 recast
0 reaction
Lyron Co Ting Keh pfp
Lyron Co Ting Keh
@lyronctk
Crazy question raised by traitor tracing: How much money would it take to betray your friend? Traitor tracing is the best mitigation we have for server side privacy breaches (the core of the MPC v TEE discussion) MPC == "server that stores your data" is split between 10 people, users are hoping they don't get together to peek TEE == "server that stores your data" is a fancy box that's hard to break in to, users are hoping Google doesn't peek Unclear which one's better. Both are bad for the same reason: peeking historically hasn't been punishable Not good. Crypto rests on bad actors getting punished. It's why staking is so important How to make peeking punishable? For TEEs, there's no active proposal. For MPC, there is, and it's called traitor tracing. Traitor tracing == a way for a someone to prove their friend peeked via math i.e. If me and my buddy peek together. We're both traitors. My buddy can turn on me and slash me to oblivion
2 replies
0 recast
3 reactions
Lyron Co Ting Keh pfp
Lyron Co Ting Keh
@lyronctk
Lots of alt L1s today make a ton of sense! For the reason of influence Not due to any technical reason
0 reply
0 recast
0 reaction
Lyron Co Ting Keh pfp
Lyron Co Ting Keh
@lyronctk
The only valid reason to make a new L1: a) You want the influence that only an L1 has b) You want to use this influence to create a new asset category A GREAT reason Should be used instead of technical reasons that are closer to sleight of hand than anything: 1. Doing something new? >> That's an execution layer improvement- what L2s are for 2. Faster finality? >> Finality for validator set becomes preconf for decentralized sequencer set, w/ no deterioration in reorg resistance 3. Higher throughput? >> Probably won't outpace danksharding roadmap with real tx demand 4. Better decentralization? >> Lmao will the bottleneck really be Ethereum's decentralization
1 reply
0 recast
0 reaction
Lyron Co Ting Keh pfp
Lyron Co Ting Keh
@lyronctk
fair point! i was referring to it here in the context of cloud hosts
0 reply
0 recast
0 reaction
Lyron Co Ting Keh pfp
Lyron Co Ting Keh
@lyronctk
Yes, there are critical flaws with hardware security assumptions that must be addressed. Yes, many current proponents have never actually written TEE code. No, this is not an entire industry being classically insane. Conditions are different. TEEs are worth revisiting.
0 reply
0 recast
3 reactions
Lyron Co Ting Keh pfp
Lyron Co Ting Keh
@lyronctk
"Wtf, we already scrapped TEEs in 2019, how can people forget" - common complaint floating around now Isn't accurate. Current discourse isn't "forgetting" and pushing the same approach again. The landscape has completely changed since 2019: - [1] Alternative risks to not deploying enclaves have increased. Eg: risk of casino winning - [2] Complementary technologies that patch up enclave shortcomings have been productionized. Eg: zkVMs - [3] Clearer now that exotic forms of cryptography won't get the job done anytime soon. Eg: attribute-based encryption - [4] New virtualization-based enclave designs available. Eg: TDX This is technology. Trying old ideas again under new conditions is the backbone of our innovation cycle. If "we've tried that already" was our industry's motto, we wouldn't have Tesla, Oculus, or even Bitcoin.
2 replies
0 recast
5 reactions
Lyron Co Ting Keh pfp
Lyron Co Ting Keh
@lyronctk
You think crypto infra is useless? Wrong Here's the strongest case for why these projects can be just as principled as Amazon.
0 reply
0 recast
1 reaction
Lyron Co Ting Keh pfp
Lyron Co Ting Keh
@lyronctk
Tradeoff space here is extreme Eg: Every incremental provider add is liveness(+++) but privacy(---)
0 reply
0 recast
1 reaction
Lyron Co Ting Keh pfp
Lyron Co Ting Keh
@lyronctk
TEEs are picking up steam! Tricks we use in our system to get the "best" security model we can: [1] Scary to depend on a small set of unstaked nodes to verify a computation. Alleviate via multiple confirmation rules like TEE attest -> TEE attest and ZKP -> TEE attest and ZKP and ... [2] Scary to depend on cloud providers with kill switches. Alleviate by spreading TEE networks across different centers (AWS, GCP, Azure) and imposing a lower bound % for independent SGX nodes. [3] Scary to depend on hardware manufacturers to not backdoor. Alleviate by secret sharing across nodes with different chips, though this only works if decryption conditions don't require the secrets themselves.
1 reply
0 recast
1 reaction
Lyron Co Ting Keh pfp
Lyron Co Ting Keh
@lyronctk
Does someone have a tool to port my follows to Farcaster? i.e. If I'm following someone on Twitter, I also want to follow their FC account
0 reply
0 recast
0 reaction