Alex retweetledi
Alex
5 posts

Alex retweetledi

"Secure Computation vs. Centralization: The Battle That Will Define the Internet"
New Academy talk! Alex Ozdemir x @Stanford explains why secure computation may be the key to decentralisation
Full below👇
📽️ by @0xfarbey & @BabyBitProd
English
Alex retweetledi

As a performative math major when I finally got to making my personal site I wanted it to actually reflect that. Not just another CS portfolio. So I recreated the Desmos UI from scratch and would love to hear what you think
ahmedxahmed.com
English
Alex retweetledi

@AbdelStark Wdym by STARKs? If you mean transparent SNARKs, then which kind, if you’re talking about hash based (FRI) then sure it’s post quantum secure but not scalable enough for interesting use cases
English

The fact that STARKs are plausibly quantum-resistant is such an underrated feature.
And I think particularly in the context of verifiable AI / AI safety.
Secure Enclaves/TEEs (like Intel SGX or NVIDIA H100 confidential compute) use classical PKI with elliptic curve-based cryptography.
That attestation chain is fundamentally quantum-vulnerable.
You might think: "Yeah, but they can just switch to post-quantum crypto in the coming years."
Not really. The problem is that the provisioning keys are burned into the silicon during hardware fabrication.
So not only do you need to move to PQ crypto for the whole chain of trust, but you have to wait for an entirely new generation of hardware to be built and deployed.
To make things worse, this makes these systems vulnerable to « retroactive forgery » / « loss of non repudiation ».
If a quantum computer cracks the manufacturer's root key in 2035, an attacker can forge a perfectly valid, backdated TEE attestation from today.
So for any use case where the verifiable execution evidence needs to outlive the hardware that produced it, TEE attestation has a structural expiration date that ZK STARKs do not have.
When STARKs become practical for provable inference, they will really unleash their full supremacy.
Pure math-based, scalable, private, post-quantum.
English


