Ian Miers

CS Prof. Security and applied cryptography.
Some highlights: Zerocash (zcash, et al. ), Zexe (Aleo, Aztec, etc ), zk-creds/zk-promises(...)

[Ian Miers @secparam]

@perrymetzger For "public-key" like crypto, if you're stuck with Needham-Schoeder like schemes, you're stuck with a pretty big back door. You want to add forward secrecy/PCS and TEEs to safe guard keys in centralized parties.

[Ian Miers @secparam]

The most "may you live in interesting times" hypothetical crypto timeline:
* AI breaks lattice crypto
* classic public key encryption is untrusted because of AI/Quantum.
But we have:
* hash based zero-knowledge proofs.
* hash based signatures.
* symmetric encryption.

[Ian Miers @secparam]

@colludingnode Its eve more bizare. The sender is secret unless you need to tell them who you are. So if buying an ebook, you say nothing. Depositing to your account, not.
The sender is revealed by you talking to them, the amount as well.

[Ian Miers @secparam]

@pratiks_crypto Thats slightly less interesting than quantum is always "10 years away", AI keeps getting better, and we really just don't know if we are in minicrypt or not. The twilight zone between minicrypt and Cryptomania, between ionized quantum vapor and vaporware, ...

[Ian Miers @secparam]

@colludingnode Interesting you mentioned banks. Zk payment schemes work. As do hash-based signatures.
So you can send private transactions, we have authenticity, but the secrecy of anything additional you tell the bank (amount, optionally your identity) depends on in person symmetric keys.

[Ian Miers @secparam]

@perrymetzger Needham-Schoeder is part of what makes it really interesting, since that means you need trusted intermediaries with existing relationships to communicate securely. Steampunk isn't the right word, More cyber than cyberpunk?

[Ian Miers @secparam]

@Keleesssss Context and bandwidth. An AI in the loop with full document access, maybe could distinguish. But by what criteria? Decided by who? What if it gets just a paragraph at a time? Can it communicate its confidence score and reasoning robustly, in the limited bits of a watermark?

[Ian Miers @secparam]

I understand the appeal of watermarking text. I did some of the 1st work on watermarks... in 2024. In 2026, even the idea of "pure human text" feels like a relic. A watermark can't differentiate a between human with AI assistance and just AI. images...
https://t.co/QuxR0QtAoO

[Ian Miers @secparam]

Images and video are a different story. There's a real distinction between content claiming to depict real events vs. memes, clipart, etc. So watermarks or signed media (like C2PA) could genuinely help distinguish . But watermarked text? No. Even this tweet went through an LLM.

[Ian Miers @secparam]

The choice between Zelle or Venmo is hilarious. Venmo of course had a on by default social payment feed. Zelle is run by Early Warning Services, LLC, a fintech company jointly owned by seven major banks that, as the name suggests, shares data for anti fraud.

[Ian Miers @secparam]

Old and busted: ai alignment for human safety.
New hotness: pm alignment for human safety.

[𝚑𝚎𝚗𝚔 𝚟𝚊𝚗 𝚎𝚜𝚜 @henkvaness]

Tonight I typed just one sentence into Google Earth and put refugees near the Mexican border. Then I planted a nuclear plant in Iran. What on earth is Google doing? Check my latest post here: https://t.co/N0h6CVECqS

[Ian Miers @secparam]

I agree, zk-ml makes sense on smaller models. But that doesn't change the core question: when are you willing to pay for zk verifiability vs a tee vs (for open models) running it yourself.
For you example with prompt injection: if a lab offers you 1% more per token for zk verified not injectable, 0.5% for TEE, and 0.01% for "we have a FIPS/ISO certified process" who will pay for zk?
I think you have to find classes of use cases where TEEs don't work

[Ian Miers @secparam]

Whats the premium the market will pay for zk inference vs inference in a TEE? For payments there's a clear reason we don't trust TEEs. And for identity, there's simply an availability problem: not enough programable trusted hardware on devices.
Are there similar cases for inference?

[Ian Miers @secparam]

The nth committee member just sat there. No laptop. No phone. Just... watching the thesis defense. Like a psychopath. Unaware his participation was to meet the dean's committee size requirement.

[Ian Miers @secparam]

Yes. And interestingly, it seems not to be a stylometry attack, its Vitalik's somewhat distinctive thinking and notation. Semantic stylometry, almost.

The same attack might not catch Vitalik if he, e.g., was posting political commentary.
But it might catch, e.g., a journalist, posting anonymously content they could not safely publish under their own byeline.

[Ian Miers @secparam]

This is a much more interesting case of AI deanonymization than the paper that was making the rounds. Here, it was someone trying to remain anonymous, but elements of their style made it through in detectable ways.

Whereas the previous paper took non-anonymous posts, redacted the username, and simply observed folks writing under their own name included identifying details.

[vitalik.eth @VitalikButerin]

And ... we have a winner!

My method when writing the post in 2024 was: I wrote it in Chinese, used qwen2.5 locally to translate it to English, then manually fixed all the bugs in the translation.

Notice that the stylistic hints that his AI picked up on were intellectual habits and style of math and algorithm explanation, which bypassed my obfuscation strategy (which only covered prose) completely.
https://t.co/d04MwOjkHP

[Ian Miers @secparam]

@luhelminger There were criticisms of differential privacy in the census. But unclear if those relate to what happened or not. And I haven't seen any explanation of why the decision was made.
https://t.co/e8f0Y8fyhz

[Matthew Green @matthew_d_green]

Demographers hate this One Weird Trick.

[Ian Miers @secparam]

On the scale of American privacy emergencies, where does "The census no longer has differential privacy" fall?
Its an honest question. To me, it clearly isn't in the top 10 privacy issues. But perhaps its in the top 10 which are plausibly solvable?

[Ian Miers @secparam]

@zkDragon @thedrekal @ZecHub @Zcash Threshold signatures for a chain with zk proofs are ... somewhat unnecessary long term. You can just prove you have sufficient signatures from keys needed to authorize the transaction. Works with existing signing hardware and captures richer semantics.

[Ian Miers @secparam]

Or perhaps the emergency is "we cannot reject differential privacy, its the best privacy solution we have." But then, we can ask the same question: For the top 10 privacy problems we have, how many is it a solution for?