AI just killed Crypto...
TL;DR
Scott Aaronson says the quantum threat is now a 2029 problem, not a distant one — In a new blog post, the UT Austin quantum theorist says trusted experts in quantum hardware and error correction now believe a fault-tolerant machine capable of breaking deployed crypto systems could arrive around 2029.
This is bigger than Bitcoin: Shor’s algorithm threatens internet identity, software signing, banking, satellites, and government secrets — Wes explains that the real danger is public-key cryptography like RSA and elliptic curves, which underpin website certificates, software updates, military archives, and crypto wallets.
Google and Cloudflare are acting like the deadline is real — Google’s March 25, 2026 post sets 2029 as its target for migrating internal infrastructure to post-quantum cryptography, while Chrome, Cloud, and Android are already integrating protections; Cloudflare is aiming at the same date.
Crypto is uniquely exposed because blockchains are public, permanent, and built around immutable keys — Dormant wallets, including Satoshi-era coins, could become stealable if exposed public keys can be reverse-engineered by quantum systems, while chains like Ethereum face a governance nightmare over who gets to “change the locks.”
AI is part of why this threat may be accelerating — Wes ties DeepMind’s AlphaQubit work on quantum error correction to the broader story: AI helped reduce one of quantum computing’s biggest bottlenecks, making large-scale reliable quantum systems more plausible sooner.
Aaronson’s warning lands harder because he’s not a hype man — Wes stresses that Aaronson has spent years correcting exaggerated quantum claims, so when someone with Google quantum-supremacy credentials, OpenAI superalignment ties, and National Academy of Sciences status says “heed my warning,” the industry should take it seriously.
The Breakdown
Why Wes takes Scott Aaronson seriously
Wes opens by grounding the whole video in Scott Aaronson’s credibility: Google’s old quantum-supremacy project, early OpenAI superalignment work, and a reputation for being almost too smart to follow in every interview. He adds his usual personality — sunglasses because ChatGPT told him he had “L riz,” a Matrix tangent, and a joke that there were obviously no Matrix sequels — but the point is clear: this is not some random doomer blog post.
The warning: 2029 could be the year deployed cryptography starts breaking
The core claim comes from Aaronson’s new post, where he says experts in quantum hardware and error correction he trusts more than himself now think a fault-tolerant quantum computer that can break real cryptosystems should be possible by around 2029. Wes lingers on that because it changes the framing from “someday” to “less than three years,” and he compares the momentum to the AI race: nobody is slowing down to let security catch up.
What quantum actually breaks — and why it’s not about your password
Wes pauses to explain the basics in plain English: the danger is public-key cryptography, not every form of encryption and not your average password. He uses a locked-box analogy, then says quantum doesn’t “pick the lock” so much as make the math behind the lock stop being hard, with Shor’s algorithm as the key historical breakthrough from the 1990s.
The two-sided attack: yesterday’s secrets and tomorrow’s trust
One of the stickier ideas in the video is Wes comparing the threat to Tenet’s “temporal pincer attack.” On one side, governments have long been saving encrypted traffic now in hopes of decrypting it later; on the other, future trust online breaks if attackers can impersonate software signers, banks, or even Satoshi and move coins from dormant wallets.
Google’s behavior is the biggest tell
Wes keeps returning to one point: if Google is setting a 2029 migration target for post-quantum cryptography, everyone else should stop pretending this is a 2040 issue. He cites Google’s March 25, 2026 post, its work since 2016, active PQC efforts across Chrome, Cloud, and Android, and notes Cloudflare is on a similar timeline — which makes the threat feel less hypothetical and more like a countdown.
Why crypto may get hit hardest first
For Bitcoin, Wes highlights the nightmare scenario around old addresses and dormant coins: once a public key is exposed on-chain, a future quantum machine might derive the private key and steal the funds. Ethereum may have more flexibility because Vitalik Buterin and the ecosystem can coordinate changes, but that opens a governance problem Aaronson cares about deeply: who gets to rewrite the rules when the old locks stop working?
Google disclosed the risk without handing out the exploit
Wes points to Google Research saying future quantum machines may break elliptic-curve cryptography with fewer qubits and gates than previously thought. He likes the zero-knowledge-proof framing here: Google is basically saying, “we know the password,” without publishing the actual password or an attacker’s manual.
Why this is still an AI story
In the final stretch, Wes connects the dots back to AI through DeepMind’s AlphaQubit, which he says helped tackle one of quantum computing’s biggest bottlenecks: error correction. That’s the punchline of the whole video — AI isn’t just automating software jobs or making chatbots smarter; it may also be accelerating the arrival of quantum systems that force the entire internet, from Coinbase to governments to your phone, onto post-quantum security.