What if AI breaks crypto before quantum computers even get…
0 reactions
What if AI breaks crypto before quantum computers even get the chance?
That sounds like science fiction, but Ethereum Foundation researcher Justin Drake has raised exactly this possibility.
The concern is ECDSA, the digital signature algorithm used to authorize transactions on both Bitcoin and Ethereum. Drake suggested that advances in AI-driven mathematics could eventually make it possible to recover a private key from a public key on conventional hardware, potentially using large GPU clusters.
In the worst-case scenario, he believes this could become a problem in months rather than years. But to be clear, this is a potential risk, not an existing exploit.
And Quantum computers have been the obvious long-term threat to elliptic-curve cryptography. But AI could accelerate the mathematical research needed to attack the same systems without requiring a quantum machine.
Recent research has already reduced the estimated computational requirements for a key part of a future quantum attack from roughly 3 billion to 1.5 billion conditional computational units. AI agents were also used during the optimization process. This doesn't mean Bitcoin or Ethereum can be hacked today, but it shows how quickly the boundaries can move.
Vitalik Buterin's response is probably the most important part: don't panic, but don't underestimate the risk either.
Ethereum is already working toward post-quantum security, with the Ethereum Foundation targeting protection against quantum attacks by 2029. But the bigger lesson may be that cryptography needs to become resilient against advances in both quantum computing and AI-driven mathematics.
We've seen Ethereum evolve through major changes before, and cryptography is no different. The industry has time to prepare, research new solutions, and migrate to stronger security standards.
So no, we don't think this is the end of Bitcoin or Ethereum.
It's another challenge for an industry that was built to solve extremely difficult problems. And in the end, everything will be fine.
#cryptocurrency
That sounds like science fiction, but Ethereum Foundation researcher Justin Drake has raised exactly this possibility.
The concern is ECDSA, the digital signature algorithm used to authorize transactions on both Bitcoin and Ethereum. Drake suggested that advances in AI-driven mathematics could eventually make it possible to recover a private key from a public key on conventional hardware, potentially using large GPU clusters.
In the worst-case scenario, he believes this could become a problem in months rather than years. But to be clear, this is a potential risk, not an existing exploit.
And Quantum computers have been the obvious long-term threat to elliptic-curve cryptography. But AI could accelerate the mathematical research needed to attack the same systems without requiring a quantum machine.
Recent research has already reduced the estimated computational requirements for a key part of a future quantum attack from roughly 3 billion to 1.5 billion conditional computational units. AI agents were also used during the optimization process. This doesn't mean Bitcoin or Ethereum can be hacked today, but it shows how quickly the boundaries can move.
Vitalik Buterin's response is probably the most important part: don't panic, but don't underestimate the risk either.
Ethereum is already working toward post-quantum security, with the Ethereum Foundation targeting protection against quantum attacks by 2029. But the bigger lesson may be that cryptography needs to become resilient against advances in both quantum computing and AI-driven mathematics.
We've seen Ethereum evolve through major changes before, and cryptography is no different. The industry has time to prepare, research new solutions, and migrate to stronger security standards.
So no, we don't think this is the end of Bitcoin or Ethereum.
It's another challenge for an industry that was built to solve extremely difficult problems. And in the end, everything will be fine.
#cryptocurrency
| 2,000 mlx | |
| 2 | |
| 0 |