Ethereum Devs to AI: 'Write Code Faster So We Can Audit Slower, Please'

Ethereum's mastermind, Vitalik Buterin, has given a nod to an AI experiment that somehow managed to vibe-code a prototype of the blockchain's entire 2030 roadmap in a matter of weeks. "This is quite an impressive experiment. Vibe-coding the entire 2030 roadmap within weeks," Buterin mused on X, probably while the rest of us were still trying to get a simple smart contract to compile.

The whole saga started in February when a brave (or foolhardy) developer made a bet with Buterin that a single person, armed with AI, could crank out a reference implementation of Ethereum's grand plan. Buterin noted that AI is "massively accelerating coding" and suggested that the degens should "be open to the possibility that the Ethereum roadmap will finish much faster than people expect, at a much higher standard of security than people expect." So, the future is coming fast, and it's written in AI-generated spaghetti code.

For the uninitiated, vibe coding is the art of using AI to generate application code, enabling software to be created at the speed of a thought—or at least the speed of your GPU. Its popularity has skyrocketed alongside better AI models, though some party-poopers warn that AI-generated code can be about as secure as a screen door on a submarine.

Buterin, ever the voice of reason, flagged some "massive caveats." He pointed out that the breakneck speed means the code "almost certainly [has] lots of critical bugs, and probably in some cases 'stub' versions of a thing where the AI did not even try making the full version." He added, "But six months ago, even this was far outside the realm of possibility, and what matters is where the trend is going." In other words, the code is a mess, but it's a futuristic mess.

He cautioned that instead of just chasing the dragon of raw speed, the ecosystem should double down on security. "The right way to use it is to take half the gains from AI in speed, and half the gains in security: generate more test-cases, formally verify everything, make more multi-implementations of things." It's the classic crypto trade-off: you can have it fast, or you can have it secure. Why not try for a bit of both?

Buterin couldn't hide his personal excitement that the dream of bug-free code, "long considered an idealistic delusion," could first become possible and "then a basic expectation." Imagine that—software that actually works as advertised. What a time to be alive.

The project reportedly churned out roughly 700,000 lines of code covering 65 roadmap items in about two weeks, with the prototype even managing to sync to Ethereum's mainnet. Buterin was quick to clarify that AI isn't ready to replace human devs just yet: "Do not assume that you'll be able to put in a single prompt and get a highly-secure version out anytime soon; there WILL be lots of wrestling with bugs... But even that wrestling can happen 5x faster and 10x more thoroughly." So, the future of development is a faster, more thorough bug-hunt. Joy.

He argues this AI-assisted approach could turbocharge Ethereum's long-term roadmap, while admitting that achieving 100% perfect security is about as likely as finding a bug-free yield farm. "There are many specific cases, where specific security claims can be made and verified, that cut out >99% of the negative consequences that might come from the code being broken." So, we're aiming for "mostly secure," which, in crypto, is basically Fort Knox.

In a separate but equally important note, Buterin confirmed that Ethereum's long-discussed account abstraction, or smart accounts, is finally set to arrive. "Finally, after over a decade of research and refinement of these techniques, this all looks possible to make happen within a year." The upgrade, enabled by E