la

Previously used zk proof systems always felt like building blocks; if you changed a small piece, you had to knock down the entire block tower and start over.
Even if it’s just adding a new batch of data to the AI model or tweaking a small parameter, all the previously calculated proofs would be wasted, and you’d have to rerun everything from scratch, which is both time-consuming and computationally expensive. Implementing this in real-time updating AI scenarios is extremely challenging.
Fortunately, the research team at @Lagrange Official came up with Dynamic SNARKs, directly resolving this 'deadlock.'
Now, proofs can evolve flexibly with new data, eliminating the need for redundant work: for example, when a hospital's AI diagnostic model adds new case data, there’s no need to regenerate all proofs; dynamic adjustments can yield reliable results. Similarly, in DeFi, AI trading strategies can quickly validate updated parameters without delaying real-time trading.