Introducing DeepProve—Lagrange’s zkML Library—a breakthrough in verifiable AI inference. We can now verify AI decisions instead of blindly trusting black-box models. And we can do it up to 158x faster than ever before. The future of AI is ZK. The future of humanity is Lagrange: 🧵

2/ No more guesswork. Prove with confidence that AI models are delivering results exactly as designed. DeepProve generates proofs for AI inferences up to 158x faster *and* verifies those proofs up to 671x faster than the leading alternative.

3/ What is zkML?? zkML fuses zero-knowledge proofs (ZKPs) and machine learning (ML) to verify AI computations cryptographically. With zkML you can now prove that: ✅ The AI is using the right ML model ✅ The AI arrives at the right results

4/ Why do we need zkML? 👉 To solve the AI extinction problem Superintelligence is inevitable. AI already influences finance, defense, and governance — yet today’s models are unverifiable and opaque. Without provability, we can’t know if AI is acting in our best interest. The solution? DeepProve

5/ DeepProve is the key to ensuring AI remains a tool for progress, not destruction. ✅ Proof generation up to 158x faster than the competition ✅ Proof verification up to 671x faster than the competition ✅ Proofs for CNNs, MLPs, & soon LLMs

6/ TL;DR: DeepProve lets you prove AI model inferences cryptographically—ensuring AI aligns with human intent—at unmatched speed and scalability Learn more about DeepProve, Lagrange’s zkML library, and how we’re pioneering verifiable AI: 🔗

Coral AI is the most powerful AI for documents. See the difference yourself:

There's a huge difference between "executing the model correctly" (what verifiable computation can check), and "producing outputs we expect"; never mind outputs consistent with principles of ethics and social justice. But the video conflates them.

keep cooking @Ismael_H_R 🫡

@eigenlayer this is pretty amazing! curious as to how it compares with state of the art zkml technologies?

Wen??

Wen tge? April is ok