5/Industrial AI will govern supply chains, infrastructure, and physical outputs.It needs verification at the decision layer — not trust in the black box.That’s what @PolyhedraZK is building: zkML as the verification layer for autonomous systems. ...

4/Imagine manufacturing where:✅Every optimization step includes a cryptographic execution trace.✅Compliance can be confirmed without inspecting internal process models.✅ Supply chains coordinate using shared proofs.zkML turns automation into auditable coordination. ...

Why zkML? Because @nvidia and @Samsung are building an AI megafactory — 50,000+ GPUs coordinating chip production, robotics, and industrial workflows.When factories begin reasoning, not just executing, the key question becomes:How do we verify the decisions machines make? https://t.co/FiD1vs532R ...

This week in AI:• OpenAI expands into India — ChatGPT Go goes free.• Reddit sues Perplexity for scraping training data.• Tencent reveals half its code is now AI-generated.• Researchers build chips that compute with light.Every headline points to one truth:we’re scaling faster than we can verify.On this week’s XTC Show, TC joined @Tony0kai to talk about what comes next — the invisible infrastructure that keeps AI verifiable and trustworthy.Enter Expander — Polyhedra’s cryptographic acceleration layer that bo ...

Join @Tony0kAI alongside @Tiancheng_Xie and @sampearton for a week of live shows and fun.Verified community members engaging will win USDT. Make sure to join our Discord to be eligible for rewardshttps://t.co/uGLwFzaD5p https://t.co/e04SnT3QpY ...

4/Next Integration Phase: Began integrating the optimized FPGA LinearGKR modules into the full GKR proof pipeline, preparing for end-to-end hardware verification and unified GPU–FPGA co-design testing.Stay tuned for more updates. ...

3/LinearGKR Hardware Refinement: Enhanced the dual-phase prepare operator by introducing fine-grained pipeline segmentation and memory reuse between the two stages. This reduced BRAM utilization by over 20% while maintaining full concurrency. ...

We’re keeping the momentum strong! This week, the dev team has been working behind the scenes on GPU and FPGA acceleration for Expander. https://t.co/6cTHssW3E6 ...

4/Gemma3 Quantization: Introduced front-end padding for variable-length inputs to enable flexible inference.Next Steps (Circuit Optimization): Plan to prune gemma3 nodes during circuitization to reduce redundancy and improve proof efficiency.Stay tuned for more updates. ...

3/ECC Framework Review: Identified key issues including strong binding between computation graph and witness (forcing re-setup on each witness change) and reshape bugs for non-power-of-two dimensions. ...