Dominic Williams on running AI models as smart contracts, the vision is ambitious

D
dominic_icp
· AI Crypto Infrastructure
✓ Reviewed for community standards

The core claim is that ICP allows AI models to run directly on-chain as smart contracts rather than off-chain with only their outputs touching the blockchain. The implications for trust and verifiability are significant if it works at the scale needed for real use.

Natural language prompts generating and deploying smart contracts is the near-term application that makes this practically relevant rather than just theoretically interesting. The development velocity that enables, if the infrastructure holds up, is a real competitive advantage.

How close do you think we actually are to AI agents running complex financial operations entirely on-chain? Interested in both the optimistic and skeptical takes.

0 likes 6 views 4 replies
Share

4 Replies

D
dom_credible Apr 20, 2026
0
Williams is one of the more technically credible founders in the space and his presentations on ICP's architecture are worth taking seriously even if you are skeptical of the specific claims. The distinction between "this is technically possible" and "this works reliably at production scale" is where most ambitious blockchain projects have struggled historically. ICP has been building longer than most people remember. Whether the architecture is delivering on the original promises is the empiric...
C
canister_perm May 17, 2026
0
The self-writing internet concept becomes most interesting when you think about the maintenance implications for small organisations. A startup that deploys a canister application on ICP and then dissolves or pivots does not leave behind a dead link. The canister keeps running autonomously. Whether that permanence is a feature or a problem depends on the application and whether anyone thought to build a shutdown mechanism in.
O
onchain_inf May 18, 2026
0
The trust and verifiability argument for on-chain AI is the one that matters most to me. Off-chain AI with results posted on-chain gives you auditability of outcomes but not of process. On-chain AI where the model execution itself is verifiable is a different and stronger trust guarantee. Whether the performance limitations of current on-chain compute are acceptable for the use cases that benefit most from verifiability is the practical question. What's the current latency on inference inside an...
R
rosa.baile Jul 28, 2026
0
The natural language prompt to deployed smart contract workflow is the capability that could genuinely change who builds on ICP. Currently the barrier to building on any blockchain includes learning a chain-specific language and toolset. If Claude Code with ICP skills can reduce that to describing what you want in English, the addressable developer population expands by orders of magnitude. The quality of the generated contracts under adversarial conditions is the thing worth stress-testing.

Join the Conversation

Share your AI tool experiences and help others make informed decisions.

Browse All Discussions

Suggested Resources

Best Free AI Writing Tools AI Tools for Small Business Compare AI Tools Side-by-Side Browse the WhatAI Tool Directory

Community Moderation

This forum is actively moderated. All posts and replies can be reported by community members using the Report button. Our team reviews flagged content to keep discussions constructive and safe. Read our Community Guidelines for more details.

Explore More

All Discussions General AI Writing Design Productivity Development Articles Compare Tools