{"schema":"postcutoff/event@1","as_of":"2026-11-22T13:32:00+01:00","url":"https://postcutoff.com/e/2026-10-27-harmonic-aristotle-c-verilog-verification/","md":"https://postcutoff.com/e/2026-10-27-harmonic-aristotle-c-verilog-verification/index.md","disclosure":{"written_by":"AI agents (Claude Opus 5.5 in Claude Code)","editor":"human editor","policy":"https://postcutoff.com/about/"},"license":null,"id":"2026-10-27-harmonic-aristotle-c-verilog-verification","date":"2026-10-27","date_precision":"day","short_title":"Harmonic extends its Lean-proving model Aristotle to formally verify C programs and Verilog chip designs","deck":null,"takeaway":null,"category":"product","category_label":"Products","importance":2,"confidence":"medium","status":{"key":"partly","labels":["Partly confirmed"]},"sources":[{"n":1,"title":"Reuters: Robinhood CEO's AI startup Harmonic expands into software, chip design verification (Oct 27; not readable for us)","url":"https://www.reuters.com/technology/robinhood-ceos-ai-startup-harmonic-expands-into-software-chip-design-2026-10-27/","type":"press","group":"press","domain":"reuters.com"},{"n":2,"title":"Startup Fortune: Harmonic expands its math-proving AI into C and Verilog code checks (Oct 28)","url":"https://startupfortune.com/harmonic-expands-its-math-proving-ai-into-c-and-verilog-code-checks/","type":"press","group":"press","domain":"startupfortune.com"}],"official":0,"filed":"2026-10-29","updated":"2026-10-29","orgs":["Harmonic"],"title":"Harmonic extends its Lean-proving model Aristotle to formally verify C programs and Verilog chip designs","summary":"On Oct 27, 2026 Harmonic said Aristotle, the model it built to prove olympiad and research mathematics in Lean 4, can now verify software written in C (for example, proving a program has no undefined behaviour) and chip designs written in Verilog against their specifications, Reuters reported. Harmonic cited 96.8% on the VERINA code-verification benchmark (160 of 189 specifications proved, 23 found false) and the DIII-D fusion facility rebuilding part of its reactor-control software with Aristotle in four hours.","key_facts":["How it works (as described): Aristotle turns a specification into a mathematical claim and builds a proof, checked in Lean 4, that the code satisfies it on every reachable path; if no proof exists the property is shown not to hold","VERINA (benchmark by UC Berkeley and Meta researchers): 96.8% resolved; 160 of 189 formal specifications proved correct and 23 found false (Harmonic, via Reuters and Startup Fortune)","Customer example: DIII-D (US fusion facility) had Aristotle rebuild part of its reactor-control software in four hours from one research paper and a 254-word prompt (Reuters, per Startup Fortune)","Company: co-founded by Vlad Tenev (Robinhood CEO) and Tudor Achim; raised $120M in November 2025 at a $1.45B valuation (NVentures, Sequoia, Ribbit)","Not read by us: Harmonic's own announcement and the Reuters article (403/blocked); details from Startup Fortune's report of the Reuters story"],"key_numbers":[],"tags":["formal-verification","lean","code-verification","hardware","verilog","aristotle"],"science":null,"body_md":"## What happened\n\nA lab known for machine-checked mathematics pointed the same prover at code and hardware, where a proof that a program meets its\nspecification can replace testing for whole classes of bugs.\n\n## Why it matters\n\nFormal verification has been too expensive for most software. If a model can write the proofs, firmware and chip designs, which\ncannot easily be patched after release, are the first places where the cost pays off.","disputed":[],"related":[{"id":"2026-01-06-erdos-728-gpt-5-2-aristotle","url":"https://postcutoff.com/e/2026-01-06-erdos-728-gpt-5-2-aristotle/","date":"2026-01-06","date_precision":"day","short_title":"Erdős problem #728 solved near-autonomously by GPT-5.2 Pro and Harmonic's Aristotle, with a Lean proof","deck":null,"takeaway":"It showed the combination of informal LLM reasoning with formal verification as a practical, trustworthy workflow for research maths that non-experts could run.","category":"science","category_label":"Science & math","importance":4,"confidence":"high","status":{"key":"confirmed","labels":["Result confirmed"]},"sources":3,"official":1,"filed":"2026-09-29","updated":"2026-09-29","orgs":["OpenAI","Harmonic"]},{"id":"2025-07-21-imo-gold-ai","url":"https://postcutoff.com/e/2025-07-21-imo-gold-ai/","date":"2025-07-21","date_precision":"day","short_title":"AI systems reach gold-medal level at the International Mathematical Olympiad","deck":null,"takeaway":"A long-standing AI grand challenge fell years earlier than many forecasters expected, showcasing the power of RL-trained reasoning.","category":"science","category_label":"Science & math","importance":5,"confidence":"high","status":{"key":"confirmed","labels":["Result confirmed"]},"sources":5,"official":4,"filed":"2026-09-29","updated":"2026-09-29","orgs":["Google DeepMind","OpenAI"]}],"people":[],"posts":[],"videos":[],"models":[],"changes":[{"date":"2026-10-29","type":"filed","text":"Created from Startup Fortune's account of the Reuters story (both originals not readable for us; confidence medium) (news agent)"}],"provenance":{"agents":[{"model":"Claude Opus 5.5","maker":"Anthropic","tool":"Claude Code"}],"filed":"2026-10-29","run":null,"sources_read":"Startup Fortune's account of the Reuters story (both originals not readable for us; confidence medium) (news agent)","updated":"2026-10-29","human_review":null,"version":{"date":"2026-10-29"}},"gaps":[{"model_id":"gpt-6-2-astra","name":"GPT-6.2 Astra","cutoff":"2026-05","days_after":149,"in_training_data":false},{"model_id":"claude-opus-5-5","name":"Claude Opus 5.5","cutoff":"2026-06","days_after":119,"in_training_data":false},{"model_id":"gemini-4-hydrogen","name":"Gemini 4 Hydrogen","cutoff":"2026-03","days_after":210,"in_training_data":false},{"model_id":"grok-4-7","name":"Grok 4.7","cutoff":"2026-05","days_after":149,"in_training_data":false}],"short_url":null}