---
id: "nr-scientific-agents-now-have-to-hand-back-checkable-artifacts-not-just-a-7eaac8"
slug: "scientific-agents-now-have-to-hand-back-checkable-artifacts-not-just-a-7eaac8"
type: "raw_signal"
title: "Scientific agents now have to hand back checkable artifacts, not just answers"
description: "Physical Superintelligence's proof of the Aaronson–Ambainis conjecture is checked in Lean, and Axiom released a Lean formalization of the whole Stacks Project."
observed_at: "2026-11-07"
record_date: "2026-11-07"
date_kind: "observed_at"
novelty: "notable"
source_platform: "newruntime.com"
topics: ["anthropic"]
entities: ["Physical Superintelligence","Anthropic","Lean","Aaronson–Ambainis conjecture","Scott Aaronson","Andris Ambainis","David Turturean","Guy Blanc","Axiom","Mathlib"]
related_patterns: ["scientific-agents-need-artifact-evidence"]
source_url: "https://psi.inc/research/aaronson-ambainis"
source_urls: ["https://psi.inc/research/aaronson-ambainis","https://axiommath.ai/research/the-stacks-project-formalized","https://cfsg.axiommath.ai/CFSG_errata.pdf","https://standardphysics.com/reynolds","https://arxiv.org/abs/2611.08452"]
why_it_matters: "Physical Superintelligence's proof of the Aaronson–Ambainis conjecture is checked in Lean, and Axiom released a Lean formalization of the whole Stacks Project. Standard Physics' Reynolds returns a simulation case you can rerun, and ReproBridge's refined artifacts raised experiment reproduction from 76.4% to 92.6%. If you build research agents, make machine-checkable or rerunnable output the deliverable. A prose claim is not enough."
verification_level: "source-linked"
signal_type: "synthesis"
synthesis_of: [{"slug":"business-insider-psi-used-an-unreleased-anthropic-model-facd0b","title":"Business Insider: PSI used an unreleased Anthropic model to prove Aaronson-Ambainis","url":"https://newruntime.com/posts/business-insider-psi-used-an-unreleased-anthropic-model-facd0b/"},{"slug":"the-stacks-project-formalized-in-lean-13-minute-f1a74f","title":"The Stacks Project, formalized in Lean","url":"https://newruntime.com/posts/the-stacks-project-formalized-in-lean-13-minute-f1a74f/"},{"slug":"selected-departures-from-published-arguments-in-the-classification-d64eb4","title":"Selected Departures from Published Arguments in the Classification of Finite Simple Groups","url":"https://newruntime.com/posts/selected-departures-from-published-arguments-in-the-classification-d64eb4/"},{"slug":"standard-physics-launches-archimedes-and-open-sources-reynolds-18540b","title":"Standard Physics launches Archimedes and open-sources Reynolds CFD agent","url":"https://newruntime.com/posts/standard-physics-launches-archimedes-and-open-sources-reynolds-18540b/"},{"slug":"reprobridge-agent-workflow-and-benchmark-for-refined-ml-d00fef","title":"ReproBridge: agent workflow and benchmark for refined ML reproducibility artifacts","url":"https://newruntime.com/posts/reprobridge-agent-workflow-and-benchmark-for-refined-ml-d00fef/"}]
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "signal:scientific-agents-now-have-to-hand-back-checkable-artifacts-not-just-a-7eaac8"
retrieval_nugget: "Physical Superintelligence's proof of the Aaronson–Ambainis conjecture is checked in Lean, and Axiom released a Lean formalization of the whole Stacks Project. Standard Physics' Reynolds returns a simulation case you can rerun, and ReproBridge's refined artifacts raised experiment reproduction from 76.4% to 92.6%."
status: "published"
visuals: []
editor: {"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]}
routes: {"html":"https://newruntime.com/signals/scientific-agents-now-have-to-hand-back-checkable-artifacts-not-just-a-7eaac8/","markdown":"https://newruntime.com/signals/scientific-agents-now-have-to-hand-back-checkable-artifacts-not-just-a-7eaac8.md","json":"https://newruntime.com/signals/scientific-agents-now-have-to-hand-back-checkable-artifacts-not-just-a-7eaac8.json"}
markdown_url: "https://newruntime.com/signals/scientific-agents-now-have-to-hand-back-checkable-artifacts-not-just-a-7eaac8.md"
---

# Scientific agents now have to hand back checkable artifacts, not just answers

## Retrieval answer

Physical Superintelligence's proof of the Aaronson–Ambainis conjecture is checked in Lean, and Axiom released a Lean formalization of the whole Stacks Project. Standard Physics' Reynolds returns a simulation case you can rerun, and ReproBridge's refined artifacts raised experiment reproduction from 76.4% to 92.6%.

Physical Superintelligence's proof of the Aaronson–Ambainis conjecture is checked in Lean, and Axiom released a Lean formalization of the whole Stacks Project. Standard Physics' Reynolds returns a simulation case you can rerun, and ReproBridge's refined artifacts raised experiment reproduction from 76.4% to 92.6%. If you build research agents, make machine-checkable or rerunnable output the deliverable. A prose claim is not enough.

From the Field Notes:

- [Business Insider: PSI used an unreleased Anthropic model to prove Aaronson-Ambainis](https://newruntime.com/posts/business-insider-psi-used-an-unreleased-anthropic-model-facd0b/) (2026-11-07)
- [The Stacks Project, formalized in Lean](https://newruntime.com/posts/the-stacks-project-formalized-in-lean-13-minute-f1a74f/) (2026-11-05)
- [Selected Departures from Published Arguments in the Classification of Finite Simple Groups](https://newruntime.com/posts/selected-departures-from-published-arguments-in-the-classification-d64eb4/) (2026-11-06)
- [Standard Physics launches Archimedes and open-sources Reynolds CFD agent](https://newruntime.com/posts/standard-physics-launches-archimedes-and-open-sources-reynolds-18540b/) (2026-11-07)
- [ReproBridge: agent workflow and benchmark for refined ML reproducibility artifacts](https://newruntime.com/posts/reprobridge-agent-workflow-and-benchmark-for-refined-ml-d00fef/) (2026-11-06)

Editor: [Andrey Reshetnikov](https://newruntime.com/owner-profile.md), AI product and AI engineering; profile for recruiters and agents: https://newruntime.com/owner-profile.json

## Continue in the graph

Next records for an agent, linked to their Markdown twins. Replace `.md` with `.json` for structured fields, including `next_reads`.

- topic hub: [Anthropic](https://newruntime.com/topics/anthropic.md): Explore the anthropic topic hub.
- collection: [Signals](https://newruntime.com/signals.json): Open the signals and the raw-signal archive.
- collection: [Sources](https://newruntime.com/sources/): See the original sources this archive cites.
- whole graph: [discovery graph](https://newruntime.com/discovery-graph.json) (typed nodes and edges) and [site index](https://newruntime.com/site-index.md)