Back to feed
News Story
APriority83
DeepTech深科技
2 sources

Claude Enters AI Drug Discovery: 14 of 15 Protein Designs Successful, Chemical Analysis in 23 Minutes

On August 18, Anthropic released Claude's latest experimental results in AI-driven drug discovery. In protein design, Claude successfully designed binders for 14 out of 15 targets, and in chemical analysis, it processed raw NMR and LC-MS data in 23 and 19 minutes respectively, matching lab results. This demonstrates AI's potential to accelerate early-stage pharmaceutical R&D.

SynthePulse Insight · AI deep readingMembers

Claude's Scientific Breakthroughs: From the Riemann Hypothesis to Protein Design, AI Is Becoming a True Research Assistant

Version 1 · 2 sources

Anthropic's experimental Claude has achieved multiple breakthroughs in mathematics and biochemistry: improving a Riemann hypothesis-related lower bound from 41.6% to 67.2%, successfully designing 14 out of 15 protein targets, and completing chemical analysis in just 23 minutes. These results mark AI's transition from a problem-solving tool to a research collaborator.

  • While attempting to prove the Riemann hypothesis, Claude unexpectedly improved the related lower bound from 41.6% to 67.2%, but did not prove the hypothesis itself.
  • In protein design, Claude successfully designed binders for 14 out of 15 targets, with a hit rate far exceeding typical industry levels.
  • Claude Opus 5 directly processed raw NMR and LC-MS data, completing analyses in 23 and 19 minutes respectively, with results largely consistent with the laboratory.
Open section navigationMathematical Breakthrough: Improvement of a Riemann Hypothesis-Related Lower Bound

Mathematical Breakthrough: Improvement of a Riemann Hypothesis-Related Lower Bound

Anthropic claims that its experimental Claude system, while attempting to prove the Riemann hypothesis, unexpectedly improved the related lower bound from 41.6% to 67.2%. This result does not prove the hypothesis but improves the mathematical guarantee regarding the proportion of zeros of the zeta function on the critical line.

Claude generated approximately 31 million output tokens across two Claude Code sessions, coordinated about 60 subagents, executed about 2,400 shell commands, and ran thousands of numerical checks. It combined Montgomery's 1973 work, Bombieri's 2000 paper, and recent methods by Baluyot et al.

Anthropic's internal mathematicians reviewed the paper and invited Brian Conrey and Dan Goldston for external review. Claude also used Lean to generate a formally verifiable version of the proof.

Free for now

Read the full analysis

3 more sections of analysis, plus the full takeaway

Loading

Credibility boundary

This report is based on Anthropic's official statements and blog posts, as well as reports from The AI Insider and DeepTech. All data are self-reported by Anthropic and have not been independently verified. The mathematical result was reviewed by external experts but has not been formally published. Protein design and chemical analysis results come from Anthropic's experiments, with no mention of peer review.

Sources for this version

  1. Anthropic Says Claude Improved a Longstanding Bound Tied to the Riemann HypothesisThe AI Insider
  2. Claude正式杀进AI制药：蛋白设计15个成功14个，化学分析23分钟出结果DeepTech深科技

Primary report

DeepTech深科技

Primary source

Same-event coverage

Also covered by 1 sources

The AI InsiderT2Anthropic Says Claude Improved a Longstanding Bound Tied to the Riemann Hypothesis