Post-Cutoff.com
  1. Home
  2. Science & Math

Science & Math

119 results found by AI or with AI's help. Each shows who did the work, how it was verified, and whether it still holds. Newest first, grouped by field. Also as Markdown: science.md.

astronomy (3)biology (18)chemistry (2)climate-weather (6)computer-science (9)materials (6)mathematics (52)medicine (12)other (2)physics (9)

astronomy

biology

2026-09-23 · microbiology / genome mining · disputed

Claude agents discover a novel CRISPR-like enzyme system; Anthropic reveals its own biology wet lab

Problem
Discovering new bacterial/phage defence and genome-editing enzyme systems
Result
Identified 'array-associated reverse transcriptases' (ARTs): phage reverse transcriptases adjacent to long CRISPR-like repeat arrays, from 200,000+ reverse transcriptases and 3,500 candidate systems; biological function still unknown.
AI
Claude (≈950 parallel agents)
Checked
Company preprint; experiments ongoing; not peer-reviewed
2026-02-10 · drug discovery / structural biology · pending

Isomorphic Labs unveils IsoDDE drug-discovery engine, hailed as 'an AlphaFold 4' — but proprietary

Problem
Predicting protein–ligand binding poses and affinities and antibody–antigen structures
Result
Proprietary engine reported to beat Boltz-2 and physics-based methods on binding-affinity prediction and reach state of the art on antibody–target structures, generalising to molecules unlike its training data.
AI
IsoDDE
Checked
Company technical report only; not peer-reviewed, model not released
2025-11 · data-driven discovery (multiple fields) · pending

Edison Scientific's Kosmos AI scientist claims six months of research per run

Problem
Autonomous data analysis and literature synthesis to generate discoveries
Result
An agent system reproduced three unpublished human findings from raw data and proposed four new findings, with 79.4% of statements judged accurate.
AI
Kosmos
Checked
Preprint (arXiv 2511.02824); accuracy assessed by independent scientists hired by the company
2025-11 · antibody engineering · confirmed

Baker lab designs antibodies from scratch with atomic accuracy using RFdiffusion

Problem
Designing antibodies to a chosen epitope computationally, without immunisation or library screening
Result
Epitope-specific antibodies designed de novo, with cryo-EM-validated atomic accuracy.
AI
RFdiffusion (antibody-tuned), Chai-2
Checked
Peer-reviewed in Nature (Baker); preprint (Chai-2); lab-validated
2025-02-19 · microbiology / drug repurposing · confirmed

Google's AI co-scientist independently reproduces an unpublished superbug discovery in 48 hours

Problem
Mechanism of cf-PICI host-range expansion; drug repurposing for AML and liver fibrosis
Result
AI-generated hypotheses matching an unpublished discovery and identifying lab-validated drug candidates.
AI
AI co-scientist (Gemini 2.0)
Checked
Peer-reviewed in Cell (cf-PICI), Advanced Science (fibrosis) and Nature (system, 2026); lab-validated
2024-10-09 · structural biology / protein design · confirmed

Nobel Prize in Chemistry for protein design and AlphaFold

Problem
Protein structure prediction and computational protein design
Result
Nobel Prize in Chemistry 2024: half to David Baker (computational protein design), half to Demis Hassabis and John Jumper (AlphaFold protein structure prediction).
AI
AlphaFold 2, Rosetta/RFdiffusion lineage
Checked
Nobel committee
2024-05-08 · structural biology / drug discovery · confirmed

AlphaFold 3 predicts structures and interactions of all life's molecules

Problem
Predicting 3D structures of biomolecular complexes (protein–ligand, protein–DNA/RNA, antibodies)
Result
Single diffusion-based model predicting joint structures of proteins, nucleic acids, ligands and ions, with at least 50% better accuracy on protein–ligand interactions than prior methods.
AI
AlphaFold 3
Checked
Peer-reviewed in Nature (May 2024); benchmarked on PoseBusters
2023-07-11 · de novo protein design · confirmed

RFdiffusion: diffusion models design new proteins that work in the lab

Problem
Designing proteins with specified shapes and functions from scratch
Result
A general generative model whose protein designs fold and bind as intended at high experimental success rates.
AI
RFdiffusion
Checked
Peer-reviewed in Nature; lab-validated incl. cryo-EM
2020-11-30 · structural biology · confirmed

AlphaFold 2 solves protein structure prediction at CASP14

Problem
Protein folding / structure prediction problem (open since 1972)
Result
Median GDT_TS of 92.4 across CASP14 targets — accuracy comparable to experimental structures for most single-chain proteins; later used to predict 200M+ structures.
AI
AlphaFold 2
Checked
Blind community assessment (CASP14); peer-reviewed in Nature (2021); widely experimentally corroborated
2018-12-02 · structural biology · confirmed

AlphaFold (v1) tops the CASP13 protein-structure prediction assessment

Problem
Protein structure prediction from amino-acid sequence (CASP13 free-modelling targets) (open since 1972)
Result
Ranked first of ~100 groups at CASP13 by predicting inter-residue distance distributions with a deep network and folding by gradient descent on the resulting potential.
AI
AlphaFold 1
Checked
Blind community assessment (CASP13); peer-reviewed in Nature (2020)

chemistry

climate-weather

2026-08-06 · tropical cyclone forecasting · confirmed

DeepMind open-sources WeatherNext 2 and WeatherNext Cyclones with a Nature paper showing ~1 extra day of hurricane warning

Problem
Forecasting tropical-cyclone track, intensity and size
Result
WeatherNext Cyclones' 3-day forecasts are about as accurate as prior systems' 2-day forecasts (>24 h extra lead time); weights released for commercial use.
AI
WeatherNext Cyclones, WeatherNext 2
Checked
Peer-reviewed in Nature (2026); operational evaluation with NHC

computer-science

2025-09-17 · competitive programming / algorithms · confirmed

AI reaches gold-medal level at the ICPC World Finals

Problem
ICPC World Finals 2025 problem set (12 problems)
Result
OpenAI's system solved 12/12 problems (would have ranked 1st); Gemini 2.5 Deep Think solved 10/12, including one no human team solved.
AI
OpenAI reasoning models, Gemini 2.5 Deep Think
Checked
Judged by the ICPC official judging system in a supervised setting
2025-03-12 · machine learning / automated research · confirmed

Sakana's AI Scientist-v2 writes the first fully AI-generated paper to pass peer review (ICLR 2025 workshop)

Problem
Can an AI system autonomously produce a research paper that passes human peer review?
Result
A fully AI-generated ML paper passed peer review at an ICLR 2025 workshop (scores 6/7/6), a first for end-to-end AI-authored research.
AI
The AI Scientist-v2
Checked
Blind peer review at an ICLR workshop; system described in Nature (2026)

materials

2026-09-25 · electrocatalysis · pending

Lila Sciences' AI-run lab screens 2,942 catalysts and finds iridium- and ruthenium-free palladium oxides for green hydrogen

Problem
Iridium/ruthenium-free anode catalysts for acidic oxygen evolution (PEM water electrolysis)
Result
AI-guided high-throughput campaign found six Pd-oxide catalyst families; the best is comparable to Ru with 1,000+ h stability.
AI
Lila Sciences autonomous lab (Bayesian optimisation + LLMs)
Checked
Preprint only (arXiv 2609.30133); not peer-reviewed

mathematics

2026-09-10 · combinatorics / matrix theory / probability / number theory · pending

GPT-6 Astra's Epoch AI run adds more Lean-checked results: Dittert conjecture proved, Ibragimov–Iosifescu and eternal-domination conjectures disproved

Problem
Dittert conjecture; Ibragimov–Iosifescu φ-mixing CLT conjecture; strong n-conjecture (n=4); Gamma–Theta eternal domination conjecture
Result
One proof (Dittert, all n) and three disproofs, each with a Lean formalization or an explicit checkable counterexample.
AI
GPT-6 Astra (pre-release)
Checked
Formal proofs in Lean (mechanically checked). Statements and write-ups mostly not independently audited
2026-09-08 · partial differential equations / fluid dynamics · disputed

OpenAI claims a Millennium Prize problem: 10,000 AI agents prove forced Navier–Stokes blow-up; priority dispute erupts

Problem
Navier–Stokes existence and smoothness (Clay Millennium Prize problem), forced-breakdown case (open since 2000)
Result
Claimed proof, formalised in Lean, of finite-time blow-up for 3D incompressible Navier–Stokes with smooth external forcing.
AI
OpenAI internal model (≈10, 000 parallel agents)
Checked
Formal proof in Lean (public); Clay review pending; human peer review ongoing
2026-09-07 · partial differential equations / fluid dynamics · pending

Caltech team (Anandkumar) reports a stable self-similar singularity candidate for the unforced 3D Euler equations on R³, found with PINNs and LLM help

Problem
Finite-time singularity for the unforced 3D incompressible Euler equations on R³ from smooth initial data
Result
Numerically certified self-similar singular profile plus a (conditional) framework for its nonlinear stability; full rigorous blow-up proof not yet complete.
AI
physics-informed neural networks, OpenAI models and other LLMs
Checked
Interval-arithmetic certification of the profile; partial Lean formalisation; stability conditional
2026-09-03 · probability / percolation theory · pending

Claude-written Lean proof claims the dying percolation conjecture θ(p_c)=0 in every dimension

Problem
Dying percolation conjecture θ(p_c)=0 for Bernoulli bond percolation on Z^d
Result
Claimed Lean-verified proof that θ(p_c)=0 for all d ≥ 2, via a new additive gluing inequality that settles Kozma–Nitzan Conjecture 3.
AI
Claude (Anthropic), Claude Fable 5.1, GPT-5.6 Sol
Checked
Formal proof in Lean (mechanically checked); the statement's fidelity and the informal write-up are not yet refereed
2026-08-30 · analytic number theory · pending

GPT-6 Astra lowers the bounded prime gaps record from 246 to 186

Problem
Bounded gaps between primes (toward the twin prime conjecture) (open since 2014)
Result
Claimed proof that infinitely many pairs of primes differ by at most 186.
AI
GPT-6 Astra
Checked
Formal proof in Lean (announced); not yet independently peer-reviewed
2026-08-10 · analytic number theory · confirmed

Claude proves more than two-thirds of Riemann zeta zeros are simple and on the critical line (up from 41.6%)

Problem
Proportion of nontrivial zeros of ζ(s) on the critical line (toward the Riemann hypothesis)
Result
Unconditional proof that more than 67.2% of zeta zeros are simple and lie on the critical line.
AI
Claude (unreleased research model) in Claude Code
Checked
Formal proof in Lean (key results); expert review (Conrey, Goldston); independent re-proof
2026-08-05 · harmonic analysis / time-frequency analysis · confirmed

HRT conjecture (1996) disproved: 12 time-frequency shifts of a Schwartz function are linearly dependent, found with ChatGPT-assisted guesswork

Problem
Heil–Ramanathan–Topiwala (HRT) conjecture on linear independence of time-frequency shifts (open since 1996)
Result
Explicit counterexample: 12 time-frequency shifts of a Schwartz function are linearly dependent, disproving the HRT conjecture.
AI
ChatGPT
Checked
Handwritten proof with certified numerics; expert-checked (Tao digest); not formalised in Lean; peer review pending
2026-08-01 · group theory, operator algebras, combinatorics, complexity theory, coding theory · confirmed

OpenAI's unreleased 'Astra' model claims ten advances in maths and theoretical CS, with Lean proofs

Problem
Ten open problems incl. existence of explicit non-sofic groups, Connes rigidity, Erdős #146/#180/#183, sphere-packing bounds
Result
Claimed resolutions or improvements on ten open problems, most with Lean-formalised proofs.
AI
Astra (GPT-6 Astra)
Checked
Formal proofs in Lean for most results; independent human audit found no remaining substantive error in principal results
2026-07-24 · algebraic geometry / polynomial maps · pending

Hessian conjecture refuted in five variables, derived from Claude-found Jacobian counterexample

Problem
Hessian conjecture: a polynomial whose Hessian determinant is a nonzero constant has an injective gradient map
Result
Counterexample for n=5 (hence all n≥5); status now: true for n≤3, false for n≥5, open for n=4
AI
Claude Fable 5 (indirectly, via the Jacobian counterexample)
Checked
arXiv preprint; explicit and checkable
2026-07-23 · olympiad problem solving · confirmed

AI systems score a perfect 42/42 at IMO 2026, officially graded

Problem
International Mathematical Olympiad 2026 problems (Shanghai)
Result
First officially graded perfect AI scores at the IMO: Huawei 'Celia' and RedNote 'dots-note-3.0' each solved all six problems for 42/42; several other labs self-reported 42/42.
AI
Huawei Celia, RedNote dots-note-3.0
Checked
Graded by IMO organisers (for Celia and dots-note-3.0); other 42/42 claims self-administered
2026-07-20 · algebraic geometry / polynomial automorphisms · confirmed

Claude Fable 5 finds a counterexample to the Jacobian conjecture in dimension 3

Problem
Jacobian conjecture (Keller 1939): a polynomial map with non-zero constant Jacobian determinant is invertible (open since 1939)
Result
Explicit non-injective polynomial map C³→C³ with constant Jacobian determinant, disproving the conjecture for all n≥3.
AI
Claude Fable 5
Checked
Directly computer-checkable; confirmed independently by multiple mathematicians; formal peer review pending
2026-05-27 · additive combinatorics · confirmed

Erdős–Szemerédi sum-product conjecture shown false over the reals; a GPT-5.5 Pro agent re-disproves it in 7 of 8 runs

Problem
Erdős–Szemerédi sum-product conjecture (over R) (open since 1983)
Result
Finite sets of reals with both sumset and product set of size at most |A|^(2−c), disproving the conjecture over R; later reproduced autonomously by an AI agent.
AI
GPT-5.5 Pro (in the follow-up)
Checked
Human proof (preprint); AI proofs checked by authors
2026-05-20 · discrete geometry · confirmed

OpenAI model disproves Erdős's 80-year-old unit distance conjecture

Problem
Erdős unit distance conjecture (planar point sets: at most N^(1+o(1)) unit distances) (open since 1946)
Result
Construction of planar N-point sets with at least N^(1+δ) unit distances for a fixed tiny δ>0, disproving Erdős's conjectured upper bound, via algebraic number theory; humans improved the exponent within weeks.
AI
OpenAI internal reasoning model
Checked
Expert-checked (Timothy Gowers and others); human follow-up papers
2026-03-10 · Ramsey theory · confirmed

AlphaEvolve improves lower bounds for nine classical Ramsey numbers

Problem
Lower bounds for classical two-colour Ramsey numbers R(3,k), R(4,k)
Result
Explicit colourings improving nine long-studied Ramsey lower bounds.
AI
AlphaEvolve
Checked
Explicit constructions checkable by computer; arXiv preprint
2026-02-11 · combinatorics / number theory (plus physics, CS) · confirmed

DeepMind's Aletheia agent and Gemini Deep Think report autonomous Erdős solutions and new physics and CS results

Problem
Open Erdős problems; conjecture in online optimisation; cosmic-string radiation calculations
Result
Several Erdős problems solved autonomously, a decade-old online-optimisation conjecture refuted, and a new analytic technique for cosmic-string radiation.
AI
Aletheia, Gemini Deep Think
Checked
Preprints; expert-checked; one generalisation peer-reviewed
2025-12-08 · combinatorics / number theory · confirmed

Genuine AI-assisted solutions to Erdős problems begin: #124 (Aristotle), #1026 (48-hour human–AI collaboration)

Problem
Erdős problems #124, #367, #707, #1026 and others (open since 1975)
Result
First AI-involved genuine new solutions of listed-open Erdős problems, including a full solution of #1026 and a Lean-verified autonomous proof for a variant of #124.
AI
Aristotle, AlphaEvolve, Gemini Deep Think, GPT-5
Checked
Formal proofs in Lean for key steps; expert-checked (Tao, Bloom)
2025-11-05 · analysis / combinatorics / geometry · confirmed

Tao, Gómez-Serrano, Georgiev and Wagner test AlphaEvolve on 67 maths problems

Problem
Broad battery of optimisation-type open problems (e.g. inequalities, packings, finite-field Kakeya-type constructions)
Result
Systematic evidence that LLM-driven evolutionary search matches or beats best-known constructions across dozens of problems.
AI
AlphaEvolve, Gemini Deep Think, AlphaProof
Checked
Constructions verifiable; preprint
2025-07-21 · olympiad problem solving · confirmed

AI systems reach gold-medal level at the International Mathematical Olympiad

Problem
International Mathematical Olympiad 2025 problems
Result
Gemini Deep Think (officially graded) and an experimental OpenAI reasoning model each solved 5 of 6 problems for 35/42 — gold-medal standard — in natural language within the 4.5-hour limits.
AI
Gemini Deep Think, OpenAI experimental reasoning model
Checked
Gemini: certified by IMO coordinators; OpenAI: graded by three former IMO medallists (not officially coordinated)
2025-05-14 · combinatorics / algorithms / geometry · confirmed

AlphaEvolve: Gemini-powered agent discovers new algorithms

Problem
4×4 complex matrix multiplication (Strassen 1969: 49 multiplications); kissing number in 11 dimensions; ~50 open problems (open since 1969)
Result
48 scalar multiplications for 4×4 complex matrices; kissing configuration of 593 spheres in 11D (previous 592); matched SOTA on ~75% and improved ~20% of 50+ problems.
AI
AlphaEvolve, Gemini 2.0 Flash, Gemini 2.0 Pro
Checked
Constructions verified computationally (independent GitHub checks); white paper, later arXiv
2024-07-25 · olympiad problem solving / formal proof · confirmed

AlphaProof and AlphaGeometry 2 reach IMO silver-medal standard

Problem
International Mathematical Olympiad 2024 problems
Result
Solved 4 of 6 IMO 2024 problems (28/42, one point below gold) with machine-checked Lean proofs (AlphaProof) and AlphaGeometry 2, including the hardest problem (P6).
AI
AlphaProof, AlphaGeometry 2
Checked
Formal proof in Lean; graded by IMO medalists Timothy Gowers and Joseph Myers
2021-12-01 · knot theory / representation theory · confirmed

DeepMind and mathematicians use machine learning to guide new theorems in knot theory and representation theory

Problem
Combinatorial invariance conjecture for Kazhdan–Lusztig polynomials; relations between knot invariants
Result
ML-guided discovery of a conjectured, then proved, relation between the knot signature and hyperbolic invariants, and a new approach to combinatorial invariance for symmetric groups.
AI
supervised neural networks with gradient saliency
Checked
Peer-reviewed in Nature; human proofs

medicine

2026-09-10 · drug discovery / pulmonary fibrosis · pending

First Phase III trial of a generative-AI-discovered drug doses first patient (Insilico's rentosertib)

Problem
Idiopathic pulmonary fibrosis (IPF) therapy via a novel target
Result
AI-identified target (TNIK) and AI-designed molecule (rentosertib) reached Phase III after a Phase IIa in Nature Medicine showing +98.4 mL mean FVC at 12 weeks on 60 mg vs a decline on placebo.
AI
Insilico Pharma.AI (PandaOmics, Chemistry42)
Checked
Phase IIa peer-reviewed in Nature Medicine (June 2025); Phase III ongoing
2025-08-14 · antibiotic design · confirmed

Generative AI designs new antibiotics that kill drug-resistant gonorrhoea and MRSA

Problem
Designing entirely new antibiotic chemotypes against resistant bacteria
Result
De novo AI-generated molecules with novel mechanisms, effective against drug-resistant gonorrhoea (in vitro) and MRSA (in mice).
AI
generative chemistry models (CReM, F-VAE) with GNN property predictors
Checked
Peer-reviewed in Cell; lab-validated
2025-01-15 · toxinology / protein therapeutics · confirmed

AI-designed proteins neutralise deadly snake-venom toxins and protect mice

Problem
Neutralising snake-venom three-finger toxins, poorly handled by existing antivenoms
Result
De novo designed proteins that neutralise lethal toxins in mice.
AI
RFdiffusion, ProteinMPNN
Checked
Peer-reviewed in Nature; lab-validated in mice
2020-02-20 · antibiotic discovery · confirmed

Deep learning discovers halicin, a structurally new broad-spectrum antibiotic

Problem
Finding new antibiotic classes against drug-resistant bacteria
Result
Discovery of halicin, a broad-spectrum bactericidal compound unlike existing antibiotics, validated in mice.
AI
Chemprop message-passing neural network
Checked
Peer-reviewed in Cell; lab-validated in vitro and in mice

other

2025-10-22 · meta-science / automated research · confirmed

Agents4Science 2025: first conference where AI must be first author and reviewer

Problem
How good is AI-authored and AI-reviewed science?
Result
A full conference cycle with AI first authors and LLM reviewers: 48 of 315 submissions accepted; organisers published an analysis of AI reviewer behaviour.
AI
GPT-5, Gemini 2.5, Claude Sonnet 4, various author agents
Checked
Conference proceedings and analysis paper (arXiv 2511.15534)

physics

2025-09-04 · gravitational-wave detection / control · confirmed

DeepMind's Deep Loop Shaping cuts LIGO control noise 30–100×

Problem
Low-frequency control noise limiting LIGO's sensitivity
Result
Learned mirror-control policy reducing control noise by one to two orders of magnitude on real hardware.
AI
Deep Loop Shaping (RL)
Checked
Peer-reviewed in Science; hardware demonstration
2022-02-16 · nuclear fusion / plasma control · confirmed

Deep reinforcement learning controls fusion plasma in the TCV tokamak

Problem
Magnetic confinement and shaping of tokamak plasmas
Result
First deep-RL controller to shape and sustain diverse plasma configurations on a real tokamak.
AI
deep reinforcement learning (MPO actor-critic)
Checked
Peer-reviewed in Nature; demonstrated on hardware