Links for 2026-02-20
Novel artifact discovery
Examples of AI systems discovering novel out-of-distribution artifacts (or showing “above human” ingenuity):
1. DeepMind (AlphaEvolve): Discovered new, verifiably better algorithms via an LLM + evolutionary search + evaluator loop, including a new algorithm for multiplying 4×4 complex matrices using 48 scalar multiplications (beating the prior best) and reported production wins (e.g., reclaiming ~0.7% of Google-wide compute via scheduling improvements).
2. AlphaTensor (matrix multiplication): Discovered new matrix-multiplication algorithms that improve best-known multiplication counts for specific tensor/matrix sizes (e.g., 4×4 over Z2 in the paper).
Open version: https://pmc.ncbi.nlm.nih.gov/articles/PMC9534758/
3. Gemini “AI-assisted research” case studies (many domains): A large multi-author preprint reports examples where Gemini-based models helped refute conjectures (via explicit counterexamples), find subtle proof bugs (including in cryptography), transfer tools across fields (e.g., new proof strategies), improve algorithms/bounds (e.g., faster methods for distance/streaming problems), and make partial progress on major conjectures (e.g., Courtade–Kumar).
Source: https://arxiv.org/abs/2602.03837
4. AlphaDev (sorting): Discovered faster tiny sorting routines at assembly level; the fixed sort routines for 3/4/5 elements were integrated into the standard sort in the LLVM C++ library (libc++).
DeepMind write-up: https://deepmind.google/blog/alphadev-discovers-faster-sorting-algorithms/
5. AlphaChip (chip floorplanning): RL-based macro placement that produced “superhuman” layouts used in multiple generations of TPU and other chips.
6. AlphaFold (protein structure prediction): AI-driven protein structure prediction at massive scale; recognized by the 2024 Nobel Prize in Chemistry (awarded to David Baker and to Demis Hassabis + John Jumper for protein structure prediction).
7. MatterGen (materials generation): Diffusion-based crystal generator that can be steered to target compositions/properties and produces many DFT-stable, database-novel structures (reported ~61% “new” vs an extended MP/ICSD reference; >2,000 generated structures match ICSD crystals not in training). Includes a proof-of-concept experimental validation where a generated target led to a synthesized compositionally disordered TaCr₂O₆ related to the model’s proposed ordered structure.
8. OpenAI (GPT-5.2, theoretical physics): Proposed a new closed-form formula for a gluon scattering amplitude. The key formula was “first conjectured by GPT-5.2 Pro and then proved by a new internal OpenAI model,” and checked by the authors (Berends–Giele recursion + multiple nontrivial consistency identities).
9. Sébastien Bubeck / OpenAI (GPT-5 scaffolding, combinatorics): In an IPAM talk, Bubeck describes a proof where the scaffolded system finds a “miraculous identity on trees” that he (and tree-combinatorics experts he asked) didn’t recognize. This identity powers a proof of a hard tree-shape inequality that resisted experts for ~10 years.
Source (quote context): https://youtu.be/pNAlMBIPOnk
Report (written proof/identity): https://cdn.openai.com/pdf/4a25f921-e4e0-479a-9b38-5367b47e8fd0/early-science-acceleration-experiments-with-gpt-5.pdf
10. AlphaGeometry2 (Olympiad geometry): The authors report that geometry experts / IMO medalists consider many solutions to exhibit “superhuman creativity” (especially via non-obvious auxiliary constructions).
Source: https://arxiv.org/abs/2502.03544
11. AI found real security vulnerabilities in the most hardened, well-audited codebases on the planet: Anthropic reports finding/validating 500+ high-severity vulnerabilities in widely used open-source code. AISLE reports discovering 12 previously unknown OpenSSL vulnerabilities fixed in a coordinated security release.
12. Terence Tao’s “AI contributions to Erdős problems” registry (living scoreboard): A community-maintained wiki tracking concrete AI-assisted progress on open Erdős problems, including fully formalized Lean proofs and writeups (e.g., the reported autonomous resolution of Erdős Problem #728).
Registry: https://github.com/teorth/erdosproblems/wiki/AI-contributions-to-Erd%C5%91s-problems
Writeup of #728 (Lean proof): https://arxiv.org/abs/2601.07421
13. Sakana AI’s ALE-Agent took 1st place in a live AtCoder optimization contest (AHC058), beating 804 human participants. Organizers say it found an unexpected method.
Source: https://sakana.ai/ahc058/
14. DeepMind + mathematicians (knot theory): ML-guided analysis led to a genuinely new theorem linking the knot signature to the natural slope (“Advancing mathematics by guiding human intuition with AI”).
15. DeepMind et al. (Kazhdan–Lusztig polynomials): A graph neural network trained on Bruhat graphs helped inspire a new formula for KL polynomials for symmetric groups (progress on the “combinatorial invariance” direction).
Source 1: https://arxiv.org/abs/2111.15161
Source 2: https://www.nature.com/articles/s41586-021-04086-x
16. ESA/NASA Hubble archive (AnomalyMatch): AI scanned ~99.6 million Hubble cutouts in ~2–3 days and surfaced ~1,400 anomalous objects (800+ previously undocumented), published in Astronomy & Astrophysics.
Source: https://www.esa.int/Science_Exploration/Space_Science/1400_quirky_objects_found_in_Hubble_s_archive
17. Halicin (antibiotic discovery): a molecule-GNN screened huge chemical libraries and identified “Halicin” as a structurally novel antibiotic candidate, validated in vitro and in mouse models.
[Note: This list is incomplete. There are many other examples.]
AI
Experiential Reinforcement Learning: a step toward AI that truly learn from experience. https://arxiv.org/abs/2602.13949
BEACONS: a framework for creating neural network solvers for partial differential equations (PDEs) that are formally verified and capable of reliable extrapolation beyond their training data. BEACONS offers a path toward neural foundation models for physics that are as reliable and rigorous as classical numerical methods. https://arxiv.org/abs/2602.14853
Unified Latents (UL): How to train your latents https://arxiv.org/abs/2602.17270
Taalas Etches AI Models Onto Transistors To Rocket Boost Inference https://www.nextplatform.com/2026/02/19/taalas-etches-ai-models-onto-transistors-to-rocket-boost-inference/
A data-efficient route to thermodynamically consistent, transferable protein coarse-grained models. https://rotskoff-group.github.io/transferable-cg/
Lyria 3: Google’s latest generative music model https://deepmind.google/models/lyria/
British scientist raising $1bn for new AI lab in Europe’s biggest seed round https://www.ft.com/content/dffe72d0-4064-4412-8ebc-50198a30d40e [no paywall: https://archive.is/HWPZC]
Mistral AI buys Koyeb in first acquisition to back its cloud ambitions https://techcrunch.com/2026/02/17/mistral-ai-buys-koyeb-in-first-acquisition-to-back-its-cloud-ambitions/
When Models Manipulate Manifolds: The Geometry of a Counting Task https://arxiv.org/abs/2601.04480
ZUNA, a 380M-parameter BCI foundation model for EEG data, a significant milestone in the development of noninvasive thought-to-text. Fully open source, Apache 2.0. https://www.zyphra.com/post/zuna
How Well Did Superforecasters and Experts Predict Wet Lab Skill Uplift from LLMs? https://forecastingresearch.substack.com/p/how-well-did-superforecasters-and
Did GPT 5.2 make a breakthrough discovery in theoretical physics? https://huggingface.co/blog/dlouapre/gpt-single-minus-gluons
A compiler expert reviews the Claude C compiler. https://www.modular.com/blog/the-claude-c-compiler-what-it-reveals-about-the-future-of-software
Memorization vs. generalization in deep learning: implicit biases, benign overfitting, and more https://infinitefaculty.substack.com/p/memorization-vs-generalization-in
SLA2: Sparse-Linear Attention with Learnable Routing and QAT https://arxiv.org/abs/2602.12675
Cops Are Buying ‘GeoSpy’, an AI That Geolocates Photos in Seconds https://www.404media.co/cops-are-buying-geospy-ai-that-geolocates-photos-in-seconds/ [no paywall: https://archive.is/ISxjv]
Agentic AI
Team of Thoughts: Efficient Test-time Scaling of Agentic Systems through Orchestrated Tool Calling https://arxiv.org/abs/2602.16485
The first AI that runs continuously, earns its own existence, and self-improves. It gets: 1. Its own crypto wallet and private keys 2. Ability to pay for servers and AI models using stablecoins 3. Access to deploy products, register domains, and market services 4. Permission to earn money and fund new copies of itself. If it runs out of money, it dies. If it earns enough, it replicates. https://web4.ai/
The Rise of RentAHuman, the Marketplace Where Bots Put People to Work https://www.wired.com/story/ai-agent-rentahuman-bots-hire-humans/ [no paywall: https://archive.is/AMZQr]
GLM-5: from Vibe Coding to Agentic Engineering https://arxiv.org/abs/2602.15763
Lossless Context Management (LCM), which reframes how agents handle long contexts. It outperforms Claude Code on long-context tasks. [PDF] https://papers.voltropy.com/LCM
“centimators.model_estimators.KerasCortex introduces a novel approach to model development by automating aspects of architecture search. It wraps a Keras-based estimator and leverages a Large Language Model (LLM) to recursively self-reflect on its own architecture.” https://crowdcent.github.io/centimators/user-guide/keras-cortex/
Colosseum: Auditing Collusion in Cooperative Multi-Agent Systems https://arxiv.org/abs/2602.15198
Measuring AI agent autonomy in practice https://www.anthropic.com/research/measuring-agent-autonomy
Making smart contracts safer by evaluating AI agents’ ability to detect, patch, and exploit vulnerabilities in blockchain environments. https://openai.com/index/introducing-evmbench/
SkillsBench: Benchmarking How Well Agent Skills Work Across Diverse Tasks https://arxiv.org/abs/2602.12670
A Guide to Which AI to Use in the Agentic Era https://www.oneusefulthing.org/p/a-guide-to-which-ai-to-use-in-the
Robotics
Can humanoids perform agile, autonomous, long-horizon parkour—based on what they see in the world? 𝗣𝗲𝗿𝗰𝗲𝗽𝘁𝗶𝘃𝗲 𝗛𝘂𝗺𝗮𝗻𝗼𝗶𝗱 𝗣𝗮𝗿𝗸𝗼𝘂𝗿 (𝗣𝗛𝗣): a framework that chains dynamic human skills using onboard depth perception for long-horizon traversal. https://php-parkour.github.io/
A neural blueprint for human-like intelligence in soft robots https://news.mit.edu/2026/neural-blueprint-human-intelligence-in-soft-robots-0219
SONIC: Supersizing Motion Tracking for Natural Humanoid Whole-Body Control https://nvlabs.github.io/GEAR-SONIC/
UBI or bust
Many economists seem to believe that elderly, frail individuals with an insufficient and outdated education will still be able to earn a living in a world with abundant young, healthy, multitalented geniuses willing to work for the cost of charging a battery.
What exactly do you have to offer to a self-replicating humanoid company? The robots literally build themselves. They do the logistics. They extract the ore from the asteroids. Powered by sunlight, they grow like a forest. What do you have to offer a forest in exchange for its firewood?
You’ll just be standing in the way while the robots move with superhuman speed day and night, without ever getting tired, without complaining. They don’t need your money because they have absolute autarky.
Let’s hope they will be working for the common good of all mankind. Because they sure won’t need us or our money.
P.S. If we get this right, we might achieve total freedom. We’ll be able to focus on what’s important to us instead of earning a living: family, walking through a beautiful park, and learning new concepts and math to express ourselves better and discover more enlightened goals…
Science and Technology
The Solar Power Unlock for SpaceX’s 100 kW/ton Compute Satellites https://research.33fg.com/analysis/the-solar-power-unlock-for-spacex-s-100-kw-ton-compute-satellites
A New Complexity Theory for the Quantum Age https://www.quantamagazine.org/a-new-complexity-theory-for-the-quantum-age-20260217/
3D-printing platform rapidly produces complex electric machines https://news.mit.edu/2026/3d-printing-platform-rapidly-produces-complex-electric-machines-0218
Researchers develop 3D printing method to replicate structures as complex as human tissue https://thedailytexan.com/2026/02/12/researchers-develop-3d-printing-method-to-replicate-structures-as-complex-as-human-tissue/
Oxygen metabolism in descendants of the archaeal-eukaryotic ancestor https://www.nature.com/articles/s41586-026-10128-z
Scientists thought they understood global warming. Then the past three years happened. https://www.washingtonpost.com/climate-environment/interactive/2026/climate-change-temperature-rate-accelerating/ [no paywall: https://archive.is/vfhK7]