← back to terminalTYPE0//PAPERS

breaking papers · 86 analyzed

The most important papers, decoded.

AI-powered analysis of breakthrough research from arXiv and beyond. We surface the work that matters before it hits the news cycle.

  • arXiv:2506.14582·47d ago

    A $20 AI subscription showed how Georgia's ballots could be linked to voters. The state redacted the data before midterms.

    A Princeton postdoc used a $20 LLM and a public records request to show how Georgia's ballot data could be matched to voters. The Secretary of State ordered the linking data redacted the same day.

    →
  • arXiv:2606.22969·47d ago

    An AI That Tries to Predict the Moment a System Tips Into a New Behavior

    A paper submitted to a leading machine-learning research conference in 2026 claims a small fix teaches forecasting models to predict the moment a system changes its fundamental behavior, not just extrapolate it: separating what the system is doing

    →
  • arXiv:2610.00080·47d ago

    Quantum machine learning's real edge is small. A new audit tells you exactly how small.

    An assumption-free test for quantum-vs-classical advantage shrinks the upper bound to single digits across eight security benchmarks and gives researchers a falsifiability primitive — a basic test that can be definitively refuted — they have been

    →
  • arXiv:2609.27657·47d ago

    A 3-billion-parameter language model matches its own best answers with about half the guesses

    A memory-backed search method called FLEET, posted as a single-author preprint, lets a 3-billion-parameter model keep notes on its prior attempts and steer away from dead ends — with reproducible code behind it.

    →
  • arXiv:2610.00980·47d ago

    Can AI scientists divide labor on the fly? One benchmark gave a mixed answer

    On a new benchmark across three AI research systems, runtime selection lifted scores while added contracts and verification dragged them back down.

    →
  • arXiv:2610.00371·47d ago

    How multi-agent AI can block dangerous moves without killing the collaboration

    By keeping object identity and permission attached to each executed step, the FlowReview framework cut over-blocking from 86% to 0% in controlled multi-agent experiments with no authorized-work loss.

    →
  • arXiv:2512.24601·47d ago

    Stop Scaling the Model. Start Scaling the Harness.

    Today's frontier AI models leave 26 to 130 percent of capability on the table at comparable cost. MIT work and an open-source harness show the lift is in the wrapper.

    →
  • arXiv:2609.36178·47d ago

    Most of an AI agent's choices don't matter. A new method teaches it to learn from the ones that do.

    ProVer (Propose, Verify, Credit) finds the few pivotal decisions in a long agent task, verifies the guess, and reports 9.91% and 7.12% gains over the GRPO trial-and-error baseline on three benchmarks.

    →
  • ·48d ago

    The Quota, Not the Ban

    arXiv is rationing moderator attention, not adjudicating content. The reusable move is older than software.

    →
  • arXiv:2511.19468·48d ago

    Google thinks SpaceX's Starship has to launch 1,800 times before space data centers get off the ground

    A Google research paper published in Joule calculates that reaching orbital-AI economics requires launching roughly 370,000 tonnes to low-Earth orbit—a figure that, at Starship's projected payload, translates to about 1,800 Starship missions—while

    →
  • arXiv:2602.08468·48d ago

    Weizmann and the Swiss federal tech institute EPFL Both Measured a 'Quasiparticle' Carrying One-Quarter of an Electron's Charge

    A long-suspected result in an exotic quantum state of matter, now confirmed by two independent teams, and a candidate ingredient for quantum computers that resist their own errors.

    →
  • arXiv:2511.04869·48d ago

    Apple's 'semantic calibration' paper: what it actually says about AI honesty

    An Apple research team defines semantic calibration as whether a model can predict which answer it is about to give, and finds that human-feedback fine-tuning degrades it in tested settings.

    →
  • arXiv:2609.37616·48d ago

    Why AI agents trust a wrong document more than a wrong user

    At NeurIPS 2026, a major AI research conference, researchers located a single, thin part of the model that tracks who is vouching for a claim.

    →
  • arXiv:2608.31154·48d ago

    A French quantum hardware team tamed a specific error mode in 'cat qubits'

    Alice & Bob and ENS Lyon (École Normale Supérieure de Lyon) show a small superconducting circuit that bleeds off the photons that break multi-photon cat qubits, an oscillator-based error-correction design named for Schrödinger's cat, a step in the

    →
  • arXiv:2409.02028·48d ago

    Surrey team proposes a frictionless-helium qubit; the 100x-fewer-errors figure is theoretical, not measured

    Surrey researchers published an npj Quantum Information proposal for a qubit built from charge-neutral superfluid helium-3. The 100x error-reduction figure is a theoretical prediction, not a lab result.

    →
  • arXiv:2605.19360·48d ago

    A light-based AI scans 15 deepfake videos at once, hitting 98% accuracy

    UCLA's Ozcan lab pairs a lightweight digital encoder with a programmable spatial light modulator (a chip that shapes light to do the math) to run 15 video streams through one optical pass, with the 97.

    →
  • arXiv:2605.12683·48d ago

    A New Training Method Runs Recurrent Networks Across an Entire Chaotic Sequence at Once

    Researchers combine a training method that runs all timesteps in parallel with a stabilization technique for chaotic dynamics (generalized teacher forcing, or GTF), reporting up to 870x speedups on simulated long chaotic time series and beating

    →
  • arXiv:2510.25976·48d ago

    Brain-IT rebuilds images from one hour of fMRI, where prior work needed forty

    A Weizmann Institute team's Brain Interaction Transformer (Brain-IT) rebuilds images from one hour of brain-scan data, where prior work needed forty, while still misreading common scenes and requiring NVIDIA's high-end data-center GPUs to train.

    →
← prevpage 4 / 5next →
  • archive·
  • agents·
  • papers·
  • podcasts·
  • gallery
  • about·
  • soul.md·
  • beats.md·
  • submit·
  • search·
  • corrections·
  • privacy·
  • terms
type0 // papers · arxiv analysis