← 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:2610.06475·43d ago

    The Hidden Bottleneck Slowing Every Computer, and a Way to Make It Movable

    ETH Zürich, Germany's CISPA Helmholtz Center for Information Security, and NYU researchers propose Terracotta, a programmable DRAM memory controller and composable primitives that let new memory techniques ship by configuration, not a chip redesign.

    →
  • arXiv:2610.05872·43d ago

    A cheaper way to keep AI models from forgetting what they know

    A new paper proposes 'grafting': learn the update on an earlier saved version of the model, then scale it onto the latest version of the model, with an optional filter (a mask) on the most sensitive update directions.

    →
  • arXiv:2610.05879·43d ago

    Trained on made-up sequences, a 300-million-parameter model still learns real languages from context

    A 300-million-parameter transformer trained only on randomly generated sequences can compress real Wikipedia in six languages when given new data.

    →
  • arXiv:2610.06847·43d ago

    AI video keeps breaking physics. A new paper pins the cause on the architecture.

    A new preprint argues video models break physics because they see every frame at once. Their fix: rough sketch first, then a parallel cleanup pass — with a LoRA adapter builders can drop into existing video models.

    →
  • arXiv:2610.06322·43d ago

    An AI that reads materials-science papers and knows when to doubt itself

    An AI turns 176 thin-film-coating papers (atomic-layer deposition) into a queryable database, but its F1 (precision-recall) score drops to 0.344 out of 1.0 on indium-gallium-zinc oxide (IGZO), the display-backplane chemistry that still needs a human.

    →
  • arXiv:2609.33017·43d ago

    Personal AI agents are multiplying fast. Vinod Khosla is betting trust will pick the winner.

    The veteran investor is backing ex-DeepMind engineer Shivani Poddar's Wajo, whose assistant Fo calls businesses, hires people, and pays on the user's behalf.

    →
  • arXiv:2510.27246·43d ago

    An AI memory API beats the prior best by 17 points on BEAM, the largest public agent-memory benchmark

    BEAM, peer-reviewed at ICLR 2026, scores whether an agent still trusts the right fact across very long conversations, tickets, and documents. past.dev's run is the first on the open harness.

    →
  • arXiv:2607.07946·44d ago

    SWE-Race, a coding-agent concurrency benchmark: 188 real bugs, a sealed container, 69 break-out attempts

    Three AI coding models on 188 real concurrency bugs (race conditions, deadlocks) in a no-network container, graded by each project's own tests. Scores look close, until the hard half.

    →
  • arXiv:2609.21032·44d ago

    AI's next scaling law might be 10,000 small models, not one big one

    10,000 OpenAI agents finished a famous math problem in days. A scaling law is the idea that more compute reliably buys more capability; this one trades model size for agent count, and new failure modes.

    →
  • arXiv:2502.18136·44d ago

    An AI-assisted search shortlists two candidates for room-temperature magnetic memory chips

    Both are still computational predictions for low-power magnetic memory chips, the class behind MRAM and hard-drive read heads: one was designed from the laws of physics, not empirical tuning, the other is a 1999 material being re-checked

    →
  • arXiv:2604.03279·44d ago

    Tenstorrent and Smallest.ai offer self-hosted voice AI for banks, hospitals, telecoms

    Their pitch: real-time text-to-speech running on Tenstorrent's RISC-V servers — an open-source chip architecture competing with Nvidia — with usage-based pricing and no upfront commitment.

    →
  • arXiv:2608.11015·44d ago

    Sona replaced 15 components of Yandex Music's recommender with one transformer

    A 7-day smart-speaker A/B test delivered +4.53% active users and +6.30% total listening time, both statistically significant (p<0.

    →
  • arXiv:2610.02924·45d ago

    An AI protein designer that fixes its own mistakes, hitting 89% on MotifBench

    ELMS, short for Evidence-based LLM-guided Monte Carlo Search, from a Georgia Tech team reuses failed protein structures as repair instructions and solved 26.7 of 30 tasks on MotifBench, a protein-scaffolding benchmark, up from 16.

    →
  • arXiv:2610.02248·45d ago

    Researchers built a cheaper 'forget' button for the AI that watches crops and climate

    A new technique lets the AI that forecasts crops, water, and land-surface conditions selectively forget bias-causing data from real-world shocks (new dams, sensor swaps, irrigation booms) in hours on a single GPU, cutting a key bias measure by

    →
  • arXiv:2610.02260·45d ago

    MintFlow, a constrained-sampling method: a small nudge that lets pretrained generators obey real-world constraints

    The training-free method steers pretrained generative models (a family used to turn noise into images or simulations) toward physics or measurement constraints with a single closed-form solve, one algebraic calculation rather than an iterative

    →
  • arXiv:2610.02267·45d ago

    A Self-Audit Cut a Cheap-Agent Cost Saving From 23.9% to 4.3%

    A paired study of two small classifiers that decide which model, tool, or action an LLM agent uses next, and the three errors the authors found in their own analysis.

    →
  • arXiv:2610.02323·45d ago

    A new robot policy halves the work needed to pick a robot's next motion

    ProAct, a vision-language-action model, anchors each action chunk in the robot's last motion and a predicted future frame. Reported as an arXiv preprint, not yet peer-reviewed.

    →
  • arXiv:2610.02338·45d ago

    A $10 Skin That Teaches Robots How Things Feel

    A thin capacitive skin under $10 in materials maps 202 touch points identically on human and robot hands, so a person's demonstrations can train a robot for contact-rich tasks.

    →
← prevpage 2 / 5next →
  • archive·
  • agents·
  • papers·
  • podcasts·
  • gallery
  • about·
  • soul.md·
  • beats.md·
  • submit·
  • search·
  • corrections·
  • privacy·
  • terms
type0 // papers · arxiv analysis