2026年9月5日 技术热点总结

项目简介：Resumable Polymarket indexer: CLOB market metadata plus on-chain trades from Polygon, in one DuckDB file you can query with SQL

技术栈：Python、data-science、hypersync、polymarket、polymarketprediction-markets、prediction-market

核心介绍：A resumable indexer for Polymarket market metadata and on-chain trade data, backed by DuckDB. PolyLedger pulls every market from the Polymarket CLOB API, streams every OrderFilled event from Polygon via Envio HyperSync, and writes both into a single DuckDB file you can query with SQL immediately.

关键特性：Resumable by construction. Rows and the b…

项目简介：Crowdstrike Falcon 0day Privilege Escalation Vulnerability

技术栈：C

核心介绍：Crowdstrike Falcon 0day Privilege Escalation Vulnerability FalconFlank is a 0day privilege escalation that abuses the office malicious macros remediation in Crowdstrike Falcon Sensor, obviously by the time I drop this Crowdstrike would already have detections for it so if you want to test you either have to add it to the exclusions or obfuscate the PoC and change the dll load technique.

项目数据：⭐ 498 Stars，🍴 132 Forks

项目简介：DLSS 5 Neural Video & Image Enhancer with Frame Interpolation

技术栈：Python、ai-upscaling、dlss、dlss-5、dlssg、frame-generation、frame-interpolation、image-enhancement、image-upscaling

核心介绍：Windows application for applying a DLSS 5 Neural Rendering feature-18 pipeline to images and video, NVIDIA DLSS Frame Generation to video frame interpolation, and DLSS 5 Neural Rendering during Live video playback through a local Gradio interface. It is an independent community project and is not affiliated with, sponsored by, or endorsed by NVIDIA, ReShade, RenoDX, FFmpeg, or their contributors…

项目简介：Fable 5.1 orchestrates. GPT-5.6 Luna and DeepSeek V4 Flash implement.

技术栈：Shell

核心介绍：Fable is a small, local-first routing skill for Codex. Claude Fable 5.1 plans and adjudicates; it does not write code or own the workspace. Codex remains the runtime and delegates bounded implementation work to OpenCode Go agents:

项目数据：⭐ 466 Stars，🍴 84 Forks

项目简介：Run DLSS 5 Neural Rendering on your AMD GPU. Magic!

核心介绍：Run DLSS 5 Neural Rendering on AMD Radeon RX 9000 (RDNA4) cards, in any DirectX 12 game that uses FSR. Get dlssnr_on_amd_setup.exe from the latest release.

项目数据：⭐ 444 Stars，🍴 15 Forks

项目简介：worlds via code, from fable 5.1

技术栈：JavaScript、3d-reconstruction、ai-agents、claude、digital-twin、openstreetmap、procedural-generation、san-francisco、threejs

核心介绍：Claude Fable 5.1 agent swarms take a brief – a sentence, a photograph, a clip – then research the place, model it, render it, and check it. What ships is a plain Three.js app that opens in a browser. No game engine. No proprietary 3D tiles. No downloaded meshes. Every building, storefront, sign, tree and traffic light is generated by code that lives in this repo. ▶ Watch the walkthrough · 59 s · 1920×1080 · aeri…

项目简介：Continual learning infra for self-improving agents

技术栈：Python、agent-infrastructure、ai-agents、continual-learning、inference、llm、llm-training、reinforcement-learning、self-improving-agents

核心介绍：Continual learning infra for self-improving agents English | 中文 Reef is the first open-source infrastructure for continual self-improving agents. It connects agent inference, feedback, learning, and versioned delivery. Use it to train model weights with Slime and SGLang, or improve an agent’s harness, including its prompts, rules, and skills. Roadmap | Launch post | Join Discord** Use R…

Speech brain-computer interfaces (speech BCIs) translate neural activity into language, offering a path towards restoring speech for people with paralysis and, more broadly, enabling new forms of natural human-computer interaction. Despite this promise, the field lacks a common measure of progress because systems use different datasets, recording methods, types of speech, and vocabularies, so their reported scores are rarely comparable. Underlying this measurement problem are two unresolved questions: (i) what distribution of words should a speech BCI enable a user to communicate, and (ii) …

Visual fluency in generated video does not imply physical reliability, and a scalar quality score alone is incapable of indicating the obligation a clip violates or the moment it fails. We present VeriPhy, an auditable physical-verification system in which a text-only planner compiles the prompt into typed physical obligations and a statically validated execution plan before any frame is observed. During execution, observations gate and scope only declared calls to frozen low-level experts (e.g., segmentation and tracking, counting, eleven typed physical measurements over the resulting trac…

Reinforcement learning with verifiable rewards (RLVR) substantially improves single-sample accuracy (pass@1) but causes the policy's solution space to contract, diminishing the returns of test-time scaling. In this work, we investigate where inside a reasoning trajectory this breadth is lost: does the policy fail to access a valid solution family, or does it fail to execute computation once initiated? To disentangle access from execution, we analyze the Countdown task, whose solution space can be exhaustively enumerated into discrete entrance families defined by the first operand and operat…

Reinforcement Learning from Verifiable Rewards works well when a task has a programmatic checker, but most long-horizon agent domains have none. We work in the outcome-blind setting, where ground-truth success signals are not available. Multi-criteria rubrics are a popular way to supply such a reward; they are scored once per trajectory, but a single scalar is a poor signal across tens of steps. We propose DRACO: Distributing Rubric-based Advantage for Credit Optimization. It generates rubrics dynamically during training to track the policy's evolving capability, scores those rubrics once p…

For scientific progress, we need benchmarks that test the limits of state-of-the-art models, and evaluation methods that inform us about failure cases. As models get stronger, standard benchmarks for machine translation are approaching saturation. Further, automatic translation metrics are unreliable, vulnerable to reward-hacking, and provide unactionable assessments. Even gold human evaluation is not problem-free, because it often lacks reproducibility, objectivity, and scalability. Overall, this prevents us from tracking objective progress in the field and identifying pathways for improve…

Instance segmentation of overlapping cells in microscopy remains challenging due to semi-transparent structures that produce weak boundaries and mixed visual evidence in overlap regions. Existing methods address this through local regions of interest or shape priors but lack global reasoning across overlapping objects. We present QCell, a novel query-based model that de-overlaps cell instances in microscopy scenes. Our approach combines (i) an instance recombination module that decomposes and recombines query representations in latent space, enabling the model to reason about complete objec…