2026年8月10日 技术热点总结

项目简介：let your agent control your phone

技术栈：Python

核心介绍：Connect an LLM directly to your real iPhone with a thin, editable harness. No jailbreak, no Xcode, no WebDriverAgent. The Mac’s iPhone Mirroring window is the whole transport: screencapture + Vision-framework OCR for eyes, HID-level CGEvents for hands. Nothing between the agent and the phone. The agent writes what’s missing during execution in agent-workspace/agent_helpers.py.

项目数据：⭐ 827 Stars，🍴 64 Forks

项目简介：[Official Repo] JoyAI-Video-Edit: Real-Time Open-Ended Video Editing with Autoregressive Diffusion

技术栈：Python

核心介绍：JoyAI-Video-Edit Real-Time Open-Ended Video Editing with Autoregressive Diffusion

项目数据：⭐ 601 Stars，🍴 25 Forks

项目简介：An interactive five-chapter night walk through a Kyoto mountain temple, rendered live in Three.js.

技术栈：HTML、creative-coding、generative-art、interactive-storytelling、japanese-design、threejs、webgl

核心介绍：An interactive five-chapter night walk through a Kyoto mountain temple, rendered live in Three.js and layered with cinematic generated imagery. View the live project · View the source · Read the build prompt Kage is a deliberately small static site. index.html contains the document structure, CSS, procedural scene construction, scroll choreography, and interaction …

项目简介：Create smooth, responsive interactive web animations.

技术栈：Python

核心介绍：Oil Motion 是一个 Agent 通用的交互动画 Skill。它负责设计动作、生成连续画面、整理动画资源，并把最终效果接入页面滚动、鼠标、拖动、触摸或设备方向。 使用时只需要说明希望表达什么、提供哪些素材，以及动画应该跟随什么操作。具体的生成、检查、压缩和前端实现由 Agent 完成。

项目数据：⭐ 571 Stars，🍴 57 Forks

Selecting a complete 3D object from a reconstructed scene with minimal user effort is essential for practical scene editing and embodied interaction. Existing 3DGS-based methods either retrain the Gaussian representation to embed per-object labels, or build dense multi-view SAM observations, both requiring heavy computation and dense viewpoint coverage that is rarely available in practice. We present GaussianSelector, a training-free framework for interactive 3D object selection from sparse views and sparse scribble guidance. Operating directly on native Gaussian primitives, we coarsen dens…

Robot learning is splitting into two bets: policies that bake competence into frozen weights (vision-language-action, or VLA, models), and agents that write and refine their own executable skills as code. This survey organises the field around that axis of weights versus skills. Its central analytical contribution is a deep-dive that arranges code-as-policy methods by their degree of self-improvement, from zero-shot program synthesis, through closed-loop self-repair and persistent skill memory, to the sparsely populated cell in which execution feedback, skill memory, and evolutionary search…

World models have attracted significant attention for their ability to capture and predict the structure and dynamics of the physical world. In this emerging landscape, Joint Embedding Predictive Architectures (JEPA) offer a particularly compelling direction.
We study a largely unexplored regime: populous, crowded, and chaotic Global South urban environments, which we call DENSEWORLD. Unlike the lower-density, lane-structured settings that dominate existing evaluations, these scenes exhibit soft spatial boundaries, extreme agent heterogeneity, persistent occlusion, and rapid social negoti…

Latent diffusion modeling (LDM), a prominent paradigm, utilizes tokenizers to map input signal to compressed representation. This dependency positions tokenizer as an integral part of generation process itself, since it affects learning speed, quality of synthesized samples and lay foundation for later applications. This report presents series of KVAE tokenizers for audio, image and video, all designed for subsequent text-conditioned generation: KVAE-Audio, a continuous full-band 48 kHz tokenizer with a 50 Hz latent of 64 channels; KVAE-3D — two causal video tokenizers for 4x16x16 and 4x8x…

Computer-use agents pay full frontier inference to re-derive routines their user has already performed, because an agent's memory today records what the user said, not what the user did. We compile passively captured screen activity into agent memory with a deterministic, zero-model pipeline: it segments a local capture stream into typed activity frames, bounded episodes carrying application, site, timing, input volume, and evidence pointers back to the raw rows, with no model in the loop, so the output is byte-identical, cacheable, and mechanically auditable. On one professional's single-u…

Data agents enable natural-language analytics over organizational workspaces, where relevant evidence may be scattered across databases, structured files, long documents, and multimedia. Existing benchmarks largely isolate structured querying, retrieval, or open-ended analysis, leaving heterogeneous evidence discovery, complete tabular outputs, and deterministic evaluation insufficiently unified. We introduce DataSpace, a benchmark in which data agents produce verifiable tabular results from task-local heterogeneous workspaces. It contains 410 cross-language tasks and 7,439 artifacts totali…