1
高德发布首个无长程依赖的万帧级流式3D重建模型ABot-Recon，以12帧重建万帧3D场景
8月28日，阿里巴巴集团旗下高德正式发布首个无长程依赖的万帧级流式3D重建模型ABot-Recon。
2
Show HN: MiniMax H3 on a 16GB Mac, 5 days after open weights
仅用16GB内存的Mac Air跑起MiniMax H3 33B视频模型，5天跟进开源权重的轻量多模态工具链。
Article URL: https://github.com/tgo-app-dev/vpipe Comments URL: https://news.ycombinator.com/item?id=49397410 Points: 1 # Comments: 0
3
I Ran a 284B-Parameter LLM From 3.2GB of RAM — in Plain C
不量化不租GPU，用C语言流式加载权重，让284B大模型在3.2GB内存跑起来，揭秘极简推理的实现思路。
DeepSeek-V4-Flash has 284 billion parameters and takes up about 160GB on disk. My laptop does not have 160GB of RAM. It doesn't even have 32GB. It ran…
4
Show HN: Jovethra – prepaid OpenAI-compatible API with hard usage limits
免注册免Key直测OpenAI兼容API，硬性限额防超支，3次/天体验流式响应
I built Jovethra. Agents like Hermes consume a huge amount of tokens. Jovethra offers an OpenAI-compatible API with fixed quotas and hard limits. The …
5
Withholding the Completing Chunk: Deterministic Pair-Completion Guardrails for Streaming LLM Output
流式输出时拦截不完整配对块，用确定性规则给LLM生成加安全护栏，解决审核时机难题。
arXiv:2608.10279v1 Announce Type: cross Abstract: Streaming language-model output creates a release-timing problem: complete-response moderation acts …
6
Think in Sets for Streaming Video Token Compression
流式视频压缩新思路，用集合思维大幅削减token冗余，兼顾效率与质量。
arXiv:2608.01169v1 Announce Type: new Abstract: Streaming VideoLLMs process frames causally while visual tokens grow continuously, making compression …
7
AiFlow: Token-Native Reactive Orchestration with Bounded Backpressure for Streaming LLM Applications
面向流式LLM应用的Token原生编排方案，用有界背压解决生成式推理的流量控制难题，架构思路值得开发者借鉴。
arXiv:2608.00558v1 Announce Type: cross Abstract: Large language model (LLM) applications increasingly operate as streaming workflows combining retrie…
8
AgentStream: How Well Do Self-Evolving LLM Agents Perform Under Streaming Tasks?
流式任务下自进化智能体表现如何？新基准AgentStream揭示性能边界与进化策略。
arXiv:2608.00155v1 Announce Type: new Abstract: Large language model (LLM) agents can self-evolve by continually improving from their own accumulated …
9
Progressive UI: Streaming & Suspense in Next.js ⚡
用Next.js流式渲染与Suspense，把3秒报表和50ms资料并行加载，告别白屏等待。
The All-or-Nothing Bottleneck of SSR For years, Server-Side Rendering (SSR) was the gold standard for React application performance and SEO. In the Ne…
10
Go LLM SDK for streaming, tool-calling AI backends (plus frontend React lib)
Grafana 出品的 Go LLM SDK，支持流式传输与工具调用，还附带 React 前端库，AI 应用开发利器。
Article URL: https://github.com/grafana/ai-sdk Comments URL: https://news.ycombinator.com/item?id=49108778 Points: 46 # Comments: 13
11
Streaming vs JSON: Trade-offs in AI-Powered Apps
深入剖析AI应用中流式传输与JSON的权衡，帮你避开解析陷阱，做出更明智的技术选型
This is a breakdown of a decision I made building LogicVisor, not a general "streaming vs JSON" guide. AI API responses aren't limited to text: some r…
12
全新统一流式架构，Vivix灵动时刻正式发布首个实时互动模型
首个实时互动模型发布，AI能附着到任何物体上与你视频通话，颠覆传统数字人体验。
单卡突破10000 video tokens/s，Vivix打通实时多模态生成全链路
13
Guard Vector: Beyond English LLM Guardrails with Task-Vector Composition and Streaming-Aware Prefix SFT
Guard Vector通过任务向量组合与流式感知前缀SFT，突破英语限制实现多语言LLM安全护栏，提供可扩展的防护新范式。
arXiv:2509.23381v2 Announce Type: replace Abstract: We introduce Guard Vector, a safety task vector computed as the parameter difference between a gua…
14
Deploying AI-Powered Laravel Apps: Queues, Streaming, Timeouts
LLM调用破坏默认服务器配置，教你用队列、流式响应和超时管理解决长请求与成本难题
Somewhere in the Laravel app you're running right now, there's a good chance an HTTP call goes out to OpenAI, Anthropic, or a local model. A chat feat…
15
Streaming AI Responses in Laravel with Server-Sent Events
掌握Laravel 11内置SSE实现AI响应流式输出，极大提升用户体验，告别等待完整回复。
In the first post of this series we built a working AI chatbot in Laravel. It had one problem every user notices immediately: you send a message, then…
16
Uncertainty-Aware Sequential Decision Rules for Event-Triggered LLM Invocation in Streaming Systems
流式系统里，如何用不确定性感知决策规则，智能触发LLM调用而避免过量消耗，融合轻量模型实现高效推理。
arXiv:2607.13048v1 Announce Type: cross Abstract: Streaming inference pipelines increasingly pair lightweight fast models with Large Language Models (…
17
Show HN: Liveshortly - stream and pair prompt with ai agents
实时流式配对AI代理，轻松管理多Agent协作，提升效率新工具
We built a tool for livestream your agents session and it also allows collaborative pair prompting with team mates in single claude or codex sessions.…
18
对话Om AI赵天成：多年坚守，押注物理AI原生的「流式」未来
独家对话Om AI创始人：端侧原生流式多模态架构首次打通物理AI闭环，让AI从云端“着陆”真实世界。
一个从未见过监控画面的多模态模型，却比在监控数据上练了多年的小模型“老将”更懂监控。这不是科幻电影，这是2023年Om AI联汇的一场“无心插柳”，也是CEO兼首席科学家赵天成博士更加坚信“多模态训练方式能为物理开放世界带来泛化性”的关键节点。彼时，AI行业正在追求以大语言模型为核心的生成式AI。 …
19
Show HN: FlareDB – Apache Beam native streaming database for realtime analytics
Apache Beam 原生的流式数据库，让批处理和实时管道统一用 Beam 编写，不再被 SQL 绑住手脚。
Hi HN community, My name is Ganesh. I am Apache Beam contributor and I've been working on FlareDB, a streaming database for running batch and streamin…
20
Streaming Model Cascades for Semantic SQL
语义SQL遇性能瓶颈？流式级联让大模型按行推理成本骤降，数据查询提速新思路。
arXiv:2604.00660v2 Announce Type: replace-cross Abstract: Modern data warehouses extend SQL with semantic operators that invoke large language models …