Show HN: Tiny-vLLM, a C++ and CUDA LLM Inference Engine
Hacker News (AI keywords)16 days agoNew ToolTiny-vLLM is a Show HN project described as a high-performance LLM inference engine implemented in C++ and CUDA. From the provided title alone, the project appears aimed at developers or ML engineers interested in GPU-accelerated local or server-side inference. No further claims about supported models, benchmarks, APIs, licensing, deployment targets, or production readiness are stated in the source.
CAPTCHAs can still detect AI agents★ 72
Hacker News (AI keywords)16 days agoPaperRoundtable argues that CAPTCHA image recognition is largely solved, but process-level behavior still separates humans from AI agents. Their CogCAPTCHA30 benchmark combines CAPTCHA with cognitive psychology tasks to test not only outputs, but how answers are produced. Results suggest frontier models like Claude, GPT, and Gemini are not necessarily more humanlike than smaller or cognition-trained models.
Has the hunt for AI compute uncovered the next Cerebras?
TechCrunch AI17 days agoHardwareTechCrunch reports that General Compute has raised a $15 million seed round at a $60 million post-money valuation to build an AI inference neocloud. The company is ordering $300 million of SambaNova SN50 chips, betting they can outperform GPUs and rival specialized chips for inference. The story frames inference speed, deployment flexibility, and lower power needs as key battlegrounds in AI infrastructure.
ESMFold2: The Bitter Lesson Is Coming for Proteins★ 74
Latent Space18 days agoCommentaryLatent Space interviews Biohub’s Alex Rives about ESMFold2 and the broader ESM protein modeling stack. The discussion centers on datasets versus inductive bias, and whether protein biology is entering its own Bitter Lesson era. The key implication is that large-scale evolutionary sequence data and open models may become foundations for structure prediction, interaction modeling, and programmable biology.
New AI Infra Decacorns: Fireworks, Baseten, and OpenRouter★ 78
Latent Space18 days agoBusinessAI infrastructure startups Fireworks and Baseten have reportedly reached massive valuations, reflecting intense investor interest in developer-focused inference and deployment platforms. OpenRouter, the popular LLM API aggregator, is also on a rapid growth trajectory. This funding wave highlights a major capital shift toward cost-effective, developer-friendly API and hosting solutions.
Reachy Mini goes fully local
Hugging Face Blog19 days agoHardwareHugging Face published a tutorial for running Reachy Mini conversations without cloud audio processing or API keys. The setup uses its speech-to-speech library as a cascaded VAD, STT, LLM, and TTS pipeline exposed through a Realtime API-compatible WebSocket. Recommended defaults include llama.cpp with Gemma 4, Silero VAD, Parakeet-TDT, and Qwen3-TTS, while allowing swaps to vLLM, MLX, Transformers, or hosted Responses API providers.
Shipping a Trillion Parameters With a Hub Bucket: Delta Weight Sync in TRL
Hugging Face Blog19 days agoTutorialBased on the title, this Hugging Face Blog post focuses on Delta Weight Sync in TRL. It likely discusses moving or synchronizing weight differences at very large model scale using a Hub bucket-related workflow. Without the full article, implementation details, benchmarks, APIs, and stability claims cannot be confirmed.
Millions of AI agents imperiled by critical vulnerability in open source package★ 78
Ars Technica AI19 days agoIncidentArs Technica reports that Starlette, a Python package with about 325 million weekly downloads, has a critical vulnerability called BadHost. The flaw can let crafted Host headers confuse request.url.path, potentially bypassing middleware-based path authorization. AI infrastructure using FastAPI or Starlette, including vLLM, LiteLLM, MCP servers, LLM proxies, and agent frameworks, should upgrade Starlette and audit custom middleware.
3D-printable humanoid legs let robotics experiments run wild
Ars Technica AI19 days agoHardwareArs Technica reports that Hugging Face has introduced a roughly $2,500 bipedal humanoid robot project built around 3D-printable legs. The effort targets builders and researchers rather than mainstream consumers, lowering the hardware barrier for hands-on robotics experiments. Its broader significance is in open, reproducible embodied AI research, where models and control systems need physical platforms for testing.
Some ideas for what comes next, May 2026
Interconnects (Nathan L.)19 days agoCommentaryNathan Lambert argues that 2026 AI progress is becoming higher-stakes, with model capabilities, work patterns, economics, and real-world risks all escalating. He says open models still lack a true Claude Code and Opus 4.5-style agent moment, and Gemini has no clear competitor to Claude Code or Codex yet. The essay also tracks Mythos, American open-model momentum, frontier-lab competition, and mounting intervention from governments and other power structures.
專業化勝過規模：大多數 AI 採購決策忽略的關鍵戰略變數★ 75
Hugging Face Blog23 days agoOpinionIn the current wave of enterprise AI adoption, most decision-makers fall into the "scale myth" when making AI procurement decisions — the belief that the…
給 AI Agent 一台電腦：專訪 Daytona 執行長 Ivan Burazin，談 74% 月成長、裸機沙盒與全新 Agent Cloud★ 75
Latent Space24 days agoNew ToolIn this Latent Space interview, the hosts hold an in-depth conversation with Ivan Burazin, co-founder and CEO of Daytona. Daytona originally started as an…
Datasette Agent: An Extensible AI Assistant for Datasette★ 70
Simon Willison's Weblog24 days agoNew ToolSimon Willison announced the first release of Datasette Agent, merging his 'llm' Python library with Datasette. The tool provides a conversational interface to query SQLite databases, with plugin support for generating charts and running code in sandboxes. It runs efficiently on lightweight models like Gemini 3.1 Flash-Lite and supports local open-weight models via LM Studio.
datasette-agent 0.1a3 版本發布：優化 SQL 查詢檢視與截斷回應處理
Simon Willison's Weblog24 days agoReleaseSimon Willison's open-source AI assistant tool for Datasette, `datasette-agent`, has recently released version 0.1a3 in alpha. Datasette is an open-source…
Qwen 3.7 Max 現已支援 Vercel AI Gateway
Vercel Changelog24 days agoReleaseVercel officially released an update announcing that its AI infrastructure service, Vercel AI Gateway, now formally supports Alibaba Cloud's latest flagship…
datasette-agent-charts 0.1a1 發布：更豐富的色彩、互動式工具提示與權限檢查
Simon Willison's Weblog25 days agoReleaseSimon Willison has released the 0.1a1 early alpha version of datasette-agent-charts for his Datasette ecosystem. This plugin is designed to make it easier for…
Datasette 插件 datasette-llm-accountant 發布 0.1a4 版本，修正回應鏈追蹤錯誤
Simon Willison's Weblog26 days agoReleaseSimon Willison, the creator of the well-known open-source data analysis tool Datasette, today released version 0.1a4 of the ecosystem plugin…
AI2 發表 OlmoEarth v1.1：更高效的開源地球觀測模型家族★ 70
Hugging Face Blog26 days agoReleaseThe Allen Institute for AI (AI2) has officially released OlmoEarth v1.1 on Hugging Face. This is a brand-new family of open-source foundation models designed…
Hugging Face 推出 Ettin Reranker 重排模型家族：大幅提升 RAG 檢索精度與效率★ 80
Hugging Face Blog27 days agoReleaseIn building Retrieval-Augmented Generation (RAG) systems, accurately locating the most relevant information from a vast document collection has always been the…
Hugging Face 與 IBM 聯合推出 Open Agent Leaderboard：開源 AI 智能體效能評測全新基準★ 80
Hugging Face Blog27 days agoReleaseHugging Face and IBM Research have jointly announced the launch of the "Open Agent Leaderboard," aimed at establishing an objective, standardized, and fully…
Import AI 457：AI 版 Stuxnet 震網病毒、神祕的 Muon 優化器，以及積極對齊（Positive Alignment）★ 78
Import AI (Jack Clark)27 days agoCommentaryThis issue of Import AI 457, written by Jack Clark, delves into three forward-looking and stylistically distinct topics in the field of artificial…
英國政府數位服務局（GDS）介入 NHS 退出開源之爭，呼籲公共部門應「預設保持開源」
Simon Willison's Weblog28 days agoCommentaryThis report stems from Simon Willison's compilation of Terence Eden's follow-up coverage. The incident began when the UK's National Health Service (NHS), upon…
最新開放模型動態 (#21)：開放模型大爆發！Gemma 4、DeepSeek V4、Kimi K2.6、MiMo 2.5、GLM-5.1 等，以及 CAISI V4 評估分析★ 85
Interconnects (Nathan L.)29 days agoCommentaryThis is Issue #21 of the "Open Artifacts" column by well-known AI commentator Nathan Lambert, exploring the explosive growth in the open-weights and…
datasette-agent 0.1a2 發布：引入工具權限控制提升 AI 代理安全性
Simon Willison's Weblog31 days agoReleaseSimon Willison, the founder of the open-source data analysis tool Datasette, recently released the latest alpha version of the AI agent plugin datasette-agent…
解鎖連續批次處理（Continuous Batching）中的非同步機制★ 75
Hugging Face Blog32 days agoReleaseAs the demand for deploying large language models (LLMs) in production environments surges, how to improve inference efficiency and reduce costs has become a…
[AINews] 微調的終結？探討 Fine-tuning 在大模型時代的未來與轉變★ 75
Latent Space32 days agoOpinionAs AI technology continues to iterate at a rapid pace, the developer community is confronting a profound rethinking of the question: "Is fine-tuning heading…
探討開源模型生態系的複利效應：中國「開源優先」與高參與度 AI 生態的啟示★ 75
Interconnects (Nathan L.)33 days agoOpinionThis article delves into how the open-source AI model ecosystem achieves exponential growth through "compounding effects," using China's highly engaged…
在 AWS 上進行基礎模型訓練與推理的建構基石 (Building Blocks)★ 75
Hugging Face Blog34 days agoTutorialIn the era of generative AI, training and deploying foundation models with billions of parameters faces enormous computational and architectural challenges…
Superset 如何在 Vercel 上構建 AI Agent 專屬的 IDE
Vercel Changelog35 days agoCommentaryAs AI agents rise to prominence, traditional code editors can no longer meet developers' needs for debugging, observing, and orchestrating agents. Superset is…
來自中國頂尖 AI 實驗室的內部觀察筆記★ 75
Interconnects (Nathan L.)38 days agoCommentaryThis in-depth piece from Interconnects founder Nathan Lambert documents his key observations after personally visiting several of China's top AI laboratories —…