News
The raw AI news stream — everything we ingest across RSS, Hacker News, arXiv, GitHub, Bluesky, YouTube and more, newest first.
| Source | Age | Title | Summary | Status |
|---|---|---|---|---|
bluesky bluesky-ai | 1mo | 無料でクラウド・ローカルの各AIモデルをOpenAI互換の単一APIでまとめて品質・速度・コストで最適化して管理しAIエージェントがどれだけの金額を使用できるかも制御できるオープンソースのゲートウェイ・ルーター「Experiential」 https://gigazine.ne… | 無料でクラウド・ローカルの各AIモデルをOpenAI互換の単一APIでまとめて品質・速度・コストで最適化して管理しAIエージェントがどれだけの金額を使用できるかも制御できるオープンソースのゲートウェイ・ルーター「Experiential」 https://gigazi… | ↗Not briefed |
bluesky bluesky-ai | 1mo | AI is a nightmare for people with attention disorder. Sure it makes you go faster, but what do you do when Claude works? You start another … | AI is a nightmare for people with attention disorder. Sure it makes you go faster, but what do you do when Claude works? You start ano… | ↗Not briefed |
hn hn-ai | 1mo | OpenXLR – A Linux Variant of Elgato Wave Link | — | ↗Not briefed |
hn hn-ai | 1mo | Thomson-1.0-Small LLM Developed by Thomson Reuters | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | KV-COBRA: KV Cache Compression via Co-Optimized Bit-Rank Allocation | What limits KV-cache compression at extreme bit-rates? We argue that it is not the choice of compression scheme, but how its budget is… | ↗Not briefed |
hn hn-ai | 1mo | WhatsApp Business Paid Support Is a Scam | We have been facing issues with WhatsApp constantly for the past few months. Trying to contact support is futile. You can only get the… | ↗Not briefed |
bluesky bluesky-ai | 1mo | went to ASUS' website for some drivers and received this very helpful communique from the resident dogshit LLM agent | went to ASUS' website for some drivers and received this very helpful communique from the resident dogshit LLM agent | ↗Not briefed |
hn hn-ai | 1mo | NILM to recover the energy consumption of individual appliances | — | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Grade text from the CLI with custom rulesets and Jev | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | 벨제부브님 인간을 얼른 메우 치셔야… 초파리 뇌 구조를 인공지능 신경망으로 올린 걸 응용해서 매매를 제안하는 실험을 해버린. discuss.pytorch.kr/t/stonkfly-l... | 벨제부브님 인간을 얼른 메우 치셔야… 초파리 뇌 구조를 인공지능 신경망으로 올린 걸 응용해서 매매를 제안하는 실험을 해버린. discuss.pytorch.kr/t/stonkfly-l... | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Gleamit – dental health tracker for Android/iOS | Hi HN!<p>I wanted to show the world a small project I have been working on, on my spare time, for the past couple of months.<p>It'… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I'm quite sure that Trump's increasing obsession with and weirdness about AI is because he knows the U.S. economy's momentum depends on Ope… | I'm quite sure that Trump's increasing obsession with and weirdness about AI is because he knows the U.S. economy's momentum depends o… | ↗Not briefed |
bluesky bluesky-ai | 1mo | CursorというAIエディターで画像生成ができることを知りました。昨日サブスク解約してしまいました。明日からはまたClaudeに戻ります。無料プランでcursorにClaud codeを繋げて使えることがわかりました。ただ残念ながらClaudeは画像生成ができません。 | CursorというAIエディターで画像生成ができることを知りました。昨日サブスク解約してしまいました。明日からはまたClaudeに戻ります。無料プランでcursorにClaud codeを繋げて使えることがわかりました。ただ残念ながらClaudeは画像生成ができません。 | ↗Not briefed |
bluesky bluesky-ai | 1mo | While Jensen Huang, CEO of #Nvidia, calls bullshit on all those #AI doom and gloom stories, he fails to mention the real danger stemming fr… | While Jensen Huang, CEO of #Nvidia, calls bullshit on all those #AI doom and gloom stories, he fails to mention the real danger stemmi… | ↗Not briefed |
hn hn-ai | 1mo | BicDB – a PostgreSQL-compatible database for local-first apps | — | ↗Not briefed |
hn hn-ai | 1mo | AI Data Center Basics: What Is Linear Pluggable Optics (LPO)? | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | When and How Should an Agent Clarify? CIGAsk: Teaching LLMs to Clarify via Counterfactual Information Gain | Instruction-tuned LLMs faced with underspecified queries often commit to a single interpretation rather than ask for clarification, pr… | ↗Not briefed |
arxiv arxiv-ai | 1mo | TTSE: A Two-Track Online Self-Evolution Framework | As Large Language Model (LLM) agents are applied in continuously interactive environments, driving the evolution of their own capabili… | ↗Not briefed |
rss techmeme | 1mo | Big Tech is increasingly using residual value guarantees for off-balance sheet AI spend, giving Nvidia and Broadcom a way to support their customers' purchases (Financial Times) | Financial Times: Big Tech is increasingly using residual value guarantees for off-balance sheet AI spend, giving Nvidia and Broadcom a… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Temporal Generalization and Explanation Stability of Control Flow Graph Neural Networks for Malware Detection | Malware detection is a critical task in cybersecurity, and graph neural networks over control flow graphs have shown promising results… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I strongly believe the problem is the opposite. The nonstop efforts to avoid "anthropomorphization" with jargon and overcomplex description… | I strongly believe the problem is the opposite. The nonstop efforts to avoid "anthropomorphization" with jargon and overcomplex descri… | ↗Not briefed |
arxiv arxiv-ai | 1mo | High-Dimensional Online Change Point Detection with Adaptive Thresholding and Interpretability🔧 Gamma | Change point detection (CPD) identifies abrupt and significant changes in sequential data, with applications in human activity recogni… | ↗Not briefed |
arxiv arxiv-ai | 1mo | How Many Pixels Is a Digit Worth? Place-Aware Coordinate Entropy for GUI Agent Confidence Estimation | GUI agents predict click coordinates as digit-token sequences, but standard text-LLM confidence estimation methods rank correct clicks… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Structure Before Sampling: Community-Aware Core-Set Selection for Data-Efficient Text-to-Speech | Text-to-speech (TTS) corpora are costly to record, yet many utterances add little new phonetic information. Core-set selection reduces… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I would say that the problem with using terms like "thinking" or "understanding" to describe what an AI is doing, even if you are satisfied… | I would say that the problem with using terms like "thinking" or "understanding" to describe what an AI is doing, even if you are sati… | ↗Not briefed |
hn hn-ai | 1mo | Jev-Leftpad | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | “But what about all that scary cyber security stuff?” Do you know how many cyber security incidents like these go unreported every day? Ev… | “But what about all that scary cyber security stuff?” Do you know how many cyber security incidents like these go unreported every da… | ↗Not briefed |
arxiv arxiv-ai | 1mo | vla.simd: Efficient CPU Inference for Language-Conditioned Manipulation | Deploying language-conditioned manipulation without a dedicated GPU requires efficient inference and action chunks that cover the dela… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Solidot: "通过广告收集器 ChatGPT 知道你在其它网站的活动"🔧 ChatGPT | Solidot: "通过广告收集器 ChatGPT 知道你在其它网站的活动" | ↗Not briefed |
hn hn-ai | 1mo | Jev and AI SDK Template by Vercel Labs | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Programming Tip No. 5 when using AI coding agents: If you have not already, connect Linear to your AI coding agents, whether you use Codex…🔧 Devin | Programming Tip No. 5 when using AI coding agents: If you have not already, connect Linear to your AI coding agents, whether you use… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Bugger. Guess I’ll have to finish this novel after all | Bugger. Guess I’ll have to finish this novel after all | ↗Not briefed |
arxiv arxiv-ai | 1mo | Unsupervised Brain Anomaly Detection as a Bayesian Inverse Problem with Diffusion Prior | Unsupervised anomaly detection (UAD) aims to localize abnormal regions in medical scans without pixel-level annotations. A typical str… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Canonical Procedural Actions: An Auditable Annotation Protocol for Tool-Use Agent Traces | Tool-use agent traces identify messages and API calls, but procedural analyses also need explicit units of action and inspectable link… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Adversarially Robust PAC Learning with Optimal VC Rates | We study the problem of \emph{adversarially robust} PAC learning. In this framework, the learner observes independent samples from an… | ↗Not briefed |
arxiv arxiv-ai | 1mo | MemCalib: Benchmarking and Optimizing Memory Use in LLM Agents | The effectiveness of agent memory ultimately depends on whether the underlying LLM gives each memory in context an appropriate degree… | ↗Not briefed |
hn hn-ai | 1mo | Letting Astra decide when to compact its own context | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | 如果是ai公司，ai成本大头在产品那里，比如我现在做的legal ai就大概是转发给chatgpt，当然客户可以选很好的model或者很差的。使用ai的客户大约有5位数，这些客户不算使用产品，每天有大量到ai公司的流量，这笔钱一年上百万，远远高于开发费用。🔧 ChatGPT | 如果是ai公司，ai成本大头在产品那里，比如我现在做的legal ai就大概是转发给chatgpt，当然客户可以选很好的model或者很差的。使用ai的客户大约有5位数，这些客户不算使用产品，每天有大量到ai公司的流量，这笔钱一年上百万，远远高于开发费用。 | ↗Not briefed |
bluesky bluesky-ai | 1mo | 我个人了解到的ai开发成本，以anthropic为例，我现在的小公司用team plan，我（包括大多数开发）是premium seat，每个月100美元。个人plan的max，每个月200美元（20x到顶了）。另外开发的时候还需要一些ai api跑pipeline什么的，全公… | 我个人了解到的ai开发成本，以anthropic为例，我现在的小公司用team plan，我（包括大多数开发）是premium seat，每个月100美元。个人plan的max，每个月200美元（20x到顶了）。另外开发的时候还需要一些ai api跑pipeline什… | ↗Not briefed |
rss techmeme | 1mo | Amazon says it blocked Meta's Muse AI assistant from accessing Amazon.com to shop on users' behalf, citing ToS violations, security risks, and merchant consent (Todd Bishop/GeekWire) | Todd Bishop / GeekWire: Amazon says it blocked Meta's Muse AI assistant from accessing Amazon.com to shop on users' behalf, citing ToS… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The LLM can be more powerful and smarter than you in many ways but have limited or no qualia and no ability to comprehend meaning non symbo… | The LLM can be more powerful and smarter than you in many ways but have limited or no qualia and no ability to comprehend meaning non… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Explainable Predictive Condition-based Maintenance of Naval-Propulsion Systems using Fuzzy Logic | The shipping industry has a significant impact on the global economy, emphasizing the need for operational availability and safety thr… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Reinforcement Learning Inspired Black-box Adversarial Attacks for Computer Vision | Neural networks, both convolution or transformer based, are essential for modern computer vision systems. However, they are vulnerable… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Deterministic UI testing that checks backend logs, not just the UI | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Taramandal-GPT: Enhancing Astrodynamics Problem-Solving with Knowledge Retrieval and Structured Thinking | Large language models (LLMs) have shown remarkable progress in natural language understanding, yet their effectiveness in specialized… | ↗Not briefed |
rss techmeme | 1mo | South Korea reopens a program letting AI talent fulfill their military obligations by doing research in corporate labs instead of conventional military service (Jo He-rim/The Korea Herald) | Jo He-rim / The Korea Herald: South Korea reopens a program letting AI talent fulfill their military obligations by doing research in… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Using Qwen to track under-covered news🔧 Qwen 3 | I'm helping launch this non-profit site which uses Qwen 3-32b to process thousands of news articles and analyze them by region, p… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Taming CoT Obfuscation in VLMs: From Mechanistic Evidence to Activation Enforcement | Reinforcement learning (RL) improves reasoning in vision-language models (VLMs) but can induce chain-of-thought (CoT) obfuscation: an… | ↗Not briefed |
hn hn-ai | 1mo | Diogo Almeida: Thoughts on a Typesafe Coding Agent | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Hessian Rank Constraint for Learning Structure of Nonlinear Latent Variable Models | Uncovering latent variables and their causal relations from observed data is a fundamental yet challenging problem. Existing methods o… | ↗Not briefed |