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 |
|---|---|---|---|---|
arxiv arxiv-ai | 1mo | Second-order optimization of variable projection SVM models and road abnormality detection | We introduce a novel second-order optimization framework for minimizing so-called variable projection functionals. We demonstrate that… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Residual Learning in Empirical Asset Pricing | Shallow models are special cases of deep models, and deep models theoretically have the potential to outperform the shallow ones. Howe… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Sequential Pretraining Favors Large Models | Large neural networks often acquire capabilities that small models fail to learn. Does this stem from large models learning more repre… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Decoupled Optimization for Teacher-Student Semi-Supervised Learning via a Pioneer Student | Semi-supervised learning (SSL) relies on two core mechanisms: self-training under the Teacher-Student (T-S) framework and joint optimi… | ↗Not briefed |
bluesky bluesky-ai | 1mo | "woah did u see all the new math openai dropped?" "no, did you?" "no" | "woah did u see all the new math openai dropped?" "no, did you?" "no" | ↗Not briefed |
arxiv arxiv-ai | 1mo | Which Language Should a Skeleton Speak? Language Choices in Multilingual Reasoning | Skeleton-based reasoning prompting is a promising training-free approach for structuring LLM reasoning, but prior work largely assumes… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 人間なので絶対否定しない誰かによしよし♡あなた頑張ってます♡全部大丈夫ですよ♡して欲しくなる時もあるが、24時間365日のそいつを求めてLLMに行くと「終わる」だろうなというのが自分で見えているのでやらない(:˒[▓▓] | 人間なので絶対否定しない誰かによしよし♡あなた頑張ってます♡全部大丈夫ですよ♡して欲しくなる時もあるが、24時間365日のそいつを求めてLLMに行くと「終わる」だろうなというのが自分で見えているのでやらない(:˒[▓▓] | ↗Not briefed |
arxiv arxiv-ai | 1mo | Lightweight and Versatile Learned Optimization by Recombination of Gradient History | This paper presents a lightweight and versatile learned optimizer that dynamically recombines gradient history, represented as average… | ↗Not briefed |
bluesky bluesky-ai | 1mo | just to remind you what this is: this single box with two SSDs. is running deepseek v4.1 flash (a 552B model) in its native MXFP4, ~550GB… | just to remind you what this is: this single box with two SSDs. is running deepseek v4.1 flash (a 552B model) in its native MXFP4, ~… | ↗Not briefed |
arxiv arxiv-ai | 1mo | SafeEvo: Deciphering the Safety Alignment Mechanism and Evolution in Language Models | Safety interpretability advances the study of Large Language Model (LLM) alignment from behavioral constraints driven by data or algor… | ↗Not briefed |
youtube youtube-twominutepapers | 1mo | Claude Opus 5.5 AI: An Incredible Leap Forward | Okay, this is stunning. Finally, I tried Cloud Opus 5.5 and it did something that no other AI system ever could. First, I asked it to… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 🚀 BlueSky trending hashtags (15m): #dlsite #art #nsfw #섹블 #humpday #gamedev #indiegame #r4today #indiedev #aiイラスト #prescaregame_muc #inkt… | 🚀 BlueSky trending hashtags (15m): #dlsite #art #nsfw #섹블 #humpday #gamedev #indiegame #r4today #indiedev #aiイラスト #prescaregame_muc… | ↗Not briefed |
bluesky bluesky-ai | 1mo | LLM禍の損害、スクラップデータと欺瞞情報の洪水になるのもあるが、自他共に"知識人"と認められるタイプは「自分と似たような考えをする者との会話」に飢えてるから、「剽窃された自分の情報と会話してる」という壁打ちというか自慰行為というかにのめり込んでどんどん変になっていく。 | LLM禍の損害、スクラップデータと欺瞞情報の洪水になるのもあるが、自他共に"知識人"と認められるタイプは「自分と似たような考えをする者との会話」に飢えてるから、「剽窃された自分の情報と会話してる」という壁打ちというか自慰行為というかにのめり込んでどんどん変になっていく。 | ↗Not briefed |
arxiv arxiv-ai | 1mo | COPC: Coupled Off-Policy Correction for Asynchronous LLM Reinforcement Learning | Asynchronous RL accelerates large language model post-training by decoupling rollout generation from optimization, but trains on stale… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: GitHub Action to check AI-written code changes no files, without review | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Structured pre-generation elicitation versus single-shot prompting in AI-assisted enterprise decision-making: a randomised online experiment | Generative AI speeds, and mostly improves, professional work, but there is concern that users who delegate both the production and the… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Scalable Logistic Gaussian Process Density Regression with Kinetic Langevin Sampling | Conditional density estimation targets the full distribution of a response given covariates, as required, for example, for per-galaxy… | ↗Not briefed |
bluesky bluesky-ai | 1mo | They can also approximate edit distance in near-linear time! github.com/openai/math/... @jnalanko.bsky.social | They can also approximate edit distance in near-linear time! github.com/openai/math/... @jnalanko.bsky.social | ↗Not briefed |
arxiv arxiv-ai | 1mo | Learning Situation-Conditioned Thinking Policies for Long-Term LLM Agents | Long-running autonomous agents must reuse accumulated reasoning experience without allowing explicit historical memory and LLM context… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Dual- versus Single-Suggestion AI Support for Radiographic Interpretation in Residents: Randomized Multireader Study🔧 GPT-5 | Purpose: To compare dual- and single-suggestion AI support for radiographic interpretation by residents, particularly when the shared… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Adaptive Code Generation for Controlling Robots | Deploying robots as Complex Adaptive Systems (CAS) in unknown and dynamic environments necessitates a transition from rigid command li… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 常に「I'm fine.Thank you.」と答えるタイプの人間にはLLM(生成AI)の挙動は自分と同レベルなので「賢い！」ってなるんだよ。「知性がある」「賢い」って「自分への近さ」の評価だから。 | 常に「I'm fine.Thank you.」と答えるタイプの人間にはLLM(生成AI)の挙動は自分と同レベルなので「賢い！」ってなるんだよ。「知性がある」「賢い」って「自分への近さ」の評価だから。 | ↗Not briefed |
bluesky bluesky-ai | 1mo | Indeed from my point of view the enemy is not OpenAI, it’s the colleagues and the bureaucrats who think that everyone should use AI (which … | Indeed from my point of view the enemy is not OpenAI, it’s the colleagues and the bureaucrats who think that everyone should use AI (w… | ↗Not briefed |
hn hn-ai | 1mo | HN has lost the joy for me | There was a time I would come to HN to get some respite from the depressing state of reality. Ukraine war or climate crisis would driv… | ↗Not briefed |
bluesky bluesky-ai | 1mo | "I code as good as the LLM" so bad? | "I code as good as the LLM" so bad? | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Neptune, a free and open-source terminal built in Rust | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Collaborative Reasoning Distillation via Cross-Feedback and Coherent Curation | Reasoning capabilities are critical for advancing Large Language Models, yet current approaches either require massive computational b… | ↗Not briefed |
bluesky bluesky-ai | 1mo | "I code as good as the LLM, but it types a lot faster than me" is... well, you already know how I feel about that. ah well. | "I code as good as the LLM, but it types a lot faster than me" is... well, you already know how I feel about that. ah well. | ↗Not briefed |
bluesky bluesky-ai | 1mo | 「How are you?」 と聞かれたとき、人間はこれが「自分の近況や体調を聞いてる」と理解し、「先日赤痢になって生還しました」と英語ですらない言葉で答えられるんだが、 LLM(生成AI)の情報処理は「返答として確率が最も高い答え」を返すので、 身内が死のうが破産してようが末… | 「How are you?」 と聞かれたとき、人間はこれが「自分の近況や体調を聞いてる」と理解し、「先日赤痢になって生還しました」と英語ですらない言葉で答えられるんだが、 LLM(生成AI)の情報処理は「返答として確率が最も高い答え」を返すので、 身内が死のうが破産し… | ↗Not briefed |
bluesky bluesky-ai | 1mo | *Enshittification intensifies* gizmodo.com/openai-decid... | *Enshittification intensifies* gizmodo.com/openai-decid... | ↗Not briefed |
bluesky bluesky-ai | 1mo | The best thing we could do would be to start promoting the continuing value of mathematics, targeting 14-16 year olds. We need an advertisi… | The best thing we could do would be to start promoting the continuing value of mathematics, targeting 14-16 year olds. We need an adve… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Correct Answers, Unsupported Findings: Evidence Binding in Forensic Reconstruction of LLM Agent Logs | Forensic reconstruction of LLM-agent actions requires not only recovering the correct value, but establishing which preserved record s… | ↗Not briefed |
bluesky bluesky-ai | 1mo | does anyone else feel like they're like a LLM where everything that comes out of their mouth is just a prediction based on the last word th… | does anyone else feel like they're like a LLM where everything that comes out of their mouth is just a prediction based on the last wo… | ↗Not briefed |
bluesky bluesky-ai | 1mo | French AI company Mistral has announced the public preview of its latest model, Mistral Large 4, called “Le Chonk”, with 1 trillion paramet…🔧 Mistral Large | French AI company Mistral has announced the public preview of its latest model, Mistral Large 4, called “Le Chonk”, with 1 trillion pa… | ↗Not briefed |
bluesky bluesky-ai | 1mo | - איפה אני נמצא - בכדור פורח - אתה מתמטיקאי, נכון? - נכון! איך ידעת? - רואים שבכית הבוקר בעקבות זה ש-OpenAI פרסמו על מאות בעיות שהמודל שלהם… | - איפה אני נמצא - בכדור פורח - אתה מתמטיקאי, נכון? - נכון! איך ידעת? - רואים שבכית הבוקר בעקבות זה ש-OpenAI פרסמו על מאות בעיות שהמודל… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Online Resource Allocation with an Endogenous Markov State: Fewer LP Solves Earn More | We study finite-horizon online resource allocation with i.i.d. requests and an endogenous Markov state on a finite state space: each a… | ↗Not briefed |
arxiv arxiv-ai | 1mo | PEACE: Covariant learning of nonadiabatic manifolds with parity-resolved Hamiltonians | Nonadiabatic molecular dynamics provides mechanistic insight into light-driven processes and informs the design of molecules and mater… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Oggi su Linkiesta esce un mio reportage dalla Svizzera: da Apertus, LLM sovrano, al super computer Alps, dono agli investimenti in robotica… | Oggi su Linkiesta esce un mio reportage dalla Svizzera: da Apertus, LLM sovrano, al super computer Alps, dono agli investimenti in rob… | ↗Not briefed |
bluesky bluesky-ai | 1mo | It’s not about us. You and I care about the influence our work has in our community. OpenAI also cares about influence, but it’s not the ma… | It’s not about us. You and I care about the influence our work has in our community. OpenAI also cares about influence, but it’s not t… | ↗Not briefed |
arxiv arxiv-ai | 1mo | How Do LLMs Change Predictions Under Negation? | Negation is an essential feature of human language, yet large language models (LLMs) remain unreliable in processing it. We evaluate r… | ↗Not briefed |
arxiv arxiv-ai | 1mo | RELATE: An Evaluation Framework for measuring Relational Orientation of Large Language Models | Large language models (LLMs) are increasingly used for emotional support, raising concern that sustained use may draw users away from… | ↗Not briefed |
bluesky bluesky-ai | 1mo | github.com/openai/math/... 2-approximation for the shortest common superstring! Better than greedy!! | github.com/openai/math/... 2-approximation for the shortest common superstring! Better than greedy!! | ↗Not briefed |
arxiv arxiv-ai | 1mo | EvoSignal: LLM-Guided Evolutionary Design of Modular Traffic Signal Control Programs | Effective traffic signal control (TSC) requires policies that respond to changing traffic demand and network conditions while meeting… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Good thread on today's dump by OpenAI of 722(!) preprints on unsolved math problems. Slightly alleviated my sense of existential despair ab… | Good thread on today's dump by OpenAI of 722(!) preprints on unsolved math problems. Slightly alleviated my sense of existential despa… | ↗Not briefed |
hn hn-ai | 1mo | CVE-2026-63277 – LibreOffice Remote Code Execution | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | World Potential Model: Pretrained World Knowledge as Progress Potentials | Long-horizon language agents often receive supervision only from terminal task outcomes, leaving little signal for distinguishing prod… | ↗Not briefed |
arxiv arxiv-ai | 1mo | UniCSI Towards a Universal Wi-Fi CSI Encoder for Ubiquitous Human Sensing | Wi-Fi sensing promises to turn the everyday wireless signals that already surround us into ubiquitous sensors for human sensing. Howev… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Japanese publishers reportedly outraged after domestic company admits to selling large volumes of books to AI developer Anthropic https://… | Japanese publishers reportedly outraged after domestic company admits to selling large volumes of books to AI developer Anthropic htt… | ↗Not briefed |
arxiv arxiv-ai | 1mo | DrugTargetWorld: A Synthetic Biobank for Training and Benchmarking AI Scientists🔧 GPT-5 | Drug target discovery requires distinguishing molecules that causally drive disease from those that are merely associated with it. Tra… | ↗Not briefed |
bluesky bluesky-ai | 1mo | New AI Decision Model Could Transform Content Moderation thailandai.news/new-ai-decis... #AINews #ThailandAINews #ThailandAI #News #TechN… | New AI Decision Model Could Transform Content Moderation thailandai.news/new-ai-decis... #AINews #ThailandAINews #ThailandAI #News #… | ↗Not briefed |