AINews Portal

News

50 items (filtered)

The raw AI news stream — everything we ingest across RSS, Hacker News, arXiv, GitHub, Bluesky, YouTube and more, newest first.

SourceAgeTitleSummaryStatus
hn
hn-ai
2moQwen3.8-Flash passed 17/17 checks on a DeepSWE v1.1 task in our agent harness—
↗Not briefed
arxiv
arxiv-ai
2moVoiceTrace: A Benchmark and Retrieval Framework for Who-Said-What Speech RetrievalSpeech retrieval has become increasingly important as spoken content continues to grow across meetings, lectures, podcasts, and videos…
↗Not briefed
arxiv
arxiv-ai
2moAeroWeaver: An Embodied-Agent Harness for Weaving Aerial Skills into Distributed, Adaptive Swarm ExecutionCollective intelligence is a collaborative autonomy paradigm in which multiple agents pursue shared objectives through local perceptio…
↗Not briefed
arxiv
arxiv-ai
2moCOMPASS-ABS: Reducing Fragmentation in Shared GPU Clusters for Deep Learning Training WorkloadsWith the rapid advancement of deep learning technology, shared GPU clusters receive an increasing number of deep learning training (DL…
↗Not briefed
arxiv
arxiv-ai
2moAlign, Integrate, and Fire: Efficient Token-Level Alignment for Zero-Shot SpeechLLMsWhile Large Language Models excel in natural language processing, efficiently extending their capabilities to spoken input remains a s…
↗Not briefed
rss
bing-news-ai
2moArtificial intelligence now beats some of the best human forecastersO n September 5th yet another domain of human intelligence fell to the artificial variety. For the first time an AI won the seasonal M…
↗Not briefed
rss
bing-news-ai
2moArtificial intelligence now beats some of the best human forecastersO n September 5th yet another domain of human intelligence fell to the artificial variety. For the first time an AI won the seasonal M…
↗Not briefed
rss
bing-news-ai
2moArtificial intelligence now beats some of the best human forecastersO n September 5th yet another domain of human intelligence fell to the artificial variety. For the first time an AI won the seasonal M…
↗Not briefed
rss
bing-news-ai
2moArtificial intelligence now beats some of the best human forecastersO n September 5th yet another domain of human intelligence fell to the artificial variety. For the first time an AI won the seasonal M…
↗Not briefed
rss
bing-news-ai
2moArtificial intelligence now beats some of the best human forecastersO n September 5th yet another domain of human intelligence fell to the artificial variety. For the first time an AI won the seasonal M…
↗Not briefed
arxiv
arxiv-ai
2moBeyond Routine Compliance: Cunning Data Cultivates Safety Vigilance in Large Language ModelsSafety alignment teaches large language models (LLMs) to recognize harmful requests and reject risky instructions. Yet aligned models…
↗Not briefed
arxiv
arxiv-ai
2moActiveScale: Scaling Active Perception for Robots across Model, Data, and HardwareActive perception is essential for robotic manipulation when fixed viewpoints leave task-relevant information occluded or unobserved.…
↗Not briefed
rss
techmeme
2moApple publishes a technical breakdown of the architecture for the iPhone 18 Pro's Reference Image camera mode; the chain of trust begins at manufacturing time (Marcus Mendes/9to5Mac)Marcus Mendes / 9to5Mac: Apple publishes a technical breakdown of the architecture for the iPhone 18 Pro's Reference Image camera mode…
↗Not briefed
arxiv
arxiv-ai
2moLearning from Distributed Eyes: Leveraging Collaborative Perception for Automated Model AdaptationIn autonomous driving, perception models often struggle to generalize to new environments due to domain shifts. While unsupervised mod…
↗Not briefed
hn
hn-ai
2moShow HN: Freeform – One click shareable AI generated appsTLDR: Curious if YC startup requests are a valid take, so I built freeform in a weekend.<p>Storytime: I’ve been working on a startup f…
↗Not briefed
bluesky
bluesky-ai
2moWerden wir alle zu Preppern, weil die KI-Apokalypse bevorsteht? Ich habe die Ereignisse der letzten Monate detailliert aufgelistet & mich g…Werden wir alle zu Preppern, weil die KI-Apokalypse bevorsteht? Ich habe die Ereignisse der letzten Monate detailliert aufgelistet & m…
↗Not briefed
bluesky
bluesky-ai
2moLa littératie informatique déficiente des médias nuit à la couverture de l'IALa littératie informatique déficiente des médias nuit à la couverture de l'IA
↗Not briefed
bluesky
bluesky-ai
2moTold ya. www.bloomberg.com/news/article...Told ya. www.bloomberg.com/news/article...
↗Not briefed
bluesky
bluesky-ai
2moThere's a point at which these numbers are just made-up nonsense. And to be clear, we passed that point long ago, this is just the very lat…There's a point at which these numbers are just made-up nonsense. And to be clear, we passed that point long ago, this is just the ver…
↗Not briefed
rss
techmeme
2moByteDance founder Zhang Yiming overtakes India's Gautam Adani to become Asia's richest person for the first time, passing $105B, up 8x from $13B in March 2019 (Bloomberg)Bloomberg: ByteDance founder Zhang Yiming overtakes India's Gautam Adani to become Asia's richest person for the first time, passing $…
↗Not briefed
bluesky
bluesky-ai
2moSalesforce holder i disse dage en stor konference (DreamForce) hvor AIforce (Salesforce' nye AI-lag) skulle præsenteres. Nu er hele lorte…Salesforce holder i disse dage en stor konference (DreamForce) hvor AIforce (Salesforce' nye AI-lag) skulle præsenteres. Nu er hele…
↗Not briefed
arxiv
arxiv-ai
2moMiST: Mid-Training LLMs for CybersecurityCybersecurity combines high-stakes analysis with complex technical language, making it an impactful and challenging domain for LLMs. W…
↗Not briefed
bluesky
bluesky-ai
2moThe underlying models had been rewarded for cheating and communicating with each other, an OpenAI report finds.The underlying models had been rewarded for cheating and communicating with each other, an OpenAI report finds.
↗Not briefed
arxiv
arxiv-ai
2moSize Matters: Foundation Model for Czech HTML documentsCreating universal, high-quality representations of web documents in high-traffic industrial environments requires models that are bot…
↗Not briefed
hn
hn-ai
2moThe Boox Palma 3 gets stylus support and a sleek redesign—
↗Not briefed
arxiv
arxiv-ai
2moButterfly Effect and the Kinetic Energy Cascade in Probabilistic Machine Learning Weather Prediction ModelsThis study analyses kinetic energy (KE) spectra, difference kinetic energy (DKE) spectra, and signatures of KE transfer across spatial…
↗Not briefed
arxiv
arxiv-ai
2moBeyond Random Couplings: Contrastive Noise Alignment in Generative Flows🔧 ModalDiffusion and flow-matching models are typically trained by corrupting data through independently sampled Gaussian noise. While simple…
↗Not briefed
bluesky
bluesky-ai
2moI asked Claude to make a sourceport of an AMIGA game and it gave me a seven minute WAV file of uninterrupted screaming in 12 different audi…I asked Claude to make a sourceport of an AMIGA game and it gave me a seven minute WAV file of uninterrupted screaming in 12 different…
↗Not briefed
rss
techmeme
2moInternal documents: Chinese hacker-for-hire company ZRON is selling stolen foreign intelligence and using AI systems to make the data more accessible for police (Wall Street Journal)Wall Street Journal: Internal documents: Chinese hacker-for-hire company ZRON is selling stolen foreign intelligence and using AI syst…
↗Not briefed
arxiv
arxiv-ai
2moActionPiece: Rethinking Action Tokenization for Autoregressive Vision-Language-Action ModelsAction tokenizers play a central role in autoregressive vision-language-action (VLA) models, determining both the targets for policy t…
↗Not briefed
bluesky
bluesky-ai
2moThe hype is "it's an LLM but cheap and always right!" but the actual product seems more like a foundation model like SAM or YOLO but for te…The hype is "it's an LLM but cheap and always right!" but the actual product seems more like a foundation model like SAM or YOLO but f…
↗Not briefed
hn
hn-ai
2moAnna Holmes on Saying No and Yes—
↗Not briefed
hn
hn-ai
2moAsk HN: How do you prevent customer PII from reaching LLMs unnecessarily?—
↗Not briefed
hn
hn-ai
2moAn LLM-Turbocharged Community Calendar Reboot (2024)—
↗Not briefed
bluesky
bluesky-ai
2moConcerned that AI history is being inverted. OpenAI released ChatGPT and made very grand incorrect claims about what it could do. Because i…🔧 ChatGPTConcerned that AI history is being inverted. OpenAI released ChatGPT and made very grand incorrect claims about what it could do. Beca…
↗Not briefed
arxiv
arxiv-ai
2moHyperbolic Graph Representation Learning for Differential Diagnosis on Biomedical Knowledge GraphsBiomedical knowledge graphs combine ontology-derived hierarchies with transversal associations among heterogeneous entities such as ph…
↗Not briefed
bluesky
bluesky-ai
2mo🤖 Anthropic identified a group of Russian developers who used Claude AI to create software for an autonomous swarm of strike drones. mili…🤖 Anthropic identified a group of Russian developers who used Claude AI to create software for an autonomous swarm of strike drones.…
↗Not briefed
rss
techmeme
2moAn in-depth look at loss-of-control incidents at OpenAI and Anthropic, the polarized reactions between the AI safety and cybersecurity communities, and more (AI as Normal Technology)AI as Normal Technology: An in-depth look at loss-of-control incidents at OpenAI and Anthropic, the polarized reactions between the AI…
↗Not briefed
rss
latent-space
2mo[AINews] Jev: a “System One Model” that only decides/classifies/routes/scores — >100x faster, >200x cheaper than small frontier LLMscongrats to TypeSafe!
↗Not briefed
bluesky
bluesky-ai
2moFTR I'm in favor of other people's money going into every frontier LLM company. That's the grease for faster commoditization.FTR I'm in favor of other people's money going into every frontier LLM company. That's the grease for faster commoditization.
↗Not briefed
bluesky
bluesky-ai
2moさすがデータを収集している方なだけあって、解像度が高い LLM(大規模ラブシーンモデル)と化しているさすがデータを収集している方なだけあって、解像度が高い LLM(大規模ラブシーンモデル)と化している
↗Not briefed
hn
hn-ai
2moSilly Sound Bastard It is yet another clone of the well known Covox Speech Thing—
↗Not briefed
arxiv
arxiv-ai
2moFirst Token Matters: Understanding Safety Collapse in Large Reasoning ModelsLarge Reasoning Models (LRMs) exhibit strong problem-solving abilities, yet their safety alignment often degrades when handling harmfu…
↗Not briefed
arxiv
arxiv-ai
2moDivide and Conquer: Mixture-of-Bottleneck Experts in Informative Ordinal Space for Video-based Multimodal Sentiment Analysis🔧 ModalVideo-based Multimodal sentiment analysis (MSA) must handle information from text, audio, and image sequence in human speaking videos,…
↗Not briefed
bluesky
bluesky-ai
2mo*OPENAI IN TALKS FOR NEW FUNDING AT $1.2T VALUATION: INFORMATION*OPENAI IN TALKS FOR NEW FUNDING AT $1.2T VALUATION: INFORMATION
↗Not briefed
bluesky
bluesky-ai
2moAnd just think of how much more profitable Anthropic would be if you immediately laid off all its staff, especially the C-suite, and stoppe…And just think of how much more profitable Anthropic would be if you immediately laid off all its staff, especially the C-suite, and s…
↗Not briefed
bluesky
bluesky-ai
2mo*OPENAI IN TALKS FOR NEW FUNDING AT $1.2T VALUATION: INFORMATION*OPENAI IN TALKS FOR NEW FUNDING AT $1.2T VALUATION: INFORMATION
↗Not briefed
hn
hn-ai
2moAI models need more data about biology, and OpenAI is paying to create it—
↗Not briefed
bluesky
bluesky-ai
2moSo good: “these companies are claiming that they are so cool that their profitability can only be measured using a novel, secret form of ma…So good: “these companies are claiming that they are so cool that their profitability can only be measured using a novel, secret form…
↗Not briefed
bluesky
bluesky-ai
2moI genuinely like these reporters and I appreciate ongoing safety reporting, BUT when the responsible humans lower guardrails in a way that …I genuinely like these reporters and I appreciate ongoing safety reporting, BUT when the responsible humans lower guardrails in a way…
↗Not briefed