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 | HVEM velger ut hva du får vite av en LLM? Både: Ingen - utregningene til maskinen velger Og: Tusenvis av designere og programmerere vi ik… | HVEM velger ut hva du får vite av en LLM? Både: Ingen - utregningene til maskinen velger Og: Tusenvis av designere og programmerere… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The company said Sonnet 5.5 doesn't advance the frontier, but that it is better at coding and knowledge work tasks than its predecessor. | The company said Sonnet 5.5 doesn't advance the frontier, but that it is better at coding and knowledge work tasks than its predecesso… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Harnessing Large Language Models to Compile Task-Relevant Context into Bayesian Optimisation | Incorporating rich task-relevant context, such as domain knowledge and external observations, is a key capability yet remains challeng… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anthropic is making a bet that AI will transform the global economy more profoundly than industrialization, electricity and the internet, a… | Anthropic is making a bet that AI will transform the global economy more profoundly than industrialization, electricity and the intern… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Holy shit! "OpenAI agents hit UN website more than 16,000 times, used aggressive techniques, bypassed a filter." l.smartnews.com/p-8zYSam… | Holy shit! "OpenAI agents hit UN website more than 16,000 times, used aggressive techniques, bypassed a filter." l.smartnews.com/p-8… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Aperture: Merge-Consistent Rotary States for Compressed Tokens | Token compression combines content from several positions, yet rotary position embeddings usually assign the merged token one coordina… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Can’t make it up: Anthropic warns of ‘existential risks to humanity’ in IPO prospectus as.ft.com/r/21e300e9-e... | Can’t make it up: Anthropic warns of ‘existential risks to humanity’ in IPO prospectus as.ft.com/r/21e300e9-e... | ↗Not briefed |
arxiv arxiv-ai | 1mo | Code4Scene: Benchmarking Coding Agents for Constructing and Editing 3D Scenes | Frontier coding agents can now write and execute code that authors 3D environments, but whether they reliably understand 3D structure… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Beyond Conditional Independence: Root Cause Analysis with Deep Causal Models | Root cause analysis (RCA) is a critical problem in many real-world scenarios. RCA enables the identification of faulty or failing mech… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Emergent Specialization in Populations of Self-Supervised Collaborative Vision Experts Without a Shared Gate or Cross-Agent Gradients | Can a population of neural networks develop a useful division of labor without a shared gate or gradients between agents? We study a s… | ↗Not briefed |
hn hn-ai | 1mo | Fable Decides, Opus and Sonnet Do the Work: How I Route Claude Code Subagents | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | AI UX conventions rn, from the idea that regular ppl understand the implications of permissions to model selection are so absurd. | AI UX conventions rn, from the idea that regular ppl understand the implications of permissions to model selection are so absurd. | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI’s delay in their latest release of an AI modelling safety grounds is a good news story. Pharma companies pull promising new medicin… | OpenAI’s delay in their latest release of an AI modelling safety grounds is a good news story. Pharma companies pull promising new me… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Ich bin alt genug, die DOT COM Blase miterlebt zu haben. Was da an Geld verbrannt wurde, bevor die Blase platzte ist Kindergeburtstag auf d… | Ich bin alt genug, die DOT COM Blase miterlebt zu haben. Was da an Geld verbrannt wurde, bevor die Blase platzte ist Kindergeburtstag… | ↗Not briefed |
arxiv arxiv-ai | 1mo | When Can Prefixes Compile LoRA? Exact Resource-Capped Tests for Frozen Attention | Can a fixed continuous prefix replace a given low-rank adapter while the attention head stays frozen? In this research, we show that t… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Graph-Spectral Flow Matching for Multivariate Time Series Anomaly Detection | Multivariate time series anomaly detection typically relies on evaluating discrepancies between observations and outputs produced by m… | ↗Not briefed |
bluesky bluesky-ai | 1mo | When your product's still shit but you need to keep the investor dollars flowing and authorities from applying laws to you www.bbc.co.uk/ne… | When your product's still shit but you need to keep the investor dollars flowing and authorities from applying laws to you www.bbc.co.… | ↗Not briefed |
bluesky bluesky-ai | 1mo | ‼️What a way to kick off an IPO! 🥳💀 Anthropic "devoted roughly 80 pages of the 261-page main body of its prospectus to laying out risk f… | ‼️What a way to kick off an IPO! 🥳💀 Anthropic "devoted roughly 80 pages of the 261-page main body of its prospectus to laying out r… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Federated Clustering with Unknown Local and Global Cluster Cardinalities | Federated clustering methods that do not require the global number of clusters $K$ still assume that each client knows its local numbe… | ↗Not briefed |
bluesky bluesky-ai | 1mo | every time I see a screenshot of an LLM "apologizing" I can't help but laugh, the jokes about it are not exaggerated in the slightest | every time I see a screenshot of an LLM "apologizing" I can't help but laugh, the jokes about it are not exaggerated in the slightest | ↗Not briefed |
arxiv arxiv-ai | 1mo | QuantMLA: Function-Aligned Dual-Path Quantization for Low-Bit MLA KV Caching | Multi-Head Latent Attention (MLA) enables expressive multi-head attention with compact caches for its content and decoupled RoPE paths… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Dual-Mode Low-Rank Learner with Bridge-Prototype Ensemble for Vision-Language Class-Incremental Learning🔧 Modal | Benefiting from transferable visual-textual alignment, CLIP has been widely adopted for class-incremental learning (CIL). However, exi… | ↗Not briefed |
bluesky bluesky-ai | 1mo | - is token-based learning _really_ sufficient to shift the underlying alignment/trained behaviour? we know that they can change the role th… | - is token-based learning _really_ sufficient to shift the underlying alignment/trained behaviour? we know that they can change the ro… | ↗Not briefed |
bluesky bluesky-ai | 1mo | ref: www.cnbc.com/2026/08/17/a... | ref: www.cnbc.com/2026/08/17/a... | ↗Not briefed |
rss techmeme | 1mo | Coinbase gets CFTC approval for Coinbase Clearing, completing its derivatives stack, letting it list, broker, and clear fully collateralized products in-house (Danny Park/The Block) | Danny Park / The Block: Coinbase gets CFTC approval for Coinbase Clearing, completing its derivatives stack, letting it list, broker,… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Ed's schtick of "well, 2025 losses should be indicative", are easily rebutted. I'd believe that spending has increased modestly for 2026, b… | Ed's schtick of "well, 2025 losses should be indicative", are easily rebutted. I'd believe that spending has increased modestly for 20… | ↗Not briefed |
arxiv arxiv-ai | 1mo | NesTok: Nested Self-Aligned 1D Tokenizer for Autoregressive Image Generation | One-dimensional (1D) variable-length visual tokenizers enable adaptive compression by varying the number of tokens, allowing downstrea… | ↗Not briefed |
hn hn-ai | 1mo | OpenAI axes next model citing safety issues | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Does a prosody-trained representation help beyond trainable fusion? A parameter-matched study with frozen HuBERT | Explicit prosodic cues may help automatic speech recognition (ASR) of spontaneous speech, but auxiliary representations typically requ… | ↗Not briefed |
arxiv arxiv-ai | 1mo | cktFormer: Transformer-Based Approach for Automated Analog Circuit Design | Circuit design is a complex and iterative process that requires expertise in electronic engineering. It involves selecting components… | ↗Not briefed |
hn hn-ai | 1mo | Show HN: Jev Decision Layer: Save Frontier Tokens on Closed Decisions | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | If only more criminals apologised. They would get invites to parliament, explain what they did and then get carte blanche to keep breaking … | If only more criminals apologised. They would get invites to parliament, explain what they did and then get carte blanche to keep brea… | ↗Not briefed |
hn hn-ai | 1mo | OpenAI models may get more reasoning budget through the API than through Codex | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Group-Marginalized Self-Rewarding RL Drives Zero-Label Self-Evolving | Self-rewarding reinforcement learning (RL) enables large language models (LLMs) to self-evolve without human labels. Existing ensemble… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Generalizable Lifelong Model Editing via Preference Optimization | Knowledge editing enables rapid updates of specific factual knowledge in large language models (LLMs) without full retraining. However… | ↗Not briefed |
bluesky bluesky-ai | 1mo | If AI doesn't kill you, three coal fired power plants might. techrepublic.com/article/news-anthropic-apac-australia-queensland-data-center/ | If AI doesn't kill you, three coal fired power plants might. techrepublic.com/article/news-anthropic-apac-australia-queensland-data-c… | ↗Not briefed |
arxiv arxiv-ai | 1mo | EASE: Behavior-Adaptive Skill Curation for Self-Evolving Agents🔧 Kimi K2 | Agent skills provide a lightweight mechanism for self-evolving agents to accumulate reusable procedural knowledge without updating mod… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Have to agree with Misty. It has a very LLM structure but maybe they translated or used a LLM to restructure their post. | Have to agree with Misty. It has a very LLM structure but maybe they translated or used a LLM to restructure their post. | ↗Not briefed |
bluesky bluesky-ai | 1mo | I am not optimistic that Anthropic is in a position to become a profitable company in the long term | I am not optimistic that Anthropic is in a position to become a profitable company in the long term | ↗Not briefed |
arxiv arxiv-ai | 1mo | SIPO: Unifying Reinforcement Learning with On-Policy Self-Distillation | Reinforcement learning with verifiable rewards (RLVR) has become a standard paradigm for improving large language models (LLMs) on var… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anthropic’s prospectus details losses, growth, and, yes, a warning that its AI could end humanity | Anthropic’s prospectus details losses, growth, and, yes, a warning that its AI could end humanity | ↗Not briefed |
bluesky bluesky-ai | 1mo | Wer zahlt eigentlich den ganzen Mehraufwand, den ich habe, weil der kryptofaschistische Komplex überall seine LLM-Chatbots reinpusht? | Wer zahlt eigentlich den ganzen Mehraufwand, den ich habe, weil der kryptofaschistische Komplex überall seine LLM-Chatbots reinpusht? | ↗Not briefed |
rss techcrunch-ai | 1mo | Anthropic’s prospectus details losses, growth, and, yes, a warning that its AI could end humanity | In its prospectus, Anthropic just told investors it's losing tens of billions of dollars a year, but also growing like crazy, and — oh… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Distinguish or Homogenize: Last-Chance Policy Identification and Risk-Budgeted Recovery under Irreversible Resource Depletion | Under irreversible resource depletion, an agent can spend resources to distinguish among latent fault models, or to change the system… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Frontier Autolab: Organizational Memory, Adversarial Dissent and Temporal Leakage in Multi-Agent LLM Firms Across Fifty Years of Technological Change | Multi-agent LLM systems are increasingly structured like organizations, with roles, critics and shared memory, yet they are evaluated… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Efficient Offline Learning of Ranking Policies via Top-$k$ Policy Decomposition | Many recommender systems such as for e-commerce and news platforms aim to provide users with rankings they are likely to interact with… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Backpropagated Output Momentum: Relocating Optimizer History from Parameters to Task Space | Optimizer momentum is usually stored as a parameter-sized moving average of past gradients, which makes history costly and fixes each… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI do seem a whole lot more enthusiastic than Anthropic at shouting about spending a trillion dollars on data centers or whatever - I t… | OpenAI do seem a whole lot more enthusiastic than Anthropic at shouting about spending a trillion dollars on data centers or whatever… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Reconstructing the Vocal Tract with Differentiable Acoustic Simulation | The vocal tract is the region of the human body responsible for filtering one's voice to create speech. In this paper, we present a di… | ↗Not briefed |
arxiv arxiv-ai | 1mo | From Neurons to Conversation: Speech Brain-Computer Interfaces | Speech brain-computer interfaces (BCIs) aim to restore communication by transforming neural activity related to speech, language, or c… | ↗Not briefed |