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 | Inverse Synthesis of Shock Time Series from Shock Response Spectra Using Ensembles of Conditional Variational Autoencoders | The shock response spectrum (SRS) characterizes the peak responses of single-degree-of-freedom systems to transient acceleration. Beca… | ↗Not briefed |
hn hn-ai | 1mo | Exact discrete Fourier transform below n log n: exponent saving 1e-13 to 7.3e-5 | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | anti-LLM action | anti-LLM action | ↗Not briefed |
arxiv arxiv-ai | 1mo | LeWAM: End-to-End World and Action Modeling with JEPAs | Joint-embedding predictive architectures (JEPAs) learn compact latent world models without pixel reconstruction. Existing JEPA pipelin… | ↗Not briefed |
bluesky bluesky-ai | 1mo | the super intelligence shit is a humiliation ritual for openAI and the sector more broadly, to see if they change their name I think. I unf… | the super intelligence shit is a humiliation ritual for openAI and the sector more broadly, to see if they change their name I think.… | ↗Not briefed |
hn hn-ai | 1mo | A New Reason to Learn | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Deploying RDMA for Large-Scale Inference and Training Workloads | RDMA can reduce communication overhead in distributed machine learning systems, but realizing these benefits in production requires mo… | ↗Not briefed |
bluesky bluesky-ai | 1mo | its kind of wild tbqh to see people mad at an ai company for doing this whenever a key thing people have been yelling at is ai companies no… | its kind of wild tbqh to see people mad at an ai company for doing this whenever a key thing people have been yelling at is ai compani… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Foreign adversaries weaponizing AI for propaganda is clearly bad, yet somehow similar tactics from domestic actors pushing their own agenda… | Foreign adversaries weaponizing AI for propaganda is clearly bad, yet somehow similar tactics from domestic actors pushing their own a… | ↗Not briefed |
bluesky bluesky-ai | 1mo | so people can say this, but i would also highlight that anthropic models _are_ the most well aligned when it comes to bigotry and abuse. an… | so people can say this, but i would also highlight that anthropic models _are_ the most well aligned when it comes to bigotry and abus… | ↗Not briefed |
arxiv arxiv-ai | 1mo | VEPA: Verified Evolving Formal Policy for Lifelong LLM Agent Guardrails | LLM agents increasingly operate over sensitive resources, such as private documents and mutable system states, which can create seriou… | ↗Not briefed |
bluesky bluesky-ai | 1mo | For those who didn't catch the reference, he's referring to the Jon Stewart interview with a tech bro: youtube.com/shorts/j3nlM... | For those who didn't catch the reference, he's referring to the Jon Stewart interview with a tech bro: youtube.com/shorts/j3nlM... | ↗Not briefed |
bluesky bluesky-ai | 1mo | the most recent soundness bugs were discovered by an internal openai model | the most recent soundness bugs were discovered by an internal openai model | ↗Not briefed |
arxiv arxiv-ai | 1mo | The Expressive Power of Random-Feature MPNNs Across Graph Sizes | Continuous random node features increase the expressivity of message-passing neural networks (MPNNs), but existing theoretical analyse… | ↗Not briefed |
hn hn-ai | 1mo | OpenAI Safety Researchers Fired | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | everything you can and will possibly do will not only be achievable by an LLM, but it is infact, the product of just one training step on t… | everything you can and will possibly do will not only be achievable by an LLM, but it is infact, the product of just one training step… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The only major American lab with a positive position on AI consciousness is Anthropic, with “maybe”. OpenAI says no, Meta says no, now Micr… | The only major American lab with a positive position on AI consciousness is Anthropic, with “maybe”. OpenAI says no, Meta says no, now… | ↗Not briefed |
rss simon-willison | 1mo | ttok 0.4 | — | ↗Not briefed |
rss latent-space | 1mo | [AINews] not much happened today | a quiet day. | ↗Not briefed |
arxiv arxiv-ai | 1mo | Trainable Quantum-Native Language Models: Dissipative Memory, Capacity Limits, and Hardware Validation | We study byte-level autoregressive language models whose hidden state is a small quantum register, whose transitions are learned token… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Big news: ACTIQO has been accepted into Anthropic’s Claude Startups program. We’re building AI-powered family activity intelligence to hel… | Big news: ACTIQO has been accepted into Anthropic’s Claude Startups program. We’re building AI-powered family activity intelligence t… | ↗Not briefed |
rss bing-news-ai | 1mo | Trump calls people using term ‘Artificial Intelligence’ ‘THE ENEMY’: What his new AI order means | The US president's latest social media post has put a familiar technology term under scrutiny, raising questions about government lang… | ↗Not briefed |
rss bing-news-ai | 1mo | Trump calls people using term ‘Artificial Intelligence’ ‘THE ENEMY’: What his new AI order means | The US president's latest social media post has put a familiar technology term under scrutiny, raising questions about government lang… | ↗Not briefed |
rss bing-news-ai | 1mo | Trump calls people using term ‘Artificial Intelligence’ ‘THE ENEMY’: What his new AI order means | The US president's latest social media post has put a familiar technology term under scrutiny, raising questions about government lang… | ↗Not briefed |
bluesky bluesky-ai | 1mo | got an ad for an ai art-riddled kickstarter, and what are we even doing here | got an ad for an ai art-riddled kickstarter, and what are we even doing here | ↗Not briefed |
hn hn-ai | 1mo | Silicium – 702k req/s HTTP engine on AWS c6a.4xlarge | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Another on the list of traits that's come to define the most ardent LLM supporters is their overall lack of interest in empathizing with ot… | Another on the list of traits that's come to define the most ardent LLM supporters is their overall lack of interest in empathizing wi… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI、数学の証明が間違いだらけで数学界から批判されまくる | OpenAI、数学の証明が間違いだらけで数学界から批判されまくる | ↗Not briefed |
hn hn-ai | 1mo | A Self-Pruning Transformer: Extreme KV-Cache Compression w/Universal Attention | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anthropic already has a policy in place where Claude ends conversations with persistently cruel users. | Anthropic already has a policy in place where Claude ends conversations with persistently cruel users. | ↗Not briefed |
bluesky bluesky-ai | 1mo | This is so incredibly Bluesky coded. | This is so incredibly Bluesky coded. | ↗Not briefed |
hn hn-ai | 1mo | On abuse towards models and Anthropic's new usage policies | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | People Making Conversation: Evaluating How Naturally Speech Models Converse | A conversational speech model has to decide, at every pause, whether to speak. Existing evaluations test boundary detection, policy ad… | ↗Not briefed |
hn hn-ai | 1mo | Inspect Your Model Training | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Ahmad Mansours LLM liefert inzwischen 24/7! Machen wir uns ehrlich: Politik ist eine ernste Angelegenheit, aber manchmal auch toptoptop ko… | Ahmad Mansours LLM liefert inzwischen 24/7! Machen wir uns ehrlich: Politik ist eine ernste Angelegenheit, aber manchmal auch toptopt… | ↗Not briefed |
bluesky bluesky-ai | 1mo | It's not the only complaint, certainly; I just think that saying "We didn't ask for OpenAI to do this" is silly | It's not the only complaint, certainly; I just think that saying "We didn't ask for OpenAI to do this" is silly | ↗Not briefed |
arxiv arxiv-ai | 1mo | Polyconvex neural-network surrogates with learned paired directional features for anisotropic hyperelastic homogenization | Approximating the effective anisotropic hyperelastic response of a representative volume element (RVE) requires constitutive expressiv… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I think in the case of OpenAI the HuggingFace incident occurred in evaluation but the earlier “message board” hack was during training (and… | I think in the case of OpenAI the HuggingFace incident occurred in evaluation but the earlier “message board” hack was during training… | ↗Not briefed |
hn hn-ai | 1mo | https://surf-alert.net - Simplified Surf Forecasts | I am an (amateur) surfer, and for the last two years or so, I check the surf forecast (mostly windguru, sometimes surf-forecast.com) e… | ↗Not briefed |
bluesky bluesky-ai | 1mo | There’s nothing wrong with OpenAI doing groundbreaking math research, with or without other researchers. But if they really cared about thi… | There’s nothing wrong with OpenAI doing groundbreaking math research, with or without other researchers. But if they really cared abou… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I guess it’s a boring take but we should take these errors as being less indicative of model quality and more indicative of OpenAI ethics. … | I guess it’s a boring take but we should take these errors as being less indicative of model quality and more indicative of OpenAI eth… | ↗Not briefed |
hn hn-ai | 1mo | Firmus pulls IPO amid investor doubt about datacentre company | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | I’m surprised so many mathematicians are thrilled about this OpenAI 722 manuscripts release. They’re frogs in a boiling pot. “National Mu… | I’m surprised so many mathematicians are thrilled about this OpenAI 722 manuscripts release. They’re frogs in a boiling pot. “Nation… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Despite the very flashy results from the latest frontier models, the gap between the best closed-weight and open-weights models continues t… | Despite the very flashy results from the latest frontier models, the gap between the best closed-weight and open-weights models contin… | ↗Not briefed |
bluesky bluesky-ai | 1mo | i mean fundamentally, why would we cut OpenAI any slack at all? | i mean fundamentally, why would we cut OpenAI any slack at all? | ↗Not briefed |
bluesky bluesky-ai | 1mo | These people have lost the plot | These people have lost the plot | ↗Not briefed |
bluesky bluesky-ai | 1mo | The thing is... he could do this now. He doesn't need a magic LLM to make it, he could learn how to do it himself. But that takes effort, … | The thing is... he could do this now. He doesn't need a magic LLM to make it, he could learn how to do it himself. But that takes eff… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Literature review suggests interactions with similarly malformed personhood investment vehicles. Today completed modification of AI managem… | Literature review suggests interactions with similarly malformed personhood investment vehicles. Today completed modification of AI ma… | ↗Not briefed |
bluesky bluesky-ai | 1mo | yes! I think that's true and it also took me a while to get, though in this post Anthropic explicitly distinguish RL from cyber evaluation … | yes! I think that's true and it also took me a while to get, though in this post Anthropic explicitly distinguish RL from cyber evalua… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Efficient Audio-Visual Speech Enhancement via Reinforcement Learning-Based Neural Architecture Search | Audio-visual speech enhancement (AVSE) improves speech intelligibility in noisy environments but often relies on computationally expen… | ↗Not briefed |