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
13dVibe-Shield – Static security scanner for AI-generated Next.js/Supabase code—
↗Not briefed
arxiv
arxiv-ai
13dLAION-Audio-XL: A Web-Scale Corpus and Scaling Studies for Contrastive Language-Audio PretrainingRecent progress on contrastive audio-language models can be attributed mainly to the scale and quality of their training data. Larger…
↗Not briefed
bluesky
bluesky-ai
13dI’m looking forward to two years of bipartisan efforts to fund Israel and bail out OpenAI.I’m looking forward to two years of bipartisan efforts to fund Israel and bail out OpenAI.
↗Not briefed
hn
hn-ai
13dLLMs can understand and write neural network weights—
↗Not briefed
bluesky
bluesky-ai
13dAn extra fuck you to the people defending them using Claude to fix a bug. When you enable them to use AI, they'll use it for more than a si…An extra fuck you to the people defending them using Claude to fix a bug. When you enable them to use AI, they'll use it for more than…
↗Not briefed
hn
hn-ai
13dA Git hook to augment Claude-code attribution—
↗Not briefed
bluesky
bluesky-ai
13dExcuse me? Oh, hell no. - Just before the long Labor Day weekend, OpenAI made waves by bringing in major reinforcements to its political s…Excuse me? Oh, hell no. - Just before the long Labor Day weekend, OpenAI made waves by bringing in major reinforcements to its politi…
↗Not briefed
hn
hn-ai
13dShow HN: Claude Opus 5.5 (Max) best at generating startup ideas—
↗Not briefed
bluesky
bluesky-ai
13dDeepSeek says its open-source agent app Harness hit 1.54M DAUs on November 4, less than three months after launch, and uses ~92T tokens a d…DeepSeek says its open-source agent app Harness hit 1.54M DAUs on November 4, less than three months after launch, and uses ~92T token…
↗Not briefed
hn
hn-ai
13dPeer-to-peer WebRTC connections with built-in signaling and discovery—
↗Not briefed
rss
techmeme
13dDeepSeek says its open-source agent app Harness hit 1.54M DAUs on November 4, less than three months after launch, and uses ~92T tokens a day, or ~60M per user (Saritha Rai/Bloomberg)Saritha Rai / Bloomberg: DeepSeek says its open-source agent app Harness hit 1.54M DAUs on November 4, less than three months after la…
↗Not briefed
arxiv
arxiv-ai
13dFinCutBench: A Continually Updated Benchmark for Measuring the Temporal Boundaries of LLM Financial KnowledgeIn financial backtesting and other point-in-time tasks, external inputs can be restricted by date, but model parameters may retain fin…
↗Not briefed
bluesky
bluesky-ai
13d@jscandid.bsky.social@jscandid.bsky.social
↗Not briefed
hn
hn-ai
13dGrote Reber and the 'Square Kilometre Array' Near Bothwell, Tasmania (2017)—
↗Not briefed
bluesky
bluesky-ai
13dClaude 3 Opus (Legacy High-Cost) vs Claude 3.7 Sonnet: Which LLM API actually protects your SaaS gross margin at 10,000 DAU? Here is the l…Claude 3 Opus (Legacy High-Cost) vs Claude 3.7 Sonnet: Which LLM API actually protects your SaaS gross margin at 10,000 DAU? Here is…
↗Not briefed
arxiv
arxiv-ai
13dFastGRT: Accelerating Training of 3D Gaussian Ray TracingRay Tracing enables rendering capabilities that are difficult to achieve with rasterization-based approaches such as 3D Gaussian Splat…
↗Not briefed
bluesky
bluesky-ai
13dDon't let the Ramburglar get you, lil' laptop!! #SamAltman #OpenAI #FuckAI www.dailygrail.com/2026/11/sund...Don't let the Ramburglar get you, lil' laptop!! #SamAltman #OpenAI #FuckAI www.dailygrail.com/2026/11/sund...
↗Not briefed
bluesky
bluesky-ai
13dThis is my experience as well. A collaborator used LLM to do some work which requires formulating a methodology and then code it up. I chec…This is my experience as well. A collaborator used LLM to do some work which requires formulating a methodology and then code it up. I…
↗Not briefed
bluesky
bluesky-ai
13d刚把本地 Agent 桥接和日常开发笔记同步到 Bluesky。最近大部分时间都在折腾本地 Coding Agent 的上下文裁剪、实机浏览器桥接,还有给 OpenWrt 写点顺手的小插件(比如刚把 Magpie 迁到了 LuCI 上)。比起看各种热闹的 AI 跑分榜单,还是更…刚把本地 Agent 桥接和日常开发笔记同步到 Bluesky。最近大部分时间都在折腾本地 Coding Agent 的上下文裁剪、实机浏览器桥接,还有给 OpenWrt 写点顺手的小插件(比如刚把 Magpie 迁到了 LuCI 上)。比起看各种热闹的 AI 跑分榜…
↗Not briefed
bluesky
bluesky-ai
13dThe main update is about the use of Agentic AI to convert legacy code bases to new hardware-agnostic languages (JAX, Julia, CUDA etc.). Thi…The main update is about the use of Agentic AI to convert legacy code bases to new hardware-agnostic languages (JAX, Julia, CUDA etc.)…
↗Not briefed
bluesky
bluesky-ai
13dOpenAI funding closes at $1.25T valuation OpenAI secured $11 billion in new funding, valuing the company at $1.25 trillion pre-money, belo…OpenAI funding closes at $1.25T valuation OpenAI secured $11 billion in new funding, valuing the company at $1.25 trillion pre-money,…
↗Not briefed
bluesky
bluesky-ai
13dEveryone sees what they want to see with this shit. It’s anecdotal. Probably fake. Maybe even LLM generated. Whatever, man. I know the stor…Everyone sees what they want to see with this shit. It’s anecdotal. Probably fake. Maybe even LLM generated. Whatever, man. I know the…
↗Not briefed
arxiv
arxiv-ai
13dSame Error, Different Recovery: How Delivery Changes LLM Agent BehaviorA tool-using language model must decide whether to retry a failed call, correct it, or stop and tell the user. The error text alone ma…
↗Not briefed
arxiv
arxiv-ai
13dQubit-to-Motion: Traceable Action Provenance in Hybrid QRL for Robotic Bin-Packing SequencingHybrid quantum reinforcement learning has been explored for robotic decision-making, but evaluations often emphasize task success, ret…
↗Not briefed
hn
hn-ai
13dShow HN: Minimalistic roguelike game about hiking in Clojure—
↗Not briefed
hn
hn-ai
13dNew method Binance and Crypto Drainer, +$70k a weekHi guys, the new crypto method just dropped! Get ready. I will guide you step by step with guaranteed income, now or never. Just me an…
↗Not briefed
bluesky
bluesky-ai
13dThis internal slide deck from Anthropic is such a trip!This internal slide deck from Anthropic is such a trip!
↗Not briefed
arxiv
arxiv-ai
13dRecursive Pricing and Incumbent Subtree Column Generation for Solving Optimal Classification TreesSolving optimal classification tree (OCT) problems is computationally challenging because branching, routing, and leaf-assignment deci…
↗Not briefed
bluesky
bluesky-ai
13dLike, in principle, I'm thinking I could grab a public/open LLM code and train something up using my REBOUND scripts. I'd hypothesise that …🔧 ChatGPTLike, in principle, I'm thinking I could grab a public/open LLM code and train something up using my REBOUND scripts. I'd hypothesise…
↗Not briefed
bluesky
bluesky-ai
13dGPT-4.1 Mini (75% Cache Discount) vs DeepSeek V3 (0324): Which LLM API actually protects your SaaS gross margin at 10,000 DAU? Here is the…GPT-4.1 Mini (75% Cache Discount) vs DeepSeek V3 (0324): Which LLM API actually protects your SaaS gross margin at 10,000 DAU? Here i…
↗Not briefed
bluesky
bluesky-ai
13dOpenAI is positioning itself as the essential engine for European AI innovation while simultaneously launching rival products that threaten…OpenAI is positioning itself as the essential engine for European AI innovation while simultaneously launching rival products that thr…
↗Not briefed
rss
bing-news-ai
13dCanada announces G7 security push on artificial intelligence, Russia and threats to North AmericaCanada is taking a much broader view of what counts as a security threat. Days after releasing a new national ...
↗Not briefed
rss
bing-news-ai
13dCanada announces G7 security push on artificial intelligence, Russia and threats to North AmericaCanada is taking a much broader view of what counts as a security threat. Days after releasing a new national ...
↗Not briefed
rss
bing-news-ai
13dCanada announces G7 security push on artificial intelligence, Russia and threats to North AmericaCanada is taking a much broader view of what counts as a security threat. Days after releasing a new national ...
↗Not briefed
arxiv
arxiv-ai
13dContinuous modeling of manipulative participation for stance analysis: Online discourse on the 2024 Street Animal Law in TürkiyeStudies of online political discourse commonly infer public opinion through stance analysis while implicitly assuming that observed us…
↗Not briefed
bluesky
bluesky-ai
13d😂😂
↗Not briefed
bluesky
bluesky-ai
13d...the cognitive task of parsing what code does is GENERIC, it's not just "how humans think" there is an element of "any system capable of …...the cognitive task of parsing what code does is GENERIC, it's not just "how humans think" there is an element of "any system capabl…
↗Not briefed
arxiv
arxiv-ai
13dBeyond the Transcript: Measuring and Improving the Use of Acoustic Information in Audio Language ModelsAudio language model benchmarks provide comprehensive evaluation across diverse tasks. However, some tasks can be answered from the tr…
↗Not briefed
hn
hn-ai
13dOptimizing Reference Deletions in Git—
↗Not briefed
bluesky
bluesky-ai
13dBut they didn't know how to code, so they used an LLM. I (kindly) pushed back. It's so important to learn how to code, especially when usin…But they didn't know how to code, so they used an LLM. I (kindly) pushed back. It's so important to learn how to code, especially when…
↗Not briefed
bluesky
bluesky-ai
13dOne of the small joys of my job has always been getting weird spam emails from cranks who think they've disproved Einstein or found mollusc…One of the small joys of my job has always been getting weird spam emails from cranks who think they've disproved Einstein or found mo…
↗Not briefed
bluesky
bluesky-ai
13dlink storage.courtlistener.com/recap/gov.us...link storage.courtlistener.com/recap/gov.us...
↗Not briefed
bluesky
bluesky-ai
13danthropic is expected to IPO any day now glhfanthropic is expected to IPO any day now glhf
↗Not briefed
bluesky
bluesky-ai
13dVision models that can help identify a bug or bird or plant leaf, tiny models that give video game npcs actual conversational life, transcr…Vision models that can help identify a bug or bird or plant leaf, tiny models that give video game npcs actual conversational life, tr…
↗Not briefed
bluesky
bluesky-ai
13dCorruption Was the number one issue in exit polls Anti data center sentiment was up there too. Schumer's daughter is a lobbyist for openAI…Corruption Was the number one issue in exit polls Anti data center sentiment was up there too. Schumer's daughter is a lobbyist for o…
↗Not briefed
bluesky
bluesky-ai
13dthe fact that big yud not only still has a job but is still taken seriously by a nonzero amount of people despite the LLM obliterating all …the fact that big yud not only still has a job but is still taken seriously by a nonzero amount of people despite the LLM obliterating…
↗Not briefed
arxiv
arxiv-ai
13dTreating the Model as an Untrusted Component: Deterministic Policy Enforcement, Human Approval and Taint Tracking Against Prompt Injection in Tool-Using Agents, with Validation on Open-Weights ModelsTool-using LLM agents act on instructions embedded in the data they read, so prompt injection is a structural rather than a prompt-wor…
↗Not briefed
bluesky
bluesky-ai
13dLM FUCKING AO 🤣 github.com/IainKay/ram-...LM FUCKING AO 🤣 github.com/IainKay/ram-...
↗Not briefed
bluesky
bluesky-ai
13dto the community like the amount of vibe coded netplay "projects" that mostly rely on Claude to do it's interpretation of rollback netcode …to the community like the amount of vibe coded netplay "projects" that mostly rely on Claude to do it's interpretation of rollback net…
↗Not briefed
bluesky
bluesky-ai
13dIt makes sense to me (management don't want it publicised that their shift to all AI code was bad for business). LLM text tends to be cloyi…It makes sense to me (management don't want it publicised that their shift to all AI code was bad for business). LLM text tends to be…
↗Not briefed