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 | 2mo | A Dominant Diffuse Phase in the Sparse Autoencoder Phase Diagram | Sparse autoencoders (SAEs) are increasingly used to recover interpretable features from neural-network activations, yet systematic fea… | ↗Not briefed |
rss techcrunch-ai | 2mo | Viral AI assistant Instinct now has its own email address | Instinct’s new email feature lets the AI agent create and manage accounts, contact businesses, handle support requests, and do more on… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Anthropic and OpenAI’s comms strategy: | Anthropic and OpenAI’s comms strategy: | ↗Not briefed |
arxiv arxiv-ai | 2mo | Learning Intrusion Response Strategies for OT Systems | Cyberattacks against Operational Technology (OT) systems, which monitor and control industrial processes, pose an increasing threat to… | ↗Not briefed |
bluesky bluesky-ai | 2mo | What is the difference between an AI agent and an LLM? As I understand it, an AI agent can interact with other programs. For instance, on a… | What is the difference between an AI agent and an LLM? As I understand it, an AI agent can interact with other programs. For instance,… | ↗Not briefed |
arxiv arxiv-ai | 2mo | The Semantic Bottleneck: Leveraging Semantic Representations for Non-Invasive Speech Decoding | Non-invasive speech decoding remains constrained by the low signal-to-noise ratio of neural recordings, which makes fine-grained recon… | ↗Not briefed |
bluesky bluesky-ai | 2mo | OpenAI researcher: "at the current frankly terrifying pace humanity will be quite lucky if we manage to find and stay on the narrow path be… | OpenAI researcher: "at the current frankly terrifying pace humanity will be quite lucky if we manage to find and stay on the narrow pa… | ↗Not briefed |
hn hn-ai | 2mo | FinOps for LLM Spend – why your bill has changed | I wrote a post which may help in identifying what may have caused savings on the LLM bill. https://leanroute.dev/blog&… | ↗Not briefed |
hn hn-ai | 2mo | Anthropic researcher says more than 10% chance AI "could kill all humans" | — | ↗Not briefed |
bluesky bluesky-ai | 2mo | AI researcher Jacob Coxon wrote a viral social post urging colleagues to rethink their work, warning that Anthropic and OpenAI are irrespon… | AI researcher Jacob Coxon wrote a viral social post urging colleagues to rethink their work, warning that Anthropic and OpenAI are irr… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Jacob Coxon accused OpenAI and Anthropic of “racing straight to self-improving superintelligence and gambling with our lives,” even though … | Jacob Coxon accused OpenAI and Anthropic of “racing straight to self-improving superintelligence and gambling with our lives,” even th… | ↗Not briefed |
bluesky bluesky-ai | 2mo | (Hubinger is „Alignment Science Lead“ at #Anthropic) | (Hubinger is „Alignment Science Lead“ at #Anthropic) | ↗Not briefed |
bluesky bluesky-ai | 2mo | The one thing I think everyone should be able to agree on is that viewing the race to dominate AI as more important than literally every ot… | The one thing I think everyone should be able to agree on is that viewing the race to dominate AI as more important than literally eve… | ↗Not briefed |
hn hn-ai | 2mo | Show HN: Give your AI agent on-screen guides that show users where to click | Hey HN. I'm Christian, one of the founders of Frigade (YC W23). I've noticed that a lot of in-app AI agents struggle to actu… | ↗Not briefed |
bluesky bluesky-ai | 2mo | "We’re on track for a lot of the most aggressive of these scenarios where by the end of next year things could be out of control already." | "We’re on track for a lot of the most aggressive of these scenarios where by the end of next year things could be out of control alrea… | ↗Not briefed |
arxiv arxiv-ai | 2mo | GANDR: Claim Auditing for Verifiable Legal Answer Generation | In high-stakes domains such as legal practice, a language-model answer is only useful to the extent that a reader can verify each clai… | ↗Not briefed |
bluesky bluesky-ai | 2mo | 「こんにちは。ChatGPT、Gemini、Claudeです」 「西方より来たれし三賢者ね」 「こんにちは。DeepSeek、Qwen、ELYZAです」 「東方より来たれし三賢者だ!」🔧 ChatGPT | 「こんにちは。ChatGPT、Gemini、Claudeです」 「西方より来たれし三賢者ね」 「こんにちは。DeepSeek、Qwen、ELYZAです」 「東方より来たれし三賢者だ!」 | ↗Not briefed |
bluesky bluesky-ai | 2mo | Anthropic researchers warn out-of-control AI could destroy humans #TheHill | Anthropic researchers warn out-of-control AI could destroy humans #TheHill | ↗Not briefed |
bluesky bluesky-ai | 2mo | I will say it is funny watching OpenAI boast about solving a complex math problem using millions of dollars of compute (and someone else’s … | I will say it is funny watching OpenAI boast about solving a complex math problem using millions of dollars of compute (and someone el… | ↗Not briefed |
hn hn-ai | 2mo | Show HN: Adding initContainers and env vars to deployments w/o Kyverno | — | ↗Not briefed |
bluesky bluesky-ai | 2mo | social media posts by anthropic employees is why mature corporations have social media policies. | social media posts by anthropic employees is why mature corporations have social media policies. | ↗Not briefed |
arxiv arxiv-ai | 2mo | Training Trajectories Determine Circuit Removability in Annealable Soft-Prior Transformers | Soft positional priors can help small Transformers learn retrieval circuits, but it is unclear whether the resulting circuits remain f… | ↗Not briefed |
bluesky bluesky-ai | 2mo | A.I. is going to kill us all. So that's fun. apple.news/ASCNgXUEcQeW... | A.I. is going to kill us all. So that's fun. apple.news/ASCNgXUEcQeW... | ↗Not briefed |
bluesky bluesky-ai | 2mo | What the FUCK!! | What the FUCK!! | ↗Not briefed |
bluesky bluesky-ai | 2mo | „Guy from Anthropic says their AI is crazy cool and scary“ ok that’s just them advertising lol | „Guy from Anthropic says their AI is crazy cool and scary“ ok that’s just them advertising lol | ↗Not briefed |
bluesky bluesky-ai | 2mo | Oh swell and here I thought brain cancer was my biggest worry😡 I've never been in favor of AI and the way governments are racing to see wh… | Oh swell and here I thought brain cancer was my biggest worry😡 I've never been in favor of AI and the way governments are racing to s… | ↗Not briefed |
bluesky bluesky-ai | 2mo | if openai has spare capacity why arent they giving me free usage | if openai has spare capacity why arent they giving me free usage | ↗Not briefed |
rss techcrunch-ai | 2mo | ‘Gambling with our lives’: Anthropic researcher quits, warns against self-improving AI | Anthropic researcher Jacob Coxon resigned over AI extinction fears, calling for pacing agreements between labs. | ↗Not briefed |
bluesky bluesky-ai | 2mo | “Increasingly, writers who use no AI at all are beginning to worry about their work containing an inadvertent AI tell.” | “Increasingly, writers who use no AI at all are beginning to worry about their work containing an inadvertent AI tell.” | ↗Not briefed |
bluesky bluesky-ai | 2mo | Every single techbro that was sniggering behind their hands at Extinction Rebellion in 2019 is now sitting the corner rocking back and fort… | Every single techbro that was sniggering behind their hands at Extinction Rebellion in 2019 is now sitting the corner rocking back and… | ↗Not briefed |
bluesky bluesky-ai | 2mo | But just like you don’t trust the spell check on your word processing computer to be the final word on a written questionnaire, you don’t t… | But just like you don’t trust the spell check on your word processing computer to be the final word on a written questionnaire, you do… | ↗Not briefed |
bluesky bluesky-ai | 2mo | OpenAI eyes Samsung for future chip production OpenAI's Korea GM Harrison Kim revealed they are considering Samsung Foundry for next-gen A… | OpenAI eyes Samsung for future chip production OpenAI's Korea GM Harrison Kim revealed they are considering Samsung Foundry for next-… | ↗Not briefed |
bluesky bluesky-ai | 2mo | If the announcement is confirmed, it would mark the first time AI has solved a truly major open problem in math. But a fierce dispute has e… | If the announcement is confirmed, it would mark the first time AI has solved a truly major open problem in math. But a fierce dispute… | ↗Not briefed |
arxiv arxiv-ai | 2mo | Structural Fusion of Bayesian Networks with Limited Treewidth Using Genetic Algorithms🔧 Consensus | This paper introduces an evolutionary computation approach for consensus in structural Bayesian Network (BN) fusion under the constrai… | ↗Not briefed |
bluesky bluesky-ai | 2mo | . More people need to be crashing the fuxkkkkk out about the fact that #JacobCoxon a senior researcer at AI companies #Anthropic & #OpenAI… | . More people need to be crashing the fuxkkkkk out about the fact that #JacobCoxon a senior researcer at AI companies #Anthropic & #O… | ↗Not briefed |
bluesky bluesky-ai | 2mo | I know this is a little controversial on here, but I believe that these guys at OpenAI and Anthropic believe their own bullshit and are (no… | I know this is a little controversial on here, but I believe that these guys at OpenAI and Anthropic believe their own bullshit and ar… | ↗Not briefed |
hn hn-ai | 2mo | Geiger – See every AI agent on your machine and what it can touch | — | ↗Not briefed |
bluesky bluesky-ai | 2mo | Anthropic staffer says the safety plan for AI is to hope AIs are better at alignment training than the humans that created them: “Insofar… | Anthropic staffer says the safety plan for AI is to hope AIs are better at alignment training than the humans that created them: “In… | ↗Not briefed |
arxiv arxiv-ai | 2mo | KVShareArena: KV-Cache Reuse Across Contexts and Model Checkpoints | LLM serving systems already reuse KV caches, but only when the reused text sits at the very start of the prompt. Two growing workloads… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Anthropic Is building a surveillance system to monitor activists using a “pre-crime” approach that tries to predict incidents before they h… | Anthropic Is building a surveillance system to monitor activists using a “pre-crime” approach that tries to predict incidents before t… | ↗Not briefed |
bluesky bluesky-ai | 2mo | “What’s particularly notable is the terms by which engineers are working in lockstep with the department of war to engineer AI systems for … | “What’s particularly notable is the terms by which engineers are working in lockstep with the department of war to engineer AI systems… | ↗Not briefed |
hn hn-ai | 2mo | A Git Config Key Ran Code in Seven Coding Agents. The 2022 Fix Does Not Stop It | — | ↗Not briefed |
arxiv arxiv-ai | 2mo | Maverick: Private and Verifiable LLM Inference Made Practical via Matrix-Vector Multiplication Delegation | Open-source large language models (LLMs) are increasingly competitive with closed-source models while offering transparency and the ab… | ↗Not briefed |
bluesky bluesky-ai | 2mo | OpenAI claims to have created a proof of Navier-Stokes stokes equations break down using internal “gen ai” models & agents. BUT! May have… | OpenAI claims to have created a proof of Navier-Stokes stokes equations break down using internal “gen ai” models & agents. BUT! May… | ↗Not briefed |
bluesky bluesky-ai | 2mo | the AI model is always racist | the AI model is always racist | ↗Not briefed |
rss techcrunch-ai | 2mo | Shipt becomes the latest delivery app with an AI shopping assistant | Users can ask the assistant to do things like "Create a cart for my Saturday tailgate for 25 people and include some brunch items," or… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Like, an LLM isn't embodied, it doesn't *really* have a model of the outside world, it doesn't have a consistent identity over time and has… | Like, an LLM isn't embodied, it doesn't *really* have a model of the outside world, it doesn't have a consistent identity over time an… | ↗Not briefed |
bluesky bluesky-ai | 2mo | I wish I could be excited for this and Layton as well, but given that Level 5 is deep in the ragu when it comes to adopting LLM slop, I jus… | I wish I could be excited for this and Layton as well, but given that Level 5 is deep in the ragu when it comes to adopting LLM slop,… | ↗Not briefed |
hn hn-ai | 2mo | Knowledge vs. wisdom: asking AI "What mushroom is that?" | — | ↗Not briefed |
bluesky bluesky-ai | 2mo | US-Behörden warnen: Chinesische KI-Anbieter wie DeepSeek und Alibaba sollen massiv Wissen aus US-Modellen wie denen von OpenAI oder Google … | US-Behörden warnen: Chinesische KI-Anbieter wie DeepSeek und Alibaba sollen massiv Wissen aus US-Modellen wie denen von OpenAI oder Go… | ↗Not briefed |