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 | And so: "is an LLM more akin to a realistic doll or to an animal?" is I think a question worth continually picking at. Though I get why eve… | And so: "is an LLM more akin to a realistic doll or to an animal?" is I think a question worth continually picking at. Though I get wh… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Yeah, it can’t be understated that people are willingly feeding sensitive data to AI as training data (often by mistake through opt-out dat… | Yeah, it can’t be understated that people are willingly feeding sensitive data to AI as training data (often by mistake through opt-ou… | ↗Not briefed |
hn hn-ai | 1mo | CWG Chain – free arbitrage calculator that walks live order books | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | MemCo: Memory-Centric Collaboration for Generalizing LLM Agents to Unseen Environments | Large language model (LLM) agents increasingly operate in interactive environments, where they need to make sequential decisions throu… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Multigroup Fairness and Omniprediction: Separations and Equivalences | Omniprediction is a learning guarantee which requires a single predictor to be competitive relative to the best hypothesis from a benc… | ↗Not briefed |
rss techmeme | 1mo | SpaceX stock closes up 7.6% after Morgan Stanley called SPCX "cheap", reaching its highest level since mid-June and returning Elon Musk to trillionaire status (Lora Kolodny/CNBC) | Lora Kolodny / CNBC: SpaceX stock closes up 7.6% after Morgan Stanley called SPCX “cheap”, reaching its highest level since mid-June a… | ↗Not briefed |
bluesky bluesky-ai | 1mo | In a very similar vein, he told VF's Mark Guiducci, "We're willing to accept some things going wrong" (and, in another comment referring mo… | In a very similar vein, he told VF's Mark Guiducci, "We're willing to accept some things going wrong" (and, in another comment referri… | ↗Not briefed |
bluesky bluesky-ai | 1mo | “But doctor, I -AM- the AI-generated OnlyFans model that steals jokes!” | “But doctor, I -AM- the AI-generated OnlyFans model that steals jokes!” | ↗Not briefed |
bluesky bluesky-ai | 1mo | The harness teams have the unenviable position of having to role model/dogfood what high-execution software delivery teams look like in the… | The harness teams have the unenviable position of having to role model/dogfood what high-execution software delivery teams look like i… | ↗Not briefed |
bluesky bluesky-ai | 1mo | There are some use cases of the LLMs (AI), but here are 10 reasons NOT to use one for editing. quadlings.com/10-reasons-t... | There are some use cases of the LLMs (AI), but here are 10 reasons NOT to use one for editing. quadlings.com/10-reasons-t... | ↗Not briefed |
bluesky bluesky-ai | 1mo | This. Anthropic fucked up on the _process_ and they get no sympathy from me, but a hell of a lot of people here seem to think that it's sel… | This. Anthropic fucked up on the _process_ and they get no sympathy from me, but a hell of a lot of people here seem to think that it'… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI doesn't want you to talk about gay furry cryptography bloggers fouroclockfarms.club/@MontgomeryG... | OpenAI doesn't want you to talk about gay furry cryptography bloggers fouroclockfarms.club/@MontgomeryG... | ↗Not briefed |
bluesky bluesky-ai | 1mo | Mark Guiducci: “[Laura Reiley's] daughter committed suicide after speaking to ChatGPT. ChatGPT did not tell her to kill herself—” OpenAI P…🔧 ChatGPT | Mark Guiducci: “[Laura Reiley's] daughter committed suicide after speaking to ChatGPT. ChatGPT did not tell her to kill herself—” Ope… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Our company is now on the stage of AI adoption where the first bill has come due so they’re focused on “token usage optimization”. Apparent… | Our company is now on the stage of AI adoption where the first bill has come due so they’re focused on “token usage optimization”. App… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI will start watermarking ChatGPT’s text in the EU🔧 ChatGPT | OpenAI will start watermarking ChatGPT’s text in the EU | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anthropic: Could AI be conscious? Pope Leo: lol, no ‘Algorithms lack the spark of humanity’ https://www.youtube.com/watch?v=9VCWgoIrRCo&l… | Anthropic: Could AI be conscious? Pope Leo: lol, no ‘Algorithms lack the spark of humanity’ https://www.youtube.com/watch?v=9VCWgoIr… | ↗Not briefed |
bluesky bluesky-ai | 1mo | thinking more about this and I think the extra layer of disintermediation is key: property testing can and does find security-sensitive bug… | thinking more about this and I think the extra layer of disintermediation is key: property testing can and does find security-sensitiv… | ↗Not briefed |
rss techcrunch-ai | 1mo | OpenAI will start watermarking ChatGPT’s text in the EU🔧 ChatGPT | OpenAI will watermark ChatGPT and Codex text in the EU to comply with the AI Act. Editing can make the invisible marks harder to detec… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I wish people who used non-theft-based machine learning models would quit calling them "AI." Dude, I *want* to hear about *those* results.… | I wish people who used non-theft-based machine learning models would quit calling them "AI." Dude, I *want* to hear about *those* res… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Identity-Conditioned Score Fusion for Open-Set Person Re-Identification | Robust person re-identification often combines complementary cues such as face, gait, and body shape. While adaptive fusion typically… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Who Wrote It Is Not Enough: Detecting Who Contributed the Insight | As LLMs increasingly assist scientific writing and peer review, detecting who wrote the text is no longer sufficient: we need to deter… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Model AI after female intelligence or else. | Model AI after female intelligence or else. | ↗Not briefed |
bluesky bluesky-ai | 1mo | Some of you may die, but that's a sacrifice I'm willing to make. | Some of you may die, but that's a sacrifice I'm willing to make. | ↗Not briefed |
bluesky bluesky-ai | 1mo | 最初のプロトタイプ作ったBorisがAnthropicジョインしたの2024年9月なので彼でも2YoEしかありません... newsletter.pragmaticengineer.com/p/how-claude... | 最初のプロトタイプ作ったBorisがAnthropicジョインしたの2024年9月なので彼でも2YoEしかありません... newsletter.pragmaticengineer.com/p/how-claude... | ↗Not briefed |
arxiv arxiv-ai | 1mo | Dynamic Budget Allocation for LLM Evaluation under Hard Resource Constraints | We evaluate large language models (LLMs) in multi-turn interactions through their time-to-event: the number of interaction steps requi… | ↗Not briefed |
bluesky bluesky-ai | 1mo | RFK Jr. may not be so sanguine about AI in healthcare after he reads this. | RFK Jr. may not be so sanguine about AI in healthcare after he reads this. | ↗Not briefed |
bluesky bluesky-ai | 1mo | would push back against the assertion that brains do not process information or that they’re not computers, though — an LLM can meaningfull… | would push back against the assertion that brains do not process information or that they’re not computers, though — an LLM can meanin… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Evaluate the Stack, Not the Layer: Do Deterministic and LLM Gates for Agent Actions Fail Independently? | Runtime gates for agent tool calls are stacked on the assumption that their errors multiply. We test it on 1,119 labelled agent action… | ↗Not briefed |
bluesky bluesky-ai | 1mo | this is fundamentally the same relationship the materials in a training corpus have to the weights in the LLM that training process produce… | this is fundamentally the same relationship the materials in a training corpus have to the weights in the LLM that training process pr… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Towards Explainable Benchmarking for Data-driven Post-Wildfire Debris Flow Prediction | Post-wildfire debris flows (PFDFs) are destructive sediment-laden hazards triggered when intense rainfall strikes recently burned terr… | ↗Not briefed |
hn hn-ai | 1mo | Give your LLM more DSLs | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Whoa wait are you questioning the staggering genius of folks who stole the entire world’s collective intellectual property and managed to t… | Whoa wait are you questioning the staggering genius of folks who stole the entire world’s collective intellectual property and managed… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 🌎 Welcome LGND as a Bronze Sponsor of #FOSS4GNA LGND builds #AI that understands the planet — turning #satellite imagery into searchable … | 🌎 Welcome LGND as a Bronze Sponsor of #FOSS4GNA LGND builds #AI that understands the planet — turning #satellite imagery into search… | ↗Not briefed |
arxiv arxiv-ai | 1mo | A Validated Dataset and Benchmark for Coherent Multi-Diagram SysML Models | Systems engineers use several diagrams to describe the structure and behavior of systems. Engineers create these diagrams together to… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Tracking Is Not Permanence: What Video World Models Keep of a Hidden Object | Video world models track objects they can see; we ask what they keep of objects they cannot. We hide an object from a frozen V-JEPA 2… | ↗Not briefed |
hn hn-ai | 1mo | Ask HN: Feedback regarding my AI setup | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Evaluating Escalation Signals for LLM Routing: Targets, Controls, and Five Ways to Fool Yourself | Deciding when to escalate a query from a small language model to a larger one requires a cheap signal that predicts, before the large… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Trajectory-Retrieval Speculative Decoding: When Does a Model's Own History Help? | Long chain-of-thought reasoning increases sequential decoding cost while creating a growing history of potentially reusable continuati… | ↗Not briefed |
arxiv arxiv-ai | 1mo | RELACE: retrospective likelihood-based action credit estimation for long-horizon language agents | Group Relative Policy Optimization (GRPO) avoids a separate critic by estimating advantages from rollout groups. For multi-turn agents… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Agree, security issues predate and are wider than AI incursions - but also it doesn't need an LLM to break into a secure database if someon… | Agree, security issues predate and are wider than AI incursions - but also it doesn't need an LLM to break into a secure database if s… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Stepped MoE: Segment-Level Routing with Configurable Inference Complexity | Training large language models (LLMs) is resource-intensive, and adapting them for diverse deployment scenarios with varying computati… | ↗Not briefed |
hn hn-ai | 1mo | Orcah Studio: A local-first video agent that can search your videos | Hey,<p>I'm Ilias, the founder of Orcah Studio. I'm building a local-first desktop app (Mac OS available now) that can transc… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Evaluating Behavioral Context for Interpretable IAM Policy Risk Scoring in Cloud Environments | IAM policy analysis typically emphasizes the authorization capabilities encoded in a policy, but security analyst review priority may… | ↗Not briefed |
bluesky bluesky-ai | 1mo | CM Kayla Santosuosso to OpenAI exec at City Council hearing on risk of AI: "Can you tell me how many people in your company are currently… | CM Kayla Santosuosso to OpenAI exec at City Council hearing on risk of AI: "Can you tell me how many people in your company are curr… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Rationale-Guided Policy Optimization: Learning to Reason with Adaptive Rationale Scaffolding | On-policy reinforcement learning has become a central paradigm for improving the reasoning abilities of large language models. However… | ↗Not briefed |
rss theverge-ai | 1mo | This startup is issuing AI-generated acne prescriptions | People in Utah can now use AI to get a prescription for acne treatment. On Monday, healthcare startup Nolla Health announced that user… | ↗Not briefed |
arxiv arxiv-ai | 1mo | CausalBind: Causal Modeling and Learning for Protein-Molecule Virtual Screening | Protein-molecule virtual screening is increasingly cast as a problem of representation learning in a shared embedding space. Existing… | ↗Not briefed |
arxiv arxiv-ai | 1mo | A doctrine-grounded visual question answering dataset for Tactical Combat Casualty Care | Tactical Combat Casualty Care (TC3) requires responders to connect visual observations of injuries and interventions with established… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Logbook: Extremely Long-form Audio Event Understanding | Audio benchmarks are built around short, pre-segmented clips, limiting model design to brief inputs or fixed vocabularies. To close th… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 🐭I deleted my post about SDL3 having claude in it because as it turns out, when I was looking it up I somehow got directed to the *wiki's*… | 🐭I deleted my post about SDL3 having claude in it because as it turns out, when I was looking it up I somehow got directed to the *wi… | ↗Not briefed |