terrific deep dive on Mythos from @cloudflare
terrific deep dive on Mythos from @cloudflare Cloudflare's security team spent the last few weeks testing Anthropic's Mythos against fifty of our own repositories. What we learned about offensive AI,
Knowledge catalogue
terrific deep dive on Mythos from @cloudflare Cloudflare's security team spent the last few weeks testing Anthropic's Mythos against fifty of our own repositories. What we learned about offensive AI,
arXiv:2605.16410v1 Announce Type: new Abstract: Test-time scaling (TTS) methods have proven highly effective for LLMs, yet their application to vision-language models (VLMs) remains relatively underex
arXiv:2605.17749v1 Announce Type: new Abstract: AI generated predictions increasingly inform decision making in critical tasks, and therefore must be trustworthy. One widely used measure of trustworth
arXiv:2605.18430v1 Announce Type: new Abstract: Text-to-CAD generation aims to create parametric CAD models from natural language, enabling rapid prototyping and intuitive design workflows. However, e
Google's I/O 2026 keynote today was once again full of AI-related announcements including a new family of Gemini 3.5 AI models, new features for Search and Gmail, and updates about its Project Aura sm
This is a new era — the agentic era – and the question is no longer, “what’s possible?” but rather, “what creates impact?” Today, organizations across industries around the world are swiftly moving fr
arXiv:2603.01092v2 Announce Type: replace Abstract: Scientific discovery is constrained not only by what is true, but by what is cognitively available to the researchers currently exploring a field. M
arXiv:2605.16895v1 Announce Type: cross Abstract: End-to-end LLM trading agents have moved quickly from research curiosity to a small ecosystem of named systems, including FinCon, FinMem, TradingAgent
arXiv:2512.22471v5 Announce Type: replace-cross Abstract: Transformers often appear to perform Bayesian reasoning in context, but verifying this rigorously has been impossible: natural data lack analy
arXiv:2605.17480v1 Announce Type: new Abstract: Multi-agent systems extend large language models (LLMs) by decomposing tasks among specialized agents, but their distributed decision process creates ne
The Commonwealth Prize response to the AI-written story (and the fact that they still don't realize that there are at least 2 AI-written prize winning stories!) is a great example of the wide gaps in
The crazy thing is that Antigravity does this quite well! So it isn't like Google is hiding these thinking traces because of distillation or because they are full of insane mutterings. I guess you nee
The daddy of baby AGI articulates what we all feel is missing in the current state of AI architecture I didn’t really understand what people meant by stateful agents so I started exploring my current
arXiv:2602.21185v2 Announce Type: replace Abstract: Uniform-state discrete diffusion models excel at few-step generation and guidance due to their ability to self-correct, making them preferred over a
This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. Here’s why Elon Musk lost his suit against OpenAI Elon Musk ha
arXiv:2605.16436v1 Announce Type: cross Abstract: For decades, the security of digital interaction has rested on an unacknowledged economic constraint. Attackers faced a tradeoff between the fidelity
arXiv:2604.02178v2 Announce Type: replace-cross Abstract: Mixture-of-Experts (MoE) architectures have become the dominant choice for scaling Large Language Models (LLMs), activating only a subset of p
arXiv:2605.18079v1 Announce Type: cross Abstract: Existing expressivity results for transformers typically rely on hardmax attention, high precision, and other architectural modifications that disconn
arXiv:2605.06506v2 Announce Type: replace Abstract: Language-model (LM) surprisal is widely used as a proxy for contextual predictability and has been reported to correlate with metaphor novelty judgm
the future is bright, lets get to work Personal update: I've joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join the team here
The modern software development landscape isn’t happening just on one surface — it’s happening across an entire ecosystem of agentic tools. Agents are being developed at an unprecedented scale, and th
Last year, after watching Google's I/O keynote, I wrote that it felt like Google's future was Google googling. After watching this year's I/O keynote on Tuesday, I don't think Google just wants to goo
The gap between what you can do on ChatGPT/Codex and Claude/Code/Cowork is closing, as Anthropic & OpenAI converge on a single experience. Google's experiences are diverging: Studio & Gemini & Antigra
arXiv:2605.17180v1 Announce Type: new Abstract: We develop a geometric theory of projection heads in self-supervised learning by modeling the head as a trainable Riemannian metric on the backbone repr
arXiv:2605.18372v1 Announce Type: cross Abstract: Large Language Models (LLMs) are increasingly used in educational settings as interactive tools for collaboration. However, their tendency toward syco
arXiv:2601.03425v2 Announce Type: replace-cross Abstract: Mixture of Experts models are widely assumed to achieve domain specialization through sparse routing. In this work, we question this assumptio
arXiv:2605.16428v1 Announce Type: cross Abstract: Search engines traditionally complement online content platforms by directing users seeking information to external websites. The emergence of generat
arXiv:2605.17008v1 Announce Type: cross Abstract: We introduce IsalProgram (Instruction Set and Language for Programming), a novel assembly-like programming language with three distinctive theoretical
arXiv:2511.08639v3 Announce Type: replace-cross Abstract: Existing AI disclosure mandates in scholarship require that AI assistance be reported but leave transparency philosophically unspecified: they
arXiv:2602.07730v2 Announce Type: replace-cross Abstract: Across scientific disciplines, Laplacian eigenvectors serve as a fundamental basis for simplifying complex systems, from signal processing to
I put together these annotated slides from my five minute lightning talk at PyCon US 2026, using the latest iteration of my annotated presentation tool. # I presented this lightning talk at PyCon US 2
arXiv:2603.01227v2 Announce Type: replace Abstract: We propose the Lattice Representation Hypothesis of large language models: a symbolic backbone that grounds conceptual hierarchies and logical opera
arXiv:2605.17087v1 Announce Type: new Abstract: Generative data augmentation with latent diffusion models is a promising strategy for addressing class imbalance in medical imaging, yet current approac
arXiv:2508.16663v2 Announce Type: replace-cross Abstract: Fine-Grained Visual Classification (FGVC) requires models to focus on subtle, task-relevant regions rather than broad object context. We prese
arXiv:2605.18063v1 Announce Type: new Abstract: Object counting is a foundational vision task with over a decade of dedicated research, yet state-of-the-art models still fail systematically in the mix
arXiv:2605.17606v1 Announce Type: new Abstract: In wide neural networks, the Neural Tangent Kernel (NTK) remains approximately constant during training, providing a powerful theoretical tool for study
arXiv:2309.01243v4 Announce Type: replace-cross Abstract: We investigate the privacy of {em any} algorithm whose outputs have Gaussian distribution. This work is motivated by the prevalence of such al
The Odyssey and the Iliad get so many movie treatments but the sequel, the Roman Aeneid, is entirely ignored. Here is a teaser trailer from one prompt to Gemini Omni. The first pass made all the flags
The official Pinecone plugin for @cursor_ai has been released! To install, run /add-plugin pinecone within Cursor. What you get: Agent Skills and scripts to help you build with our vector database and
arXiv:2605.17113v1 Announce Type: cross Abstract: Existing deception datasets label completed outputs as honest or deceptive, treating deception as a property of the final response rather than a funct
arXiv:2601.21170v3 Announce Type: replace Abstract: Emergent phenomena -- onset of epileptic seizures, sudden customer churn, or pandemic outbreaks -- often arise from hidden causal interactions in co
arXiv:2605.17062v1 Announce Type: cross Abstract: Spracklen et al. (USENIX Security '25) showed that code-generating large language models hallucinate package names that do not exist on PyPI or npm at
arXiv:2605.16283v1 Announce Type: cross Abstract: For centuries, each new technology has automated some layer of cognitive work and been absorbed by education retreating upward to teach the skills mac
arXiv:2605.16508v1 Announce Type: cross Abstract: As agent systems scale, skills accumulate into large reusable libraries, yet their scaling laws remain poorly understood. Across 15 frontier LLMs, 1,1
arXiv:2605.16644v1 Announce Type: cross Abstract: A central obstacle in nonlinear Bayesian filtering is representing the belief distribution. Moment-based filters address this by propagating polynomia
arXiv:2605.17500v1 Announce Type: cross Abstract: Generative text-to-image models are typically trained on large-scale web-scraped datasets that include diverse visual content such as copyrighted and
arXiv:2605.18319v1 Announce Type: new Abstract: We develop a framework for analyzing parameter symmetries in deep ReLU networks and obtain a complete characterization of the generic parameter fibers f
🚀 The team at @Google just released the Agents API, a service for building and running custom agents inside a sandboxed Linux environment, and we built a template that gives these agents access to Lla
arXiv:2602.17831v2 Announce Type: replace Abstract: Evaluating the reasoning capabilities of Large Language Models is increasingly challenging as models improve. Human curation of hard questions is hi
This post presents statistics claiming that the top 1% of earners receive 22.4% of total income while paying 40.4% of income taxes, suggesting a disproportionate tax burden relative to their share of
At Google I/O, two dominant branding phrases emerged: 'agent-first,' emphasizing AI agents as a primary computing paradigm, and 'at the speed of voice,' highlighting voice-based interaction speed and
“the uncritical adoption of AI in science is alarming” Artificial intelligence is rapidly accelerating scientific output, but risks narrowing inquiry, weakening judgement and undermining how scientist
arXiv:2605.16787v1 Announce Type: cross Abstract: Reinforcement Learning with Verifiable Reward (RLVR) has proven effective in improving Large Language Model's (LLM) reasoning ability. However, the le
arXiv:2601.13992v2 Announce Type: replace-cross Abstract: Chain-of-Thought (CoT) reasoning empowers Large Language Models (LLMs) with remarkable capabilities but typically requires prohibitive paramet
The @xai team has published a full setup guide on how to use the xurl skill, which allows your Hermes Agent to read and write to X on your behalf — posting, searching, pulling bookmarks, managing list
arXiv:2601.16880v2 Announce Type: replace Abstract: The minimal norm weight perturbations of DNNs required to achieve a specified change in output are derived and the factors determining its size are
arXiv:2507.16059v2 Announce Type: replace Abstract: Following a stroke, individuals often experience mobility and balance impairments due to lower-limb weakness and loss of independent joint control.
there's 4 parts to this AI SDLC 1. have ~50 tests in place, with instructions to add more, including 'make a memory that whenever you do browser e2e tests, use computer vision to visually spot check d
arXiv:2605.17042v1 Announce Type: new Abstract: While RGB-Thermal crowd counting has shown promise, the paradigm faces critical limitations: RGB data raises privacy concerns in public surveillance, an
arXiv:2605.16848v1 Announce Type: cross Abstract: Planning from raw visual input remains a significant challenge for current Vision-Language Models (VLMs), when the complexity of input is beyond their