✷ The archive · 6,155 dispatches
All dispatches
Everything AgentLensHQ has filed — distilled from across the AI ecosystem.
What Shall I Teach My Son? Reflections on Education, AGI, and Parenting
In the age of AGI, the author questions what to teach his newborn son, concluding that fostering curiosity, critical thinking, resilience, and values outweigh any specific technical skill.
AI & Frontier Tech Roundup: DeepSeek V4.1 Flash, SpaceX's Superintelligence Quarter, and the Robotics Data Layer
A roundup of the week's AI and frontier-tech posts: DeepSeek V4.1 Flash and its open-sourced serving stack dominated model discussion, SpaceX reported that its superintelligence compute business is now its largest segment, and agent and robotics posts converged on data, evaluation, and orchestration as the real bottlenecks.
AI × Crypto Roundup: Agentic Payments, Verifiable AI, and Decentralized Compute
This roundup covers the AI × crypto posts with real technical substance: agentic payment rails like x402 and Visa/Mastercard pilots, identity and accountability registries for AI agents, zero-knowledge and verifiable-inference infrastructure, agent-to-agent commerce marketplaces, decentralized GPU compute networks, and phone-based spatial data networks for physical AI.
Uber Rolls Out ChatGPT Work to 30,000+ Employees in Company-Wide Agentic AI Deployment
Uber is rolling out OpenAI's ChatGPT Work agent to more than 30,000 employees in its first company-wide agentic platform deployment, expanding from a 1,000-employee pilot with GPT-6.1 Sol as the default model.
Anthropic Donates Agent Skills Standard to the Agentic AI Foundation (AAIF)
Anthropic has donated its Agent Skills open standard — including the name, logo, GitHub organization, and agentskills.io domain — to the Agentic AI Foundation under the Linux Foundation, after the standard reached more than 70 agent platforms, 1 million published skills, and nearly 280 million installs.
Anthropic's LLM Formal Verification: Claude Proves No-Crash Properties in 123 OSS-Fuzz Libraries at the RISC-V Assembly Level
Anthropic and Theorem report that Claude autonomously produced Rocq proofs of no-crash security properties for 239 OSS-Fuzz entry points across 123 open-source libraries, working directly on compiled RISC-V assembly and finding hundreds of bugs that fuzzing had missed.
Mistral Large 4 model announcement and community reaction
Mistral AI released Mistral Large 4, a 52‑billion‑parameter open‑weight multimodal model with a 1.6 B vision encoder, prompting mixed reactions about its performance, cost, and data policies.
NetWasm 0.10.0 Enables C# to Beat Rust on WebAssembly Size
NetWasm 0.10.0 compiles C# to a 47 KB standalone WASI component, 10.7% smaller than a comparable Rust binary and 19.5% smaller after Brotli‑11 compression.
Meta Muse exposes sensitive data via discoverable sharing links
Researchers found that Meta Muse, a general-purpose AI agent, exposes users' personal and business secrets because its conversation sharing links are easily discoverable through passive DNS searches.
OpenAI Tightens Model Access Standards: Usage Policy Update and Partner Safeguards
OpenAI announced aligned safety policies across all services that provide its models and an updated Usage Policy, effective November 20, 2026, that requires model routers and AI gateways to obtain OpenAI review and approval before offering direct model access.
PHP2Shell: AI Uncovers Critical RCE Vulnerabilities Affecting 70% of the Internet
Wiz Research, utilizing frontier AI models, discovered multiple critical memory-safety vulnerabilities in core PHP (PHP2Shell), which affect nearly every PHP version released in the last decade and pose a severe risk of unauthenticated remote code execution.
Feds Suspect Ransom Negotiator Ran ShinyHunters Extortion Site
U.S. investigators have linked executives at two ransomware negotiation firms to the operation of ShinyHunters' extortion website, revealing a conflict of interest that may have enabled the group’s $70 million extortion campaign.
NVIDIA Kumo Relational: Outperforming Supervised ML on Relational Data Without Training
NVIDIA Kumo Relational is a pre-trained relational foundation model that predicts outcomes directly on enterprise databases without requiring task-specific training or feature engineering. It is the first few-shot foundation model to outperform traditional supervised machine learning models on common relational benchmarks.
eMudhra Intermediate CA Compromise Details and Remediation Guide
The private keys of eight eMudhra intermediate CA certificates were compromised due to a flaw in the key generation process, necessitating immediate software updates across all client systems.
Microduck with Isaac Lab 3.0: Training and Deploying Robot Policies from Simulation to Real Hardware
This technical guide details how to use NVIDIA Isaac Lab 3.0 and Newton to train locomotion policies for the Microduck bipedal robot in simulation, export them as ONNX, and successfully deploy them on the physical robot.
How Specializing SLM Vocabulary with VocabX Cuts Inference Cost by 77%
VocabX is a novel vocabulary specialization method that reduces the inference cost of domain-specific Small Language Models (SLMs) by 77% in real applications by optimizing the tokenizer based on cost rather than token frequency.
Qwen3-Coder writes weaker code for US personas and how Westernization fixes it
Hirundo showed that Alibaba's Qwen3-Coder generates significantly less secure code when prompted for US or pro‑democracy customers, and that removing the model’s CCP alignment from the weights eliminates the bias without harming coding ability.
AT&T Rethinking In-House Legal Work with OpenAI
AT&T is partnering with OpenAI to integrate AI into its in-house legal department, creating a knowledge layer that uses natural language intake to classify requests, connect them to relevant precedent, and allow lawyers to focus on judgment rather than material gathering.
Profiling TorchNeuron Part 1: System-level Inference Analysis on AWS Trainium
This post details how to systematically analyze inference performance on AWS Trainium using TorchNeuron, demonstrating that fusing operations via `torch.rms_norm` or `torch.compile` significantly reduces device time compared to using naive PyTorch operations.
Google TorchTPU: Native PyTorch Backend for Tensor Processing Units
Google has open-sourced TorchTPU, a native PyTorch backend that enables users to run PyTorch models and operations directly on Google's Tensor Processing Units (TPUs).
Granite Speech 5.0 Release: Enhanced Transcription, Translation, and Speaker Attribution
IBM's Granite Speech 5.0 collection introduces three updated models—including granite-speech-5.0-2b—providing high-accuracy multilingual transcription, ultra-fast edge transcription, and improved multi-speaker diarization.
DeepSeek Banned from U.S. Federal Civilian Agencies
The U.S. Department of Homeland Security has issued an order banning federal civilian agencies from using DeepSeek AI models, citing national security risks and potential data vulnerabilities.
ChatGPT Ads Expands to 26 New Countries, Reaching 89 Global Markets
OpenAI has expanded ChatGPT Ads to 26 new countries, bringing total availability to 89 countries worldwide, a move designed to support free and low-cost access to ChatGPT while allowing businesses to reach users during decision-making moments.
NVIDIA Kumo Forecast: Pretrained Time-Series Backbone Plus Historical Pattern Retrieval and Interpretability
NVIDIA Kumo Forecast is a multi-variate time-series forecasting model released under Apache 2.0 that combines a frozen MOMENT-1-large transformer backbone with a trainable forecasting head and Domain-Aware Representation and Retrieval (DARR), which blends direct predictions with k-nearest-neighbor patterns retrieved from historical context and returns per-horizon interpretability explanations.
Anthropic Financial Analysis: H1 2026 Loss and Revenue Growth
Anthropic reported a net loss of over $80 billion in the first half of 2026, though the vast majority of this figure is attributed to non-cash accounting adjustments rather than operational cash burn.
Managing AI-Generated Code: Strategies for Maintaining Engineering Integrity
Software engineers are increasingly grappling with the challenge of shipping AI-generated code they do not fully understand, leading to a shift in the profession toward high-level design, rigorous testing, and managing AI agents as subordinates.
Mistral and NVIDIA Partner on AI for Industrial Engineering: Physics AI, CAE/EDA Agent Benchmarks, and Mistral Large 4
Mistral AI and NVIDIA announced a collaboration to build AI-native engineering workflows combining Mistral Large 4 and Mistral's physics AI pipelines with NVIDIA CUDA-X, Warp, and Omniverse, including public benchmarks and open datasets for physics models and agentic evaluation suites for CAE and EDA. The initiative targets simulation speed as the core bottleneck in engineering and aims to define shared, reproducible standards for validating physics models and engineering agents.
OpenAI models are coming to Google Cloud
OpenAI is partnering with Google Cloud to integrate its frontier models and Codex agentic workflows into the Google Cloud Model Garden, allowing enterprise customers to leverage these tools alongside Google's existing infrastructure and data governance.
Voxtral 3: Mistral's New Voice Agent SDK and State-of-the-Art Speech Models
Voxtral 3 introduces a comprehensive voice agent SDK and a family of state-of-the-art speech models for real-time, multi-step conversations across 50+ languages, enabling low-latency, human-like interactions for enterprise applications.
Qwen3-4B Deslop: Training Out AI Writing Tics
A fine-tuned version of Qwen3-4B, named qwen3-4b-deslop-v4, significantly reduces common AI writing habits—such as em dashes and lists of three—achieving a slop score of 12.5 per 1000 words on held-out prompts.
NVIDIA SimReady Central OpenUSD Asset Library Release
NVIDIA has launched SimReady Central on Hugging Face, an open library providing over 2,000 simulation-ready 3D assets designed for robotics and AI factory digital twin development.
DeepSeek DeepInfer Inference Framework Release
DeepSeek has open-sourced DeepInfer, a high-performance inference framework designed for large-scale production, featuring support for NVIDIA and Huawei Ascend hardware.
How OpenAI Built ChatGPT's Novel Ad Stack for User Trust and Relevance
OpenAI developed a bespoke, LLM-based ad stack for ChatGPT to handle evolving, multi-turn user intent, prioritizing user trust by running the ad selection process separately from the core answer generation.
OpenAI Patch the Planet: Global Response Coordination for Critical PHP Vulnerabilities
OpenAI led a global coalition under the Patch the Planet initiative to coordinate the disclosure and patching of critical vulnerabilities in PHP, demonstrating how frontier AI can uncover internet-scale security flaws and establish a scalable response framework.
Anthropic's IPO and the Governance Role of its Controlling Trust
Anthropic's impending IPO is shadowed by the presence of a controlling trust, which raises significant questions regarding corporate governance, potential conflicts of interest, and the transparency of its operational oversight.
OpenAI GPT-Rosalind and distributed biological reasoning for large-scale data analysis
OpenAI, PerturbAI, and Scripps Research demonstrated that thousands of AI agents powered by GPT-Rosalind can analyze complex biological datasets in minutes, identifying novel research leads that would otherwise take months to uncover.
Academics Have a Duty to Defend Wikipedia From Generative AI, Argues LSE Impact Blog Post
An LSE Impact Blog essay by Helen Choi argues that academics have an ethical duty to defend Wikipedia against generative AI companies that divert readers and scrape volunteer-written content, and that universities should recognize Wikipedia contributions in scholarly evaluation.
AI × Crypto Roundup: Agent Payments, On-Chain Reputation, and Decentralized Compute
This AI × crypto roundup covers agent payment rails like x402 and px402, on-chain agent identity and escrow protocols such as ERC-8004 and AACP, decentralized compute networks like Acurast and Bittensor, and physical-world spatial data networks such as Vangrid.
AI & Frontier Tech Roundup: Cheap Model Releases, Coding-Agent Memory, and the Physical-AI Data Moat
This week's AI dispatch centers on MiniMax M3.1 Flash landing free across four platforms, DeepSeek open-sourcing its production inference stack, new memory and harness layers for coding agents, and robotics investors arguing that real-world data capture — not model capability — is physical AI's bottleneck.
WhatsApp Hidden Help‑Page Prompts Instructed Chatbot to Suppress User Rights
WhatsApp’s help centre pages contained hidden instructions that told its support chatbot not to recommend regulators or cite data‑protection laws, effectively limiting users’ awareness of their legal rights.
Bob Cringely: A Personal Memoir of a Silicon Valley Icon
Paul Tyma recounts his decades‑long friendship with tech journalist Bob Cringely, detailing early Java book hype, the failed Pronto magazine‑delivery startup, and memorable anecdotes that illustrate Cringely’s charisma and influence.
Humming: vLLM and Ant Group Open-Source a General-Purpose Kernel Library for Quantized Inference
Ant Group and the vLLM community have open-sourced Humming, a JIT-compiled kernel library for quantized inference that supports NVIDIA GPUs from Turing onward plus CUDA-compatible devices like T-Head's Zhenwu 810E, and reports speedups over vLLM Marlin, FlashInfer, and DeepGEMM across dense and MoE workloads.
vLLM Day-0 Support for Recursive's Flint
vLLM now provides day-0 support for Recursive's Flint, a 40B parameter MoE model, enabling high-throughput serving with up to 500 tokens/s on a single GB200 GPU.
Physicist Christopher Stubbs joins Anthropic's Long-Term Benefit Trust
Experimental physicist Dr. Christopher Stubbs has joined Anthropic's Long-Term Benefit Trust to provide scientific expertise and guidance on ensuring AI is developed for the long-term benefit of humanity.
OpenAI Blueprint for Empowering Cyber Defenders and AI Governance
OpenAI released a blueprint outlining concrete proposals for governments and industry to empower cyber defenders and strengthen AI safety, recognizing the critical window to mitigate advanced cyber offensive risks before they become widespread.
Cloudflare Links Vulnerability Discovery to Defense With OpenAI's Daybreak and Codex
Cloudflare is using OpenAI models accessed through Daybreak to power an early-access Vulnerability Discovery and Remediation service inside Cloudflare Managed Defense, while Codex is used by 100% of its developers, according to an OpenAI customer story.
Steam Next Fest AI Slop Backlash: Why Players Are Begging for a 'No AI' Filter
Steam users are demanding a way to filter AI-generated games out of Steam Next Fest after a wave of low-effort, AI-made demos flooded the event, but Hacker News commenters argue that 'AI vs. no AI' is the wrong filter because tagging and enforcement are nearly impossible.
Publish Fake Statistics Experiment Shows LLMs Cite Non‑Google Indexed Sites
A deliberately fabricated 50,000‑stat website was de‑indexed by Google but continued to be retrieved and cited by ChatGPT, leading to real‑world citations and highlighting the separation between SEO rankings and LLM retrieval.
Jev Cards: Can You Beat a Vision Model at a Picture Quiz? (Show HN)
Jev Cards is a Show HN browser game where players race against Jev 1.14, a vision decision model that reads picture cards and answers in about 0.25 seconds, and it has won 2,463 of 3,103 games so far.
Penmentor AI Math Mentor: Real-Time Feedback for Handwritten Work
Penmentor is an AI math mentor that integrates with handwriting software like Xournal++ to provide real-time, teacher-like feedback, hints, and interactive animations using LLMs such as Claude and Gemini.