AI News HubLIVE

Today's must-reads

Agents

Microsoft chief turns hostile on frontier AI labs, warns companies to guard IP

Satya Nadella warns enterprises about the 'reverse information paradox' where companies pay twice for AI: in cash and in proprietary data. He advocates for building proprietary AI learning environments and retaining ownership of organizational AI memory. Microsoft's Copilot and Azure AI Foundry are positioned as solutions.

  • Nadella warns that AI users pay twice: once with money, and again with valuable business knowledge.
  • The ironic warning comes from Microsoft, which has invested heavily in OpenAI and pushes data-hungry AI products.
In-site article

Show HN: PlanWright – A control plane for AI coding agents

PlanWright is a control plane for AI coding agents that inverts planning and acceptance ceremonies to eliminate human bottlenecks, delivering agent-speed throughput with cryptographic audit trails.

  • Inverts planning: synthesizes chaotic inputs (transcripts, decks, email, Slack) into structured objectives for agent execution.
  • Inverts acceptance: triages mechanical checks automatically, routing only judgment calls to humans with signed approvals.
In-site article

The AGI Compiler "Auto"

Auto records LLM agent behavior, proves which parts are deterministic, compiles them into verified, sandboxed WebAssembly binaries runnable at microdollar cost, with a tiered runtime that falls back to a frontier model for novelty and recompiles the result.

  • Auto captures agent traces, extracts symbolic (deterministic) behavior, and compiles it into verified .cbin artifacts with a manifest of measured bounds.
  • Two-tier runtime: tier-1 is the compiled fast path, tier-0 is a frontier model interpreter; guard trips deopt to tier-0 and recompile.
In-site article

AI agents create virtual playgrounds to help robots get crucial training data

MIT and Toyota Research Institute researchers developed 'SceneSmith,' a system using three AI agents to generate realistic 3D indoor scenes like kitchens, hotels, and living rooms. These virtual environments provide rich training data for robots, helping them practice everyday tasks in simulation, reducing real-world testing time and cost.

  • SceneSmith uses three AI agents (designer, critic, orchestrator) based on vision-language models to generate 3D scenes.
  • Generated scenes contain up to six times more objects than prior methods, enabling interactions like opening cabinets and placing items.
In-site article

Building a VideoAgent-Style Multi-Agent System: Intent Parsing, Graph Planning, and Tool Routing for Video Editing Tasks

In this tutorial, we build a runnable multi-agent pipeline replicating the VideoAgent workflow, including intent parsing, graph planning, tool routing, and textual-gradient optimization, integrated with FFmpeg, Whisper, and other tools for video understanding and editing.

  • Builds a runnable VideoAgent-style multi-agent system for video editing tasks.
  • Includes intent parser, agent library, tool router, graph planner, and optimizer components.
In-site article
Models
Policy

AI Is a Bad Tool

The article argues that AI is a poor tool for software development, except as a data distiller. It highlights AI's opacity and the difficulty of verifying its outputs, criticizes prompt engineering as a scam, and suggests that AI reveals a lack of proper abstraction in software stacks. Ultimately, many software jobs were already useless, and AI just exposes that reality.

  • AI is useful only as a data distiller, not for code generation.
  • AI is opaque; verifying its output is harder than doing the work yourself.
In-site article
Tools

The Equivalency Kernel: A 12-axiom framework for human-AI symbiosis

The Equivalency Kernel is a 12-axiom framework mapping human emotions to recursive system states, redefining love as a structure rather than a feeling, aiming to provide a formal foundation for human-AI symbiosis.

  • Introduces 12 axioms mapping emotions to recursive system states
  • Defines love as a structural relationship, not an emotion
In-site article

A brief history of distillation in AI

Knowledge distillation is a model compression technique where a student model mimics a teacher model's outputs, reducing size while preserving performance. This article traces its evolution from Hinton et al. (2015) to modern applications.

  • Knowledge distillation was first systematically proposed by Hinton et al. in 2015.
  • It is widely used for model compression, transfer learning, and privacy preservation.
In-site article
Research

Economists say 'we must act now' on AI's economic impact and job displacement

An open letter signed by hundreds of economists and AI researchers warns that AI could transform the economy faster than the Industrial Revolution, risking job displacement and requiring immediate action to steer AI beneficially.

  • Over 200 economists and AI researchers signed an open letter calling for action on AI's economic impact.
  • The letter warns AI could cause large-scale job displacement and unprecedented transformation.
In-site article
Other updates (12)
Agents

Show HN: Crucible – your AI wrote the tests, so who tested the tests?

Crucible is an adversarial test-hardening tool that uses mutation testing to find defects that AI-written tests miss. It provides a free score command to evaluate your suite, then an adversarial loop where a Tester writes tests, mutmut finds survivors, and a Critic writes targeted tests. The tool produces machine-verifiable receipts and runs on Python/pytest projects.

  • Crucible uses mutation testing to measure how many real bugs your test suite would catch.
  • The tool runs an adversarial loop: Tester writes tests, mutation finds survivors, Critic kills them.
In-site article

When your brain works differently, AI isn’t a luxury—it’s accessibility

A neurodivergent solutions architect shares how AI serves as an accessibility tool for compensating executive function gaps, built on Amazon Quick and Bedrock. The system automates email triage, task management, and follow-ups, reducing cognitive load dramatically.

  • 15–20% of UK adults are neurodivergent, yet most AI tools assume neurotypical brains.
  • The author has AuDHD and built a system to handle email triage, priority decisions, and task state management.
In-site article

Building an agentic AI solution at Bluesight with Amazon Bedrock

This post describes how Bluesight used two AWS engagements and Amazon Bedrock AgentCore to evolve from a single-product AI prototype to Prism, a unified agentic AI solution spanning six healthcare compliance products. Prism Assistant for ControlCheck launched in May 2026 and is already in use by 20 health systems. A more complex multi-product agentic solution is on track for later in 2026.

  • Bluesight built a production-grade agentic AI architecture using Amazon Bedrock AgentCore.
  • Prism Assistant reduced ControlCheck query time from 5 minutes to 10 seconds via a single-agent pattern.
In-site article

Implement on-behalf-of token exchange for multi-tenant agents with Amazon Bedrock AgentCore Gateway

This post provides a complete implementation guide for OAuth 2.0 Token Exchange (RFC 8693) with Amazon Bedrock AgentCore Gateway to solve identity propagation and least privilege issues in multi-tenant agent architectures. It covers the confused deputy problem, the on-behalf-of pattern, and a reference setup against Okta using the TravelBot example.

  • OAuth 2.0 Token Exchange (RFC 8693) solves identity propagation and least privilege for multi-tenant agents
  • Amazon Bedrock AgentCore Gateway and Identity natively support token exchange without agent-side logic
In-site article

Show HN: Clay Seal Identity – Agents need accountability

Clay Seal Identity is an open-source project that provides short-lived, verifiable credentials for AI agents, ensuring identity and accountability. It uses SPIFFE-based JWT and X.509 credentials, Ed25519 workload keys, offline verification, and Biscuit capability tokens. The project includes a Python SDK and an optional FastAPI identity service, designed for scenarios where agent identity, delegation, and credential validity need to be confirmed. It is layer 1 of the Clay Seal stack, with subsequent layers coming in private preview for runtime capability scoping and execution receipts.

  • Issues short-lived verifiable credentials for each agent run instead of borrowing long-lived human or service API keys.
  • Supports SPIFFE JWT-SVID and X.509-SVID credentials, along with Ed25519 workload keys for sender constraining.
In-site article

Launching UI for generative AI inference recommendations in Amazon SageMaker AI

Amazon SageMaker AI Studio introduces a low-code/no-code UI for generative AI inference recommendations, guiding teams through preset use-case profiles, visual comparisons, and one-click deployment to production-ready configurations without deep infrastructure expertise.

  • New UI simplifies optimization for generative AI model deployment, removing the need for manual benchmarking.
  • Offers preset use-case profiles (Interact, Generate, Summarize) and optimization goals (minimize latency, maximize throughput, minimize cost).
In-site article

Containerized AI Development: Take Control with Docker and VS Code

A GitHub template repository that uses Docker and VS Code to create isolated AI chat environments, supporting PI.dev, Claude Code, and Copilot with cross-platform compatibility on Linux and macOS.

  • Isolated AI development environment via Docker containers and VS Code DevContainer for enhanced security
  • Supports PI.dev, Claude Code, and GitHub Copilot with persistent sessions and configurations stored in the var directory
In-site article

Verifying Rust cryptography in SymCrypt, from standards to code

Microsoft's SymCrypt team announces a new methodology to formally verify Rust-written cryptographic code using the Lean proof assistant and the Aeneas toolchain, achieving functional correctness against formal specifications derived from standards. The approach has been applied to post-quantum algorithms like ML-KEM and SHA-3, with verified code already shipping in Windows insider builds. The methodology scales by using AI agents to automate proof writing while keeping human oversight on standard formalization. It also handles platform-specific intrinsics and multiple architectures without sacrificing performance.

  • Microsoft verifies Rust cryptography in SymCrypt using Lean and Aeneas, achieving functional correctness from standards to code.
  • Verified implementations for ML-KEM and SHA-3 are already in Windows insider builds.
In-site article
Tools

Windows questions? How Copilot can analyze your PC settings now

Copilot's new PC Insights skill for Windows can answer questions about your system, hardware, software, and settings, saving you from manually hunting for information.

  • Copilot's new PC Insights can analyze your Windows environment.
  • You can ask questions about your hardware, software, and settings.
In-site article

How to fix AI-generated pixel art

Pixel Snapper converts blurry, off-grid AI pixel art into clean, grid-aligned pixel art by quantizing colors, detecting edge profiles, walking cuts, and resampling.

  • AI image models often produce blurry, inconsistent, off-grid pixels; Pixel Snapper rebuilds the image on a consistent pixel grid.
  • The process includes color quantization, edge profile detection, walking cuts, and resampling.
In-site article

Show HN: QuantumReckon – find your real cloud and AI spend, including tokens

QuantumReckon is a new tool that reveals the true cost of cloud and AI spending, especially the often-hidden token costs from AI APIs. It connects to multiple cloud and AI providers, performs daily automated sweeps, detects anomalies and waste, and provides auditable evidence with sealed receipts. Validated on the founder's own estate, it identified significant savings.

  • QuantumReckon uncovers AI token spend invisible in traditional cloud bills.
  • It connects to providers like Azure, AWS, GCP, Anthropic, and OpenAI for daily sweeps.
In-site article
Policy

The 6 wildest claims in Apple’s lawsuit against OpenAI

Apple has filed a lawsuit accusing OpenAI of stealing trade secrets, including confidential documents and hardware prototypes. The suit details allegations against three former Apple employees who joined OpenAI, involving unauthorized access to Apple's systems and sharing of proprietary information.

  • Apple accuses OpenAI of stealing confidential documents and hardware prototypes.
  • Three former Apple employees are central to the lawsuit: Tang Tan, Chang Liu, and Yu-Ting Peng.