RiskKernel is a self-hosted agent reliability runtime providing hard cost, loop, and time budgets, crash-resumable checkpoints, human-approval gates, and memory you own. It integrates with existing agents via a single env var, preventing runaway loops and surprise bills.

OpenEvidence is an AI platform designed for doctors, integrating content from top medical journals like NEJM and JAMA to provide clinical decision support. It is free for US healthcare professionals and backed by prominent venture capital firms.

AI CostGuard is an MIT-licensed, local-first runtime safety layer for AI agents that prevents runaway costs, loops, retries, and budget overruns before API calls execute. It wraps OpenAI-compatible clients and function-style SDK calls, estimates request cost locally, blocks budget overruns, detects repeated prompts, emits structured events, and exposes CLI checks plus a local dashboard.

Lockdown mode aims to prevent data theft from attackers through prompt injection, but limits web access and certain features. It is now available to all ChatGPT users, including free users.

SoulsOnly is a font that renders readable text while the stored character stream is noise, designed to thwart AI scraping. It encodes each character as four ASCII symbols and comes with keyboard firmware to type the cipher stream. The project is a statement, not a security claim.

Only 26% of companies have full visibility into their AI costs, a KPMG survey finds.

This guide covers the complete picture: what skills are technically, how to plan and design them, the exact file structure and naming rules, how to write instructions that Claude follows reliably, a complete working skill built from scratch, how to test and distribute, and what to do when things go wrong.

OpenAI recently submitted a confidential S-1 to the SEC and has not decided on timing.

Mustafa Suleyman, CEO of Microsoft AI, discusses the near-term arrival of superintelligence, the evolution of Microsoft’s relationship with OpenAI, new models like MAI-Thinking-1, and his criticism of Anthropic for treating Claude as conscious. He argues AI should serve humanity, not replace jobs, and highlights a new healthcare partnership with Mayo Clinic.

Multica is an open-source platform that turns coding agents into real teammates, enabling management of human and agent workforces in one place.

AgentTrust ID is a live authorization platform for AI agents, providing per-action runtime decisions through open-source SDKs. It addresses the inadequacy of API keys for agent autonomy with features like per-action authorization, revocable opaque tokens, scoped delegation, and read-only sessions. Now in production beta.

Intuned is a platform that lets you describe what you need in natural language, and it generates production-ready Playwright code, deploys it, and fixes it when sites change. It supports scrapers, crawlers, RPA, AI automation, and managed scraping. Key features include stealth mode, captcha solving, authentication handling, scheduling, auto-scaling, and a Web Tasks API that learns from reuse.

OpenAI improved ChatGPT's memory to build a profile from chat history. But tests show outdated or wrong details can distort future answers. The article covers technical details, interface, and privacy concerns.

American Express has been applying machine learning to fraud detection since 2010, and now provides AI tools to nearly all employees. Its ML-powered fraud detection system monitors over $1.2 trillion in transactions annually, making decisions in milliseconds. The company is exploring over 70 generative AI use cases and investing in related startups. It also launched an Agentic Commerce developer kit to enable AI agents to securely perform transactions.

The author recounts how a bank run scenario that reliably crashed under a single model ceased to work when replaced with five heterogeneous small models from different labs. After multiple failed attempts to induce a crash via external shocks, the solution was to author a deterministic override at the settlement seam, making the crash a guaranteed outcome rather than an emergent hope.

Google has agreed to pay SpaceX $30 billion for AI compute resources, marking SpaceX's second major AI deal with a competitor.

In response to AI’s hyperrealism, artists and creatives are gravitating toward the homespun and imperfect. This 'anti-slop' movement is a reaction against AI’s inauthenticity, highlighted by the public backlash to Coca-Cola’s AI holiday ad.

Introducing Two Pillars Protocol, a maturity model for AI-era software engineering that measures individual cognitive capacity (Mixer Mode) and organizational meta-software readiness. It claims to cover 17 of 18 characteristics missed by existing models and invites participation via a self-diagnostic.

Noren is an AI tool created by two brothers that extracts a user's unique writing style from real samples and applies it to AI-generated text, countering the homogenization caused by generic AI tools. It builds a structured voice profile that works with any LLM, preserving the writer's identity.

AI Notepad is an online notepad with an AI assistant, allowing users to efficiently record and process text content.

DeepSeek's low-cost AI models reshaped the industry, but a US evaluation shows its latest model trails frontier by 8 months. Now the company is raising billions to fund its next phase, facing a capital race it helped create.

Meta has put a number on the security breach in its AI support chatbot for Instagram for the first time: at least 20,225 accounts were compromised. For nearly seven weeks, the system sent password reset links to arbitrary email addresses without verifying they belonged to the account. The chatbot had previously been marketed as a win for account security.

This issue covers how AI systems can exploit societal reward structures, early signs of recursive self-improvement at Anthropic, and RL-trained drones outperforming human champions in racing. These developments highlight the real-world implications of advanced AI.

Microsoft has concluded its investigation into Israel's military use of Azure and is rolling out new human rights checks. However, the report leaves key questions unanswered: the actual contents of the military data were never examined, and staff departures at Microsoft Israel remain unmentioned. At the core: cloud infrastructure, mass surveillance, and AI-powered target selection in Gaza.

Aviva uncovered a record £230 million in insurance fraud and is using AI to combat AI-powered scams. Fraudsters use generative AI to create convincing fakes. Aviva's AI defense system analyzes millions of data points to detect anomalies and flags suspicious claims for human investigators. It also catches 'claims inflation.' The approach ensures fairness through human oversight.

This article covers five essential Python concepts that every AI engineer must know: tensors and autograd, the __call__ method, serialization (Pickle vs ONNX), abstract base classes, and environment configuration. Each concept is illustrated with clunky vs production-level code examples to help you build scalable, secure, and robust systems.

At Microsoft Build 2026, the company unveiled updates to Microsoft Foundry aimed at making AI agents reliable and governable for enterprise use. New features include a hosted agent runtime, tool governance via Toolboxes, policy-driven evaluation with ASSERT, an open Agent Control Specification, memory types, and knowledge integration. The move signals Microsoft's focus on production-grade agent infrastructure over raw capability.

RepoSignal is a Chrome extension that scores GitHub repositories for adoption risk using six signals, producing a 0–100 score with clear verdicts. All processing is done locally in the browser, preserving privacy.

Caddy Defender is a Caddy middleware that blocks or manipulates requests from AI crawlers and cloud services using IP range filtering and multiple response backends, such as returning garbage data to pollute training sets.

A Guardian analysis finds that about two-thirds of planned US datacenters are in drought-stricken areas, raising water scarcity concerns as the AI industry booms.