This article discusses the structural shift in the data job market: SQL and Python have been demoted from differentiators to prerequisites. Analysis of over 700 data scientist postings from early 2026 shows machine learning and AI skills are now top demands. Four new differentiators are identified: data modeling, performance optimization, infrastructure awareness, and practical AI skills (LLMs, RAG, prompt engineering). The article provides actionable advice for acquiring each skill.

A Melbourne psychiatrist requires new patients to consent to AI scribe use or find another provider. AI note-taking tools are popular, with two in five GPs using them, per RACGP.

Databricks launches the Analytics Engineer Learning Pathway, a curriculum for SQL practitioners covering data modeling, ETL pipelines, semantic models, and conversational AI agents. The path includes six hands-on courses to build governed, AI-ready data foundations. Available now on Databricks Academy in self-paced and instructor-led formats, included with any active Databricks Learning Subscription.

The deal is one of the largest humanoid rollouts ever, according to the companies.

IBM Research launches the Open Agent Leaderboard, an open benchmark for comparing full agent systems (not just models). It evaluates generality across six diverse benchmarks, reporting both quality and cost. Early results show general agents are competitive with specialized ones, and agent architecture is increasingly impactful. All code, data, and paper are open-source.

Coding has emerged as the most widely used enterprise AI application. xAI's Grok Build aims to compete with Claude Code and Codex.

Ukrainian drone founder Yaroslav Azhnyuk went from pet cameras to AI-guided weapons. He and guest host Noah Smith make the case that the West is asleep at the wheel.

This issue of Import AI covers three important topics: the fast16.sys virus that selectively sabotages high-precision calculation software, reminiscent of the Sophon from The Three-Body Problem; the discovery that Muon optimizer can kill neurons and the introduction of Aurora optimizer; and a position paper on 'positive alignment' that addresses how AI can help humans flourish after safety is achieved. Additionally, LLMs are now capable of autonomously optimizing the training of other LLMs, though they struggle with creativity.

Brew applies Claude's design principles to email marketing for a streamlined experience.

Intel announces 3rd Gen Core processors for 'All-AI Thin-and-Light' laptops, featuring Intel 18A process, up to 40 TOPS AI performance, and system-level innovations via the 'Firefly' plan. Over 70 partner designs expected, targeting diverse user scenarios.

Pew research shows Americans are more worried than excited about AI as graduates voice fears over jobs. Former Google CEO Eric Schmidt was booed by students at a university commencement address in Arizona when he raised the topic of artificial intelligence.

The author shares four practical projects using local language models in daily workflow: private document brain, local code review, fully offline AI assistant, and a personalized thinking partner, highlighting privacy, offline capability, and customization advantages.

Linus Torvalds says AI-powered bug hunters have made Linux security mailing list ‘almost entirely unmanageable’ due to duplicate reports from multiple researchers using the same tools.

arXiv's computer science chair Thomas Dietterich announced a new policy: papers with clear signs of AI-generated content will lead to a one-year ban for all listed authors, and subsequent submissions must pass peer review. Mathematician Terence Tao supports the move, saying it aligns with the need to prioritize digestion over generation. Princeton University also ended its 133-year honor code due to widespread AI cheating, reinstating proctored exams.

Pope Leo XIV’s first encyclical Magnifica humanitas, on preserving the human person in the age of artificial intelligence, will be released on May 25, 2026. Signed on May 15, the 135th anniversary of Rerum novarum, the encyclical will be presented at the Vatican’s Synod Hall with the Pope and several speakers.

AI research in 2025 shifted from chatbots to reasoning systems, autonomous agents, and multimodal models. Key papers include DeepSeek-R1 (reinforcement learning), Gemini 2.5 (multimodal reasoning), Qwen2.5 (open models), Large Concept Models (concept-level language modeling), ESG analysis against greenwashing, VideoWorld (world models), AI Scientist-v2 (autonomous research), SWE-Lancer (coding agent benchmark), OLMo 2 (fully open language models), and Mixture-of-Recursions (efficient reasoning).

Anthropic and other AI providers are shifting from bundled enterprise pricing to pure consumption-based billing, driven by rising compute costs, hardware shortages, and capacity constraints. Organizations must adopt cost attribution, model routing, and budget management to balance experimentation with financial control.

As the US prepares for hurricane season and record-breaking summer heat, experts fear Trump administration cuts to climate and weather data programming could undermine federal weather forecasts when they're most needed. NOAA launched AI-powered models but needs ample data to train them.

Databricks and Veeva integrate to close the commercial intelligence gap in life sciences, embedding Genie AI agents and AI/BI dashboards directly into Veeva Vault CRM. This enables sales reps, MSLs, and territory managers to get real-time answers from their data without switching tools. A case study on undiagnosed ATTR-CM patients illustrates how field teams can prioritize target physicians with better data.

AI agent harnesses like OpenClaw are transforming how we build, train, and run AI models by orchestrating multi-step tasks. Small models with good harnesses can rival larger ones, causing a Mac Mini shortage. Training is shifting to reinforcement learning for tool calling, CPU demand is surging, and GPU-LPU combinations are becoming key. Future may offload work to client devices.

Anthropic will brief the Financial Stability Board on the cybersecurity implications of its Claude Mythos AI model, which it has declined to release publicly due to fears of misuse by hackers.

ActingWeb Personal Memory is a cross-platform persistent memory layer for AI assistants like ChatGPT, Claude, and Gemini. It lets users systematically build and manage context shared across all AIs, with features like task builder, memory sharing, and full data control. Free trial available.

Now that real life has caught up with science fiction, the imminent danger isn’t malfunctioning cowboys, it’s the robots convincing us that we’re great and everything is totally fine.

Recent research reveals the real effectiveness of agent skills: curated skills improve task completion by 16.2% on average, while self-generated skills show no consistent benefit. As skill libraries grow, flat retrieval breaks down, making hierarchical organization critical. Additionally, more than one in four community skills contain exploitable vulnerabilities.

Former Google CEO Eric Schmidt was repeatedly booed during a commencement speech at the University of Arizona when discussing artificial intelligence. He contrasted AI with the rise of computers, acknowledged technology's downsides, and urged graduates to shape the future. The university defended the invitation, citing his leadership. A similar incident occurred earlier at the University of Central Florida.

This article explores the downsides of relying on AI for investment writing: it strips away the value of writing and reading—for authors, writing is a process of thinking, exploration, and growth; for readers, it builds trust and understanding of the author's perspective. AI-generated text lacks originality and personality, and cannot replace the deliberative human writing process.

The Huawei-backed openJiuwen community has launched and open-sourced JiuwenSwarm, a swarm intelligence platform that enables multiple AI agents to collaborate efficiently and autonomously evolve. It introduces the Coordination Engineering paradigm, marking a shift from single-agent to multi-agent systems.

This presentation highlights recent efforts at the Johns Hopkins Applied Physics Laboratory to advance agentic AI for collaborative robotic teams. It begins by framing the core challenges of enabling autonomy, coordination, and adaptability across heterogeneous systems, then introduces a scalable architecture designed to support agentic behaviors in multi-robot environments. The talk concludes with key challenges encountered and practical lessons learned from ongoing research and development.

Melbourne is leveraging sovereign AI compute, hyperscale data centers, and international conferences to create a research flywheel. MAVERIC supercomputer, investments by CDC and NEXTDC, and upcoming conferences position Melbourne as a global hub for AI research.

OpenAI and Dell partner to bring Codex to hybrid and on-premise environments, helping enterprises deploy AI coding agents securely across data and workflows.