Hi tech & AI enthusiasts, this week's highlights:
The Pope's First Encyclical Is About AI
Claude Opus 4.8 Lands With Better Judgement and an Effort Control
xAI Just Launched a Coding Agent in Your Terminal
The Lab: How to Use ChatGPT in PowerPoint
AI Image of the Week: Construction mini
Let’s get into it.
▲ AI MODELS
Pope Leo XIV released Magnifica Humanitas, his first encyclical as Pope. The 42,300-word document is his most direct statement on the promise and dangers of AI, and he chose to present it himself rather than delegate it to cardinals.
What you need to know
Technology is never neutral: Leo's underlying argument is that AI is not inherently evil, but it takes on the characteristics of whoever designs, funds, regulates, and uses it. Responsibility sits with people, not the tools.
Strong language on AI in warfare: He directly condemns autonomous weapons, arguing that reducing human control over weaponry makes wars easier to start. The encyclical calls the "just war" theory outdated in the age of AI.
Themes of work, truth, and dignity: Beyond warfare, the document covers AI's impact on jobs, social justice, the spread of misinformation, and the dignity of human labour. Signed on the 135th anniversary of Rerum Novarum, the 1891 encyclical on workers' rights.
Key Ideas
▲ Why it matters: When the head of an institution with 1.4 billion members makes their first major statement about AI, it changes how the conversation is held in boardrooms, families, and governments. For business leaders, the practical signal is that AI ethics is moving from a tech-industry conversation into mainstream public debate, and customer and employee expectations will follow.
▲ AI MODELS
Anthropic released Claude Opus 4.8, a meaningful upgrade to its flagship model. Available everywhere today at the same price as Opus 4.7, with a new effort control that lets you trade speed for depth on every prompt.
What you need to know
Sharper judgement on agentic work: Cursor reports tool calling is "meaningfully more efficient." Devin calls it the strongest agentic coding model they've tested. Opus 4.8 hits 84% on the Online-Mind2Web computer-use benchmark, beating both Opus 4.7 and GPT-5.5.
Significantly more honest: Opus 4.8 is around 4x less likely than Opus 4.7 to let flaws in its own code pass unremarked. Early testers consistently report it flags uncertainty and pushes back when a plan isn't sound.
New effort control and dynamic workflows: Pick how hard Claude thinks on any prompt (high, extra, or max). Fast mode is now 3x cheaper than before. Claude Code can now run hundreds of parallel sub-agents in a single session to handle codebase-scale migrations.
Key Idea
▲ Why it matters: For most business teams using Claude for analysis, writing, and decisions, the headline is honesty. Opus 4.8 is less likely to confidently invent an answer when it doesn't know, which is the single biggest blocker to trusting AI on real work. Same price, better model. If you pay for Claude, you got a free upgrade today.
▲ AI TOOLS
xAI Just Launched a Coding Agent in Your Terminal
Elon Musk's xAI launched Grok Build, a coding agent that runs from the command line. It is a direct shot at Claude Code, Codex CLI, and Gemini CLI, with the addition of "Plan Mode" and parallel sub-agents.
What you need to know
Plan Mode before execution: You see a step-by-step plan before any code is written. You can approve it, comment on individual steps, or rewrite it entirely. Every change shows up as a clean diff.
Parallel sub-agents: For larger tasks, Grok Build delegates work to specialised sub-agents that run in parallel, each in its own worktree. Useful for refactors and large migrations.
Powered by a new model: Runs on grok-code-fast-1 and the new grok-build-0.1 model, with a 256,000-token context window. Available for SuperGrok and X Premium+ subscribers.
Key Idea
▲ Why it matters: xAI is officially in the agentic coding race. With Anthropic, OpenAI, Google, and now xAI all offering CLI-based coding agents, the question for engineering teams is no longer "should we use one" but "which one fits how we work." The Plan Mode angle is a smart competitive choice, since over-eager agents writing the wrong code is the most common complaint about Claude Code and Codex.
▲ EVEN MORE NEWS
Other big ideas from the past week.
▲ ChatGPT Now Builds and Edits PowerPoint Slides for You: OpenAI launched ChatGPT for PowerPoint, an add-in that lives inside the deck. Build slides from a brief, rewrite copy, pull context from Gmail and SharePoint, and audit the narrative for weak spots. Tutorial below.
▲ OpenAI's Model Just Solved an 80-Year-Old Maths Problem: An internal OpenAI model disproved a central conjecture in discrete geometry, the "unit distance problem" first posed by Erdős. External mathematicians have verified the proof. One commenter: "AI has gone beyond being just an assistant."
▲ Google and Synaptics Ship the Coralboard for Edge AI: Google Research and Synaptics launched the Coralboard, a development board that runs AI models locally on a single chip. Built for builders who want AI in cameras, kiosks, and connected devices without sending data to the cloud.
The Lab
▲ AI EDUCATION
ChatGPT now works as an add-in inside Microsoft PowerPoint. You describe what you need and it builds the deck, rewrites the copy, or audits the narrative directly inside your file.
What you'll learn:
How to install the ChatGPT for PowerPoint add-in
How to build a full deck from a single brief
How to edit existing slides with plain-language prompts
How to audit a finished deck for narrative weak points before you present
🔬 Interested in learning about AI for business? Join The Lab waitlist.
▲ AI IMAGE SHOWCASE
Create this type of image with Midjourney
Professional stock colored photography image of a construction crane standing on the building plans, next to it is a concrete frame of a building, the background is light and monochromatic, very realistic 16k photo 3d isometric.Image Usage Suggestion: Construction firm website hero images, project pitch decks, investor presentations for development companies, architecture and engineering brochures, LinkedIn carousel covers, or any content that needs a confident, well-composed visual to anchor a business case.
💌 Reply to this email with your AI image generation or suggestion of what you’d like to see and I'll feature it in a future newsletter.
▲ SHARE YOUR THOUGHTS
Help shape the content you see here by giving feedback
Have specific feedback or want to get in touch? Reply to this email and I’ll get back to you.
Know someone who’d love this newsletter? Forward it to a friend and have them sign up here.
Thanks for reading, until next time.
Stay curious,
Matt Lok, Editor
Brought to you by Metalabs. Digital marketing consulting specialising in AI.