THE AI UPSKILL
A weekly digest for Beginner Builders
⚡ TLDR
Jensen Huang spent two hours at Nvidia's GTC conference this week telling the world that OpenClaw is the operating system of the AI agent era. We agree, and this week we're showing you exactly how to set it up, step by step, no coding required.
Meanwhile, 45,000 tech jobs have been cut in 2026 so far, with Block, Atlassian, and Oracle all pointing at AI as the reason. Morgan Stanley says a massive AI intelligence leap is coming in the first half of 2026, and most of the world isn't ready for it.
ARTIFICIAL INTELLIGENCE
THIS WEEK WE'RE WATCHING
→ freeCodeCamp's full OpenClaw tutorial for beginners: If you're going to set up your first AI agent, this is the best free resource right now. Covers installation on any platform, connecting AI models, messaging integrations, memory, and security sandboxing. Thorough without being overwhelming. Watch here.
→ Nate Herk: How I'd Learn n8n If I Had to Start Over in 2026: Nate has been our go-to n8n educator. This video is his revised learning roadmap, and his biggest takeaway is counterintuitive: don't start with AI agents. Start with the fundamentals, get your automations stable, then layer in AI. Watch here.
UPSKILL FURTHER
WANT TO GO DEEPER ON ANY OF THESE?
Join the AI-Enabled Marketer Skool community, which takes you from absolute beginner to AI-Enabled Marketer with Myosin.xyz, a community of top marketing professionals. - https://www.skool.com/the-ai-enabled-marketer/classroom
THE BUILDERS TIP
SET UP YOUR FIRST AI AGENT WITH OPENCLAW
Every build we've done so far has lived inside n8n. This week we're doing something different. We're setting up OpenClaw — the AI agent that Jensen Huang just called "the most popular open-source project in the history of humanity."
OpenClaw is a personal AI agent that lives in your messaging apps. WhatsApp, Telegram, Discord, Slack — wherever you already talk to people. You text it like a friend, and it handles tasks for you. It can manage your calendar, read and edit documents, browse the web, send emails, run terminal commands, and spawn sub-agents to handle multi-step projects. Under the hood, it connects to Claude, GPT, or local models through an open-source framework created by Austrian developer Peter Steinberger.
This is not another chatbot. You don't go to a website and type a prompt. Your agent runs 24/7, remembers your conversations, learns your preferences, and gets better at anticipating what you need over time. And as of this week, Nvidia has built an enterprise-grade reference stack around it called NemoClaw, with major companies like Perplexity, Mistral, and Cohere joining the ecosystem.
Here's how to get yours running today.
What you need:
An API key from Anthropic (Claude) at https://platform.claude.com/docs/en/api/admin/api_keys/retrieve or OpenAI https://platform.openai.com/api-keys
A computer — Mac, Windows, or Linux all work. A messaging app you already use (Telegram is the easiest to start with). About 30 minutes.
Quick start:
Open your terminal and run: curl -fsSL https://docs.openclaw.ai/install.sh | bash
The installer walks you through setup. Choose your AI provider (Anthropic recommended — Claude is the strongest model for agent work right now) and enter your API key.
Connect a messaging app. Telegram is the simplest: create a bot through @BotFather (1 minute tutorial https://www.youtube.com/shorts/lHe_DYTF1jg ), copy the token, paste it when prompted.
Send your first message. Try: "What can you do?" or "Summarize my last 5 emails."
That's it. Your agent is running.
Important: do this before anything else.
Run a security audit. In your terminal, type clawdbot security audit --deep and follow the instructions. By default, OpenClaw binds to your entire local network, meaning any device on your WiFi could reach it. Lock it down to loopback only. The audit walks you through this. Don't skip it.
Also: never share your agent with anyone. Don't add it to group chats. Keep write access limited to specific files and folders, not your entire system.
Something not working? Copy the error message and paste it into Claude or ChatGPT for troubleshooting. Most issues come from API key setup or messaging app permissions.
Make it yours:
Give it memory. OpenClaw stores everything in plain Markdown files on your disk. Open the MEMORY.md file and add facts about yourself: your name, timezone, projects, how you like communication. The more context it has, the better it works. Same principle as the Claude setup we covered in Issue 5 — but this time, the context lives with your agent permanently.
Add skills. Skills are reusable instruction sets that turn your agent into a specialist. Start with one practical skill — calendar management or email triage — and test it before adding more. The OpenClaw marketplace has dozens. Install one at a time. (See our latest video on Claude Skills https://www.skool.com/the-ai-enabled-marketer/classroom/39ff9eab?md=f4b1b4fa99e849bfa4be37cf9acd1dfd )
Automate a daily briefing. Text your agent: "Every morning at 8am, check Hacker News top stories, summarize the top 5, and send me a digest via Telegram." It will set this up and run it on schedule. Sound familiar? We built this exact workflow in n8n back in Issue 5. The difference: this time you set it up with one sentence.
Watch your costs. OpenClaw uses more tokens than a regular chat session because each task triggers multiple API calls behind the scenes. A long session can burn through tokens fast. Start a new session regularly rather than letting context pile up. Claude Sonnet is the best balance of cost and quality for daily use.
The bigger picture with OpenClaw:
Peter Steinberger, the creator, joined OpenAI in February. Sam Altman said OpenClaw will continue as an open-source foundation project. At GTC this week, Jensen Huang said every company needs an "OpenClaw strategy" the same way every company once needed an internet strategy. Nvidia built NemoClaw specifically to make OpenClaw safe for enterprise use — with sandboxed runtimes, privacy routing, and network guardrails.
Whether OpenClaw itself becomes the long-term standard or something else replaces it, the pattern it represents — a personal AI agent that runs continuously, connects to your tools, and acts on your behalf — is the direction everything is moving. Getting hands-on with it now puts you ahead.
To access additional setup guides and community support, head over to the AI Enabled Marketer Skool community through the link here — remember your first 7 days are free!
🌍 THE BIGGER PICTURE
Morgan Stanley: The AI Breakthrough Most People Aren't Ready For
A massive AI breakthrough is coming in the first half of 2026, which Morgan Stanley says most of the world isn't ready for it.
In a sweeping new report, the investment bank warns that a transformative leap is imminent, driven by an unprecedented build-up of computing power at America's top AI labs. The logic is striking: apply ten times the compute to AI training, and you effectively double its intelligence. Researchers say the data backing that claim is holding firm.
Executives at major US AI labs are already telling investors to brace for progress that will "shock" them. It is not hype. OpenAI's latest model recently scored 83% on a benchmark measuring performance on real-world economic tasks. That is at or above human expert level.
Morgan Stanley says the curve only gets steeper from here. Understanding AI now — even at a basic level — is the difference between being ready and being caught off guard.
45,000 tech jobs cut in Q1. AI is the stated reason.
Block cut 4,000 jobs in February, nearly 40% of its workforce. Jack Dorsey said AI can now automate much of that work. Atlassian cut 1,600 roles on March 11. Oracle is reportedly planning 20,000–30,000 cuts to fund AI data centers. Total across the industry: over 45,000 tech jobs gone in Q1 2026.
Not everyone buys the framing cleanly. Many of these companies over-hired during 2020–2022 and are correcting. Atlassian's CEO promised more engineering hires as recently as October. Block's headcount had tripled in three years. AI softens what might otherwise look like poor planning.
But the trend line is real. LinkedIn data shows AI job postings up 340% since 2024, while traditional software engineering roles have dropped 15%. The jobs being created require different skills than the ones disappearing, and the reskilling timeline doesn't match the displacement timeline. That gap is the thing to watch.
That's it for this week.
Building something cool with AI or have thoughts on how we can make this newsletter the best it can be? Reply to this email. 💬
P.S. If this was useful, forward it to someone upskilling in AI.