From the Author’s Desk
We’re releasing this week’s volume a little early to align with International Women’s Day. Technology shapes the world we live in, but the people who build it shape technology. Moments like this are a reminder that more voices, perspectives, and builders should be part of that future.
Last week we talked about the moment when AI stops just answering questions and starts doing things on your behalf. This week, we’re opening that black box. How does AI actually get things done? The answer is simpler than most people think. The model reasons, but tools do the work.
“Technology changes the world. Builders decide how.”
How AI Gets Things Done
Most people imagine AI as a giant brain that already knows everything.
In reality, the model itself is limited. It can generate language, reason through problems, and make decisions, but it cannot directly interact with the world.
That’s where tools come in. A modern AI system can decide to call tools that perform specific actions. Examples include:
searching the internet
analyzing spreadsheets or documents
running code and calculations
accessing internal company databases
interacting with software through APIs
When AI uses tools, the workflow looks something like this:
Goal → AI plans steps → AI calls a tool → tool returns information → AI continues reasoning.
This is how AI moves from generating answers to completing tasks.
The Mindset Shift: AI isn’t always the worker.
In many systems, it’s the orchestrator.
It decides what information is needed, which tool should fetch it, and how the results should shape the next step. The model isn’t doing every task itself. It’s coordinating them.
Once you notice that pattern, modern AI systems start making a lot more sense.
⚡ The “Spot the Tool” Exercise
This week, try giving AI a task that requires outside information. For example:
“Find the latest news about Nvidia’s AI chips and summarize the key trend.”
“Analyze this spreadsheet and tell me what stands out.”
“Search the web and explain how companies are using AI agents.”
Now watch what happens. Notice when the system pauses to search, browse, or analyze a file before responding.
It means the model didn’t already know the answer. It decided to use a tool to find it.
Once you start noticing this pattern, you’ll realize modern AI isn’t just generating responses. It’s coordinating actions! 💡
💡 International Women’s Day: MEP Lina Gálvez on Online Violence & AI Deepfakes
Ahead of International Women’s Day, EU lawmaker MEP Lina Gálvez speaks about the growing risks of online violence against women and girls in the age of AI.
Recent research found that AI systems can generate millions of sexualized deepfake images with minimal effort, the vast majority targeting women. In this conversation, Gálvez discusses why stronger accountability for AI platforms is becoming urgent.
Worth a watch if you want to understand how AI safety debates are increasingly intersecting with women’s rights and digital protection.
💡 When AI Ethics Meets National Security
Anthropic CEO Dario Amodei pushed back on parts of a Pentagon contract, refusing to allow Claude to be used for mass surveillance or autonomous weapons. The move led to the company being flagged as a supply-chain risk.
In a conversation with The Economist, he explains why the debate isn’t just political. It’s about whether today’s AI is reliable enough for the battlefield and who should ultimately control it.
Worth a watch if you want a glimpse into the growing tension between AI capability and responsibility.
💡 Who’s Most Exposed to AI at Work?
Anthropic analyzed millions of Claude interactions to see where AI is actually being used in real jobs. Their research introduces “observed exposure,” a measure of where AI is beginning to automate tasks.
Early signals show that higher-paid, knowledge-heavy roles are most exposed, with workers in these professions more likely to be older, highly educated, and female.
There’s no clear rise in unemployment yet, but hiring for younger workers in exposed fields may already be slowing. Worth a read if you’re curious how AI adoption is quietly reshaping the labor market.
She Builds: 24 Hours to Build an AI App
To celebrate International Women’s Day, Lovable is opening its platform for a 24-hour global build event. On March 8, anyone can log in and build AI products for free using Lovable, powered by Anthropic.
Participants also receive $100 in Anthropic API credits and $250 in Stripe fee credits, making it a great moment to experiment with your first AI-powered idea.
No applications. No prerequisites. Just show up, build something, and share what you create.
Perfect if you’ve been thinking about moving from using AI to actually building with it!
This week, notice when AI switches from answering to actually using a tool to complete a task.
Next week, we’ll explore what happens when multiple AI agents start working together.
And as we mark International Women’s Day, here’s to more women stepping into AI and technology, not just as users, but as builders. 🙌
Link to ➡️ Previous Volume
💛 If this helped, feel free to share it with someone learning AI. 💛
