Monday, September 28, 2026

Good morning 👋

Tech didn't exactly slow down while we took a little break.

This morning, we're looking at a bigger shift happening right in front of us:

AI isn't just answering questions anymore. It's starting to do things.

And that changes the game.

Microsoft is pushing Copilot toward something much more ambitious: an AI workspace that combines chat, coding and longer-running agents capable of carrying out tasks.

Microsoft says its new Copilot brings Home, Code and Autopilot together, with the company describing Copilot as an emerging operating system for work.

Meanwhile, Meta's Muse is already being designed to handle real-world tasks such as emailing, shopping and booking travel. Reuters reports that Meta is also testing a feature where Muse can make phone calls on a user's behalf.

This is the important bit:

The next generation of AI may not need you to tell it every step.

Instead of:

the experience could become:

That could save people enormous amounts of time.

But it also means we're going to need to think much harder about what we're allowing AI to do on our behalf.

Apple is pushing its high-end Mac Mini and Mac Studio computers as machines capable of handling demanding AI workloads locally.

Reuters reports that the new systems can cost close to $20,000 and are aimed partly at businesses and AI developers who want to run advanced workloads without continually paying cloud providers for AI usage. Apple demonstrated a trillion-parameter model running across four Mac Studios.

For years, the big AI story was:

“Send your data to the cloud.”

Apple is betting there will also be a growing market for:

“Run powerful AI right here.”

That matters because local AI can mean different economics, more control over data and less dependence on cloud usage fees.

The battle isn't just about who builds the smartest AI anymore.

It's also about where that AI actually runs.

AI agents are increasingly being built to shop and make purchases for people.

But banks are now warning that the technology could create new problems involving fraud, scams and financial-data privacy.

Reuters reports that banks including NatWest, Bank of America, ING and others have called for stronger safeguards as AI shopping agents become more common. British retailer John Lewis said searches originating from AI agents had risen from 0.3% to 2.5% in a year.

This is where the futuristic AI story suddenly becomes very ordinary.

Imagine telling an AI:

“Find me the cheapest flight and book it.”

What happens if it chooses the wrong flight?

What happens if a scammer tricks the agent?

What happens if your financial information is exposed?

And perhaps the biggest question:

Who is responsible when the AI makes the decision?

The technology is moving quickly.

The rules around it are still catching up.

The big shift we're watching isn't simply better AI.

It's AI with permission to act.

Microsoft is building agents into work.

Meta is building agents into everyday life.

Apple is building powerful local computing hardware.

And financial institutions are already asking what happens when AI starts making purchases.

The internet may be moving from a place we click through to something that increasingly clicks for us.

AI is becoming less like a search box and more like an assistant.

The question isn't just:

“What can AI answer?”

It's becoming:

“What are you comfortable letting AI do?”

That is a very different question.

— OneClick Tech