At this point we've all gotten that email. The one where you open the attachment and within three seconds you know nobody actually looked at this before they hit send. The formatting is suspiciously clean. The content says everything and nothing at the same time. It's eight pages long and somehow the point doesn't show up until page six. That's what I want to dig into today.

I've also got three stories: Anthropic just gave Claude the ability to control your computer from your phone, the Microsoft 365 roadblock that was holding Claude back in workplaces just disappeared, and a $122 billion number that should make you pay attention.

Here's something I see too often. Someone gets tasked with putting together a report. They type a vague prompt into ChatGPT, copy what comes back, and fire it off. The email has emojis in it. The report looks impressive at first glance. Full sections, metrics, charts. But the actual point of the report doesn't show up until section 8, part C. The middle three pages go off on a tangent because the prompt wasn't specific enough. Nobody who receives it knows what they're supposed to do with it. But the person who sent it? They're proud of themselves. A deliverable that should have taken three hours, done in five minutes. That's the whole point of AI, right?

Nope. They're doing it wrong. And this is the problem with copy-paste AI.

First, it's making you worse at your job.

Not in a dramatic way. In a slow, barely noticeable way. You stop thinking about what you actually need before you ask. You stop evaluating whether what came back is good. You stop building the muscle of knowing what a good output looks like. Copy-paste AI is brainrot with extra steps. You're not using a tool. You're outsourcing your thinking instead of outsourcing your tasks (there is a difference).

Here's how I think about it. Imagine a school that lets students use AI for every essay. Full access, no restrictions. But the exam is different. The teacher generates a personalized test for each student based on the topic they chose to write about. And that exam is taken without AI.

If you actually used AI well during the essay, you learned the subject. You engaged with the material. The AI helped you think, not think for you. And you pass the exam. If you just copy-pasted your way through the essay, you learned nothing. And the exam exposes that immediately.

Work is the same thing. Except nobody's giving you the exam. You just slowly get worse and don't notice. And what's worse is that everyone is telling you that you should be getting better with AI.

Second, you're leaving the best part on the table.

There's a conversation happening right now about AI taking away software engineering jobs. And I think most people are getting the story wrong. AI isn't replacing good software engineers. It's replacing bad ones (or ones without experience). A lot of people look at vibe coding and think the gap between a top-tier dev and someone who's never written code just closed. And in some ways, it has. You can build an app really quickly now, and that's cool.

But here's what people miss. AI also works as a multiplier of what you already know. A great engineer who deeply understands architecture, debugging, and systems thinking gets 10x more out of AI than someone who just types "build me an app." The expert gets a force multiplier. The beginner gets a shortcut that eventually hits a wall.

This applies to every job, not just engineering.

If you show up to AI with "make me a presentation that will make the client sign with us," you'll get something generic and forgettable. But if you show up and say "I know that X, Y, and Z work well in a client deck, so create those sections, and then help me build a tracker for metrics A, B, and C so I can improve the deck over time," now you're getting somewhere. That second person isn't just using AI. They're compounding their own expertise with it.

You don't need to understand how AI works under the hood. You don't need to know what a transformer is. But you absolutely need to understand your own workflow. You need to know your outputs. You need to come to the conversation with intent. Because the people who bring domain knowledge to AI will run laps around the people who show up empty-handed.

That's the difference between AI making you better at your job and AI slowly replacing the parts of your brain that made you good at it in the first place.

If you're thinking about rolling out AI across a team and you want to make sure people actually get better instead of lazier, that's exactly what we help with at Bridge AI. It starts with understanding workflows, not installing software.

Anthropic shipped computer use for Claude Cowork and Claude Code. Claude can now point, click, scroll, and navigate your screen to complete tasks when there's no direct integration available. Pair it with Dispatch, which lets you message Claude from your phone and come back to finished work on your desktop, and you're looking at a real shift in how AI gets work done. You can literally assign a task on the train and review it when you sit down. This is the kind of thing people imagined AI would do three years ago. It's here now.

For a long time, the biggest thing holding Claude back in workplace settings was that it couldn't talk to Microsoft. Most businesses run on Outlook, Teams, SharePoint, and OneDrive. If your AI can't access any of that, you're stuck copying and pasting context into every conversation. That roadblock just vanished. Claude now has a native Microsoft 365 connector. It can search your emails, pull from SharePoint, read your calendar, and work with your files directly. If your company runs on Microsoft and you've been waiting for a reason to try Claude at work, this is it.

OpenAI closed the largest private funding round in history. $122 billion. $852 billion valuation. $2 billion in monthly revenue. 900 million weekly users. These are absurd numbers for a company that just killed its Sora video app after burning $1 million a day running it. The money is real but so is the chaos. Four executives stepped down this week ahead of a potential Q4 IPO. They're consolidating into a ChatGPT "super app." My take hasn't changed: Anthropic is ahead on product and OpenAI is catching up with capital. Capital is powerful. But product wins in the long run.

That's it for this week.

If you found this useful, forward it to someone on your team who's been copy-pasting ChatGPT responses for the last year and calling it "using AI." They need this more than you do.

And if you want to talk about what AI could actually do for your business, not in theory, but in your workflows, with your team, on your timeline, just reply to this email. That's what we're here for.