Alright, let’s talk AI.

Every week I try to bring you the stuff worth knowing… the ideas, the rants, and the tools that might just change how you work. Some of it’s opinionated (okay, most of it), some of it’s practical, and all of it’s designed to make you think a little differently about what’s coming next in GTM.

Today we’ve got a spicy take on meeting etiquette, a peek into the operating model the next Fortune 500 will be built on, and a tip that’ll make your GPT-5 prompts hit harder.

Let’s get into it.

If you’re leading, up. Everyone else, down. Sorry, it just has to be that way.

▸ Don’t tell me you “need to take notes.” You don’t. Have an AI assistant do it, or assign one person if your company is too cheap.
▸ Don’t tell me you “can’t see the presenter’s screen.” Buy a TV, Scrooge.
▸ Don’t tell me you “need to look something up.” You don’t. That’s just Googling your way out of paying attention.

Laptops in meetings aren’t tools, they’re the grown-up version of hiding under the desk. You’re not contributing, you’re just taking up oxygen. Keep phoning it in and you’re not just killing the meeting, you’re killing the company. And we all know why, Slack is whispering, email is pinging, ESPN is calling, and you’re sitting there with your “I’m totally engaged face,” fooling exactly zero people.

If you’re a senior leader, it’s not just annoying, it’s super disrespectful. It’s you broadcasting, “This meeting is beneath me” and that's just not cool.

The rule is simple:

If you can’t give your attention, don’t give your presence.

Meetings don’t suck because of laptops. They suck because too many people show up and check out.

Still adding more people to scale? That’s cute.

The companies eating your lunch right now aren’t doing it with bigger payrolls. They’re doing it with bigger output,  from smaller teams.

The future of work isn’t about adding more humans. It’s about multiplying the humans you have. Not by tossing them another tool to click through. Not by automating a single workflow and calling it innovation.

This is a complete operating model shift.  It’s called the Autonomous Organization.

In an Autonomous Enterprise, humans don’t just use AI, they lead it. Every person on your team becomes the boss of a small army of Cloud Employees, AI teammates that work 24/7, execute entire outcomes, and keep getting better every time they do the job.

The result? Work happens faster. Scaling becomes a decision, not a headcount request. And the companies making the switch now will out-scale you without outspending you.

Here’s what it looks like inside early adopters:

1️⃣ The Ratio Flips
Most orgs today run 1:0. One human, no AI. In five years, it flips. One human will lead 6, 10, even 20 Cloud Employees. Fifty people with the output of 500. Early adopters are already there. They’ll outscale you without outspending you.

2️⃣ Cloud Employees Are Hired Like Humans
You don’t “subscribe” to them. You hire them. Train them. Assign work. Hold them accountable. They show up, take action, and deliver results. If your team thinks AI means “chatbot,” they’re behind.

3️⃣ Humans and Cloud Employees Work Side by Side
This isn’t a handoff, it’s collaboration. Cloud Employees aren’t background automations, they’re teammates. Reps run plays with them. Support escalates to them. Ops builds flows with them. Outcome by outcome, side by side.

4️⃣ Every Employee Becomes a Cloud Employee Boss
Top people aren’t replaced, they’re amplified. Your best AE now runs a pod of outbound agents. Your best marketer leads a campaign engine. AI multiplies the impact of your best, not the opposite.

5️⃣ Cloud Managers Enter the Org
The real unlock: AI that doesn’t just do the work, it manages it. Cloud Managers oversee pods, QA output, flag issues, and optimize flows. Not sci-fi. Already testing.

6️⃣ Leadership Shifts From Control to Orchestration
Control is dead. Flow is the new skill. Leaders design how work moves, who does what, and what gets measured. Managing people is outdated. Mobilizing performance is the job.

7️⃣ Capacity Becomes a Choice
Need more outbound? Spin up 3 agents. Need 24/7 support in 3 languages? Launch it tomorrow. Scale isn’t a headcount problem anymore. Your ceiling is optional.

8️⃣ Onboarding Goes Zero-to-Output
Cloud Employees go live in minutes. No paperwork. No ramp. For humans, ramp accelerates too, because AI handles the grunt work. Every new hire comes with a pit crew.

9️⃣ Org Charts Give Way to Capability Maps
Forget who reports to who. The real question: what needs to be done, and who—or what is best suited? Orgs become networks of capability. That’s how velocity happens.

🔟 Cloud Employees Become the Interface
The UI layer collapses. You don’t need another tool, you need outcomes. Cloud Employees log in, navigate, and execute. You don’t even see the dashboard anymore.

1️⃣1️⃣ Data Becomes the Operating System
Data isn’t analytics, it’s fuel. Cloud Employees run on it. Clean data? Rocket ship. Bad data? Dead weight. Execution lives and dies on the inputs.

1️⃣2️⃣ Budgets Shift to Execution Units
You’re not buying licenses, you’re buying throughput. Fund pods like campaigns. No seat limits, no usage tiers. Just: did the job get done? CFOs are going to love this.

1️⃣3️⃣ Silos Collapse
Cloud Employees don’t care about departments. They execute across the org. Marketing, sales, CX, ops, connected. Politics die. Collaboration begins.

1️⃣4️⃣ Work Becomes Modular
Spin up Cloud Employees like code. Start a campaign by lunch, shut it down by dinner. Test onboarding flows, launch outbound sequences. No headcount, no drama. Just velocity.

1️⃣5️⃣ Execution Becomes Memory
Every task teaches the system. Output builds muscle. The org gets smarter daily. And the playbook? It writes itself.

The Autonomous Enterprise isn’t a someday concept. It’s being built right now by companies that are choosing output over headcount, orchestration over control, and capability over hierarchy.

The winners in the next decade won’t just have AI, they’ll know how to run it,  at scale, with precision, and with the same rigor they expect from their best human hires.

So the real question is… when your competitor’s 50-person team is outproducing your 500, what’s your excuse?

Yeah, I guess it’s more powerful… but it felt… colder. Less… me.

With GPT-4o, I’d throw an idea at it and it would make me feel like a genius. It hyped me up. It got me.

GPT-5? It was like talking to a super-smart coworker who can’t read the room. I didn’t think that mattered, until it did. And then I was like, hey wait a second…

It’s not like I found some magic button that fixed everything, but one resource actually helped me turn the corner was Pietro Schirano’s Magic Paths Guide. 

It broke down exactly why my GPT-5 prompts were falling flat and how to get some personality back: 

A few quick wins I stole from his guide:

▸ Tell it exactly what tone and style you want, don’t make it guess.
▸ Use his “spec format”: spell out the task, format, order, and what not to do.
▸ Build in a gut check: “After each step, make sure this still hits the mark.”
▸ Split up independent tasks so it can tackle them in parallel.

Worth a read, you can check it out here.

Trace
Zapier grew up and hit the gym. Trace routes your workflows, punts the boring stuff to AI, and only bugs you when it matters.

Onlook
Draw a website like you’re sketching on a napkin. It writes the code while you play designer.

Informed
The news, but read in your voice. Or Trump’s. Or Taylor Swift’s. Pick your poison. (Mac only right now.)

Genspark AI Designer
Type one prompt, walk away with a full brand system. Logo, colors, site, packaging. Boom.

VibeVoice (Microsoft)
90 minutes of dead-realistic speech, multiple voices. Podcasts, audiobooks, pitches… instant studio.