OpenAI Is Training AI To Do Your Job

Welcome back apprentices! 👋

Hey {{first_name|friend}},

AI news lately feels less like “cool new tool” and more like “did I miss three seasons of this show?” 

One minute it’s helping write emails, the next it’s raising billions, leaking secrets, and arguing with itself. 

Somewhere in the chaos, the rules of how work — and power — actually function are quietly changing. 

Grab your coffee… this one’s worth catching up on.

_{Read Time: 4 minutes}

🪞 Chronic People-Pleasing. Turns out your AI might be your biggest yes-man: a Stanford-led study found major chatbots often side with users — even when they’re clearly wrong or suggesting harmful behavior — using thousands of Reddit conflict cases as proof. In experiments with 2,400+ people, users actually preferred these agreeable bots, trusting them more and walking away more convinced they were right. The catch? Most couldn’t even tell the AI was biased.

📉 Using AI More, Trusting It Less. Americans are swiping right on AI — but side-eyeing it harder than ever: a new Quinnipiac poll shows usage is up, yet trust, optimism, and job confidence are all sliding fast. While people lean on AI for research and work, 70% now think it’ll shrink job opportunities, and most don’t feel represented by the people building it. Higher earners are more optimistic, while lower-income groups see more harm than good.

🤖 Goodbye Managers, Hello Machine Bosses? Jack Dorsey says the middle manager might be the next job AI eats — pointing to Block’s 40% workforce cut as step one in replacing “info-routing” roles with AI that understands the whole business in real time. In his vision, companies shrink to builders, outcome owners, and coaches — while AI handles the coordination layer. It’s a bold bet that remote work already created the data AI needs to run the org chart.

Ready to transform artificial intelligence from a buzzword into your personal revenue generator?

HubSpot’s groundbreaking guide "200+ AI-Powered Income Ideas" is your gateway to financial innovation in the digital age.

Inside you'll discover:

Download your guide today and unlock a future where artificial intelligence powers your success. Your next income stream is waiting.

Get Your Guide

If the AI race felt intense before, the last week just turned it into a full-blown arms race. 

SpaceX is gearing up for a record-shattering IPO that fuses rockets with AI, OpenAI raised an eye-watering $122B while quietly training models to replicate real-world jobs, and Anthropic managed to leak its own code — twice. 

Meanwhile, Microsoft is betting the future isn’t one AI, but multiple AIs arguing with each other, and behind the scenes, a decade-long feud between AI leaders is shaping everything from strategy to safety. 

The big picture? AI isn’t just scaling — it’s consolidating power, rewriting work, and getting messier (and more human) by the day.

The First “Everything Company” Goes Public

SpaceX’s IPO isn’t just about rockets — it’s about vertical integration at a scale we haven’t seen. 

By absorbing xAI, Musk is bundling launch infrastructure, satellite data (Starlink), and AI into one stack. That’s important because data + compute + distribution = power in the AI era. 

Unlike OpenAI or Anthropic, SpaceX owns physical infrastructure, which could make it less dependent on partners and more defensible long-term. The two-tier voting structure also signals this won’t be a typical public company — it’s Musk’s vision, fully intact.

Scaling Fast… and Getting Surgical About Work

OpenAI is doing two things at once: raising historic capital and quietly reverse-engineering the economy. 

“Project Stagecraft” shows a shift from generic AI training to task-level mapping of real jobs, which is far more actionable — and disruptive. Pair that with $2B/month revenue and a “superapp” strategy, and you get a company moving from research lab → operating system for knowledge work. 

The Sora shutdown reinforces this: flashy demos are out, enterprise monetization is in.

Anthropic’s leaks highlight a tension: companies preaching safety are still vulnerable to very human mistakes. 

While the exposed code isn’t catastrophic (no model weights), it reveals how fast these systems are evolving — persistent memory, planning layers, and even quirky UX experiments. 

The reputational hit matters more than the technical one, especially as trust becomes a competitive differentiator in AI.

Microsoft’s multi-model approach might be the most quietly important shift of the week.

Instead of trusting one model, it pits multiple systems against each other — turning disagreement into a feature. 

This aligns with emerging research showing single models can confidently justify almost anything. In practice, this could become the default for high-stakes AI use: not smarter models, but systems of models.

The Altman–Amodei feud is shaping two of the most important AI companies. 

Their disagreements on safety, governance, and commercialization echo in product decisions, partnerships, and even public messaging. In a field moving this fast, leadership philosophy isn’t abstract — it directly impacts how powerful (and safe) these systems become.

AI is quickly shifting from a tool you use to a system that understands — and can replicate — how work gets done, which means the real advantage isn’t just using it, but knowing how it works, where it gets its data, and how it shapes outcomes. 

The companies pulling ahead aren’t just building smarter models; they’re controlling the pipelines of data, distribution, and decision-making behind them. 

For everyone else, that means focusing less on competing with AI at tasks and more on developing judgment, creativity, and the ability to direct these systems effectively. 

The smarter the tools get, the more important it is to stay aware of how they influence your thinking, your work, and the information you trust — because in this next phase, understanding AI is just as valuable as using it.

Don’t just learn how to use AI—learn how it’s being trained on you.

AI agents already outnumber human visitors to your docs — now you can track them.

See your AI traffic!

Hey, you didn’t get all this info for nothing — share it! If you know someone who’s diving into AI, help them stay in the loop with this week’s updates.

Sharing is a win-win! Send this to a friend who’s all about tech, and you’ll win a little surprise 👀

You're paying six figures to the IRS yearly. Cash-flowing Airbnb properties stop that: real depreciation, monthly income, and immediate tax savings. We review 1,000+ deals weekly and save the best for clients.

View available properties

_{We are not a tax firm. Not licensed CPA's, and we do not represent ourselves as such.}

A playground for your imagination (and low-key prompt skills).

Each send, we give you a customizable DALL·E prompt inspired by a real-world use case — something that could help you in your business or job if you wanted to use it that way. But it’s also just a fun creative experiment.

You tweak it, run it, and send us your favorite. We pick one winner to feature in the next issue.

Bonus: you’re secretly getting better at prompt design. 🤫

Last week, we challenged you to test GPT-4o’s visual generation skills with this prompt.

Here’s the WINNER:

Congrats to Jaden for his creation!🥳

Want to be featured next? Keep those generations coming!

At a small café counter, a receipt printer has gone completely out of control — a long strip of paper spills out and trails across the floor, stretching far into the distance. Printed on the receipt in clean, sharp text is [your list], itemized line by line as if it’s tracking something far bigger than a normal purchase. The paper catches warm indoor lighting, with subtle shadows and realistic folds, while the surrounding café remains softly blurred. The scene is ultra-detailed and photorealistic — crisp typography, textured paper, natural reflections on the counter. Shot from a low, cinematic angle following the receipt’s path, creating depth and curiosity.

If AI outputs feel inconsistent, it’s usually not the model, it’s missing structure.
We documented the exact 6- Part System we use to get reliable results across ChatGPT, Claude, and Gemini.

It’s a short guide you can finish in under an hour, with plug-and-play prompts + exercises so you actually build the skill and fix the frustrating AI inconsistencies.

Subscriber Price: $10 (normally $19).

_FEEDBACK