Greg Isenberg - Crazy week

Hey there,

Some things I can’t stop thinking about this week…

1. I stumbled across this job posting looking to hire AI agents, not human employees. Sounds like a bad joke, right? But it's legit. And now I can't stop thinking about it. The idea of companies filling roles with artificial intelligence instead of people is just blowing my mind. I mean, we all knew this was coming, but seeing it actually happening is a serious wake-up call. Wild times ahead.

2. We're about to see software companies capture value that used to belong to agencies, consulting firms, and entire departments. Companies don’t budget for tools—they budget for outcomes.

If you pitch an AI tool, they’ll compare it to existing software.
If you pitch an AI-powered team, they’ll compare it to headcount.

3. Met a founder who only builds products for one specific person: Sandra, 42, head of ops at a midsize insurance company. Not “ideal customer profile” - literally one person. Every feature decision is “would Sandra use this?” His team has a cardboard cutout of Sandra in their office. Weird strategy but they hit $1M ARR in 6 months.

4. Had a fascinating interaction at a coffee shop yesterday. Owner put up an iPad showing live comments from their Google reviews. Customers were literally ordering things they saw in the feed. Made me realize social proof isn’t just for websites - it’s becoming live entertainment. Businesses that show their love in real-time are winning.

5. Been studying Drunk Elephant’s (exit for $845M) playbook. People think they won on “clean beauty” but that’s wrong. They won by making skincare feel like mixing cocktails. Every product comes with a “mixing guide”, customers combine different “shots” for custom routines. Turned boring skin routines into a fun ritual. Now doing $450M in revenue. Product design drives behavior, behavior drives revenue.They also had an exclusive with Sephora for years. Gets you thinking how you can build an unfair advantage like into your startup.

6. Was thinking about how Figma really won: they didn’t just beat Adobe on product (they did) - they beat them on distribution. Every Figma file became a tiny viral loop. Build a great prototype, share it, suddenly your whole team wants Figma. While Adobe was selling seats, Figma was building a network. Same pattern is emerging with AI tools in 2025. The winners won’t just have the best tech - they’ll have the best sharing mechanics built into their core product. Watching Cursor take over code reviews because devs naturally share it, while better standalone AI coding tools struggle to grow.

7. Startup idea I can’t stop thinking about: “Paperwork” - generate every document a business needs to operate. Here’s why it’s going to print:

• Companies waste $8-50k/year on legal templates and policies
• Most just copy-paste from competitors or use sketchy templates
• Train AI on premium docs from top companies
• Charge $199/month for unlimited custom docs
• Target specific verticals (restaurants first - they need 30+ docs to operate)
• At 5000 customers = $12M ARR
• Distribution through restaurant POS systems, SMB accountants
• Could exit to Toast/Square who want to own more of the stack

8. I started a new Apple Note called "Questions" to capture all the unknowns swirling in my head. It's already up to 10 entries, mostly around the future of tech and the crazy opportunities emerging. In a world moving this fast, sometimes the best way to get clarity isn't to chase answers, but to double down on asking the right questions. Embracing uncertainty and wrestling with the juiciest unknowns is how we sharpen our vision and spot the whitespace others miss.

Some sample questions on the list range from "is no code dead?" to "how do you hire AI agents?" to "what's an unpopular business to start right in the tech community now and why?"

Let me know if any of these spark thoughts for you.

9. More random thoughts, tutorials and startup ideas on the podcast. Some new episodes for you to listen to

• How to build Pixar/Disney using AI (ChatGPT, Leonardo AI, Kling AI, ElevenLabs). YouTube, Spotify, Apple.
• 4 microsaas ideas that could do $100k/month. YouTube, Spotify, Apple.

Did you like this format for Greg's Letter? Let me know :)

#1

Zach 🏴 @zachpogrob Some days you have so many ideas it causes your brain physical pain 6:28 PM • Feb 4, 2025 17 Retweets 198 Likes Read 14 replies

​

I'm feeling like this more and more in this AI world right now. Maybe you are too.

#2

Logan Kilpatrick @OfficialLoganK As the internet grows and fills with AI: - search is 100x more important - private communities become key - human connection premium goes up - authenticity is differentiation What else am I missing? 11:55 AM • Feb 2, 2025 132 Retweets 1977 Likes Read 374 replies

​

Couldn't agree more.

#3

The Boring Marketer @boringmarketer I built an agent to automate video research & script generation. I'm sharing the complete workflow, setup steps, output, plus some big ideas on how it will help me: 11:53 AM • Feb 6, 2025 5 Retweets 103 Likes Read 12 replies

​

Been obsessed recently about learning how to do automations like this. Really amazing what you can do nowadays.

#4

Norgard @BrianNorgard What makes meme coins interesting is the perception everyone has a chance. That feeling has been absent from the Internet for a long time. 4:50 PM • Feb 2, 2025 1 Retweets 29 Likes Read 5 replies

​

True.

Thank you for reading Greg's Letter. I hope you it got your creative juices flowing. You can forward this email to a friend that might benefit.

You might also enjoy joining our paid membership Startup Empire to accelerate your ideas and business.

And feel free to click follow/subscribe on The Startup Ideas pod on Apple, Spotify and/or YouTube to get more startup ideas in your feed. It's free, so why not?

Until next week.

Be well,

© 2025 Late Checkout LLC
113 Cherry St #92768, Seattle, WA 98104-2205
​Unsubscribe