The Co-Founder Memo

A product co-founder in your inbox.

Happy Sunday! Today:

1. From the Lab: On taking the lead
2. From the Trenches: The never-ending ai agent
3. From the Chair: On mentors
4. From the Feed: Cool podcasts!

Let's go!

🧠 From the Lab

A group of friends got together for dinner.

2 of them would take care of cooking for everyone.

But one of the two was in a really bad mood after a bad week.

They had different menus to cook, but none of the two felt very confident about one specific meal.

They were open to trying out anything.

So they picked one idea and started working on it.

After 10-15 minutes into it, they felt uninspired.

The guy originally in a bad mood kept saying how uninspired he was and how sorry he was about that.

Now they were both in a bad mood.

So they decided to press pause and think.

What do we really want to do?

One of them said to the other "just make the call, I'm not in the right mood to do it".

If you had to make the call - what would you do?

Option A: experiment something new

Option B: do the thing you know you'll be able to take the lead on

Last week I was in a very similar situation.

And I took path A.

It went terribly bad.

I wasn't warmed up.

I felt uncertainty at all times when doing so. My bad mood wasn't helping.

And my buddy's bad mood wasn't inspiring at all.

We ended up calling it and wrapping it up.

And I took a hard lesson home.

When you work with a team, you need to create a culture of vulnerability and brutal honesty.

When someone isn't in the right mood, they should be able to talk about it, and transparently say "please take the lead for the team".

At that moment, the new leader can go pick a path where she knows she'll be able to have everyone else's back.

This isn't about risk vs safety.

It's about taking the lead.

And being able to inspire and cheer everyone else as the leader.

It might be your turn this time.

And it will be someone else's turn when you're in a bad place.

🔧 From the Trenches

We've been building multi-agent web and mobile apps at Jams for the past 6 months or so.

And I keep getting this insight over and over.

One of the biggest challenges when building AI agents, is knowing when to stop, and how to say no.

Talking to an agent through a conversational AI is just addictive.

This is why people spend so many hours on GPT or whatever.

Running small tests on any LLM takes seconds.

And they'll convince you that the LLM is capable of doing what you want it to do.

So yes, every little task you do on GPT is a good starting point for ideation.

But just know that what GPT does in its own context window has nothing to do with what a custom agent could be able to do in a specific environment (like a mobile app).

Real life example:

Every time we do product demos with our clients, they start ideating on the spot about new potential behaviors the agent could have.

It's even funnier post-demo.

One or two days later they'd keep coming back with a message or a phone call with a new idea.

"Can we get the agent to review the user's profile photo and improving it before approving the account?"

"What if the user doesn't have a photo? Could we create a sub-agent that's an avatar generator?"

"And what about the onboarding once they're in? We could create an AI video tour"

All we wanted to do in the first place was a private community of nomads with a built-in conversational agent, not Lindy.ai

And while yes, most of these things are actually viable technically, they'll keep delaying the launch and with that, the customer feedback.

I keep seeing this pattern everywhere on teams.

And it's actually pretty cool to have this level of creativity, tools and ideas around products.

Just remember a rule of thumb:

Never stop launching iterations.

And think of this as a continuous product improvement engine.

Happy AI shipping.

💭 From the Chair

There are multiple ways to get a person to mentor you on something specific.

Sometimes a coach will become a mentor.

Sometimes a close friend with experience in XYZ.

Sometimes a past client, colleague, or investor.

The point of mentors is to be able to spot them, because many times they'll show up in unsolicited contexts.

I founded mentors in past clients. People whom I've never considered mentors before. People who don't consider themselves as my mentors.

I've been thinking about this over and over, and there's a clear signal to chase in many contexts.

Try to find people who did what you're doing today, 20 years ago.

And ff you have those around you, AND they value you, AND they give you feedback, make sure you take care of that relationship.

You have a mentor right in front of you.

Take them to dinner. Gift them a good wine. Ask how they're doing. Listen.

Try to add value. And pay attention to their feedback.

This comes from a true story of me collaborating with a mentor of mine for 2 years (client of mine) before realizing how good of a mentor he's always been to me.

Look for the oldies. Stick around them. Figure out how to add value so you can take them for coffee.

Keep building that relationship from there.

Just remember to be patient. It takes a lot of time.

🔗 From the Feed

Recommended podcasts from last week:

• ​Tim Ferriss w/ David Senra from Founders​
• ​David Senra interviews Daniel Ek​
• ​The Hospitality Principles that Build Billion-Dollar Startups (this one is really good)
• ​I can't find balance by Jonathan Courtney (might be hard to get used to his podcasts but he's a founder just openly talking about stuff)
• ​How to get rich (without luck, talent or a trust fund) by MFM​

If you made it this far, really appreciate you.

Until next Sunday,

Juan

P.S. If you know someone who might benefit from reading this, please share.

The Co-Founder Memo goes out every Sunday. Join 1,000+ founders here.