|
Welcome to The Weekly Fizz, a newsletter companion to our podcast.
The Drip
Jensen just told the world every computer will have its own AI agent.
Meanwhile, a company called StrongDM launched something they're calling the "software factory," where humans literally cannot see the code. Only AI writes it. Only humans review the output. Two very different signals pointing in the same direction:
AI isn't staying in the chat window. It's moving to the edge, into the workflow, onto the device.
We break down both stories and what they mean on this week's episode.
Inside The Bottle
This is the first issue of The Fizz, so let's start with why we exist.
Kellan and Justin started Tiny Bottle AI just a few weeks ago. The name comes from a song Justin’s college band wrote. The lyric was: "I can get it all into a tiny bottle. It's as simple as this."
That line captures the bet we're making. After a few years of watching companies try to "transform" with AI, we kept seeing the same pattern: teams that spread themselves thin across dozens of use cases got nowhere.
Tiny Bottle is the antithesis of that. We don't believe everyone needs a comprehensive AI strategy. We believe choice and focus are what will set companies that get ahead apart.
We're building the company the same way. Our website went from zero to launch in two and a half weeks. Two people. No designer. No agency. We just focused on our most important problem and employed AI to the most practical extent we could.
Kellan put together a visual of our Git history and the pattern was clear: the first 80% happened fast. Tons of commits, rapid iteration, multiple design options generated in a day. Then the last 20% was slower, more specific, more deliberate. That's where the real value lives. Not the demo. The product.
That 80/20 split is the story of AI adoption right now. Most teams are stuck in the first 80%, impressed by demos but shipping nothing production-grade. The ones winning are staying for the last 20%, getting specific, iterating tightly, and knowing when to push versus when to step back and let the tools catch up.
Lab Notes
| ■ | Justin's note: My oldest son has a Python learning app on his Chromebook. I looked at it and thought: you're not going to need that. The skill that matters isn't writing code. It's knowing what to build and why. That's not going away. |
| ■ | Kellan's note: Building with AI is night and day depending on the medium. Code-based design (like our SVG logo) is easy to iterate on — tight feedback loops, specific changes. Image generation? You have to speak its visual language or you'll bounce off it. The gap is closing, but it's not close yet. |
Latest Article
Why We Started Tiny Bottle AI
What Stopped Our Scroll
| ■ | What 81,000 people actually want from AI — Anthropic's massive qualitative study across 159 countries. Cuts through the Twitter discourse to show real hopes and fears. Primary source gold for content and strategy. |
|
