Apple blocked my fun AI palm reading app because the category was considered “spam.”
Which is funny because I built it for the least spammy reason possible: nostalgia.
Growing up in Sri Lanka, palm reading was always part of the background magic of life. An elder, astrologer, or family friend would take your hand, study the lines, and turn them into a story about who you were and where life might take you.
Then I saw what new vision models could do and went down a rabbit hole.
So I built Chirogram, a private TestFlight experiment to see if AI can make that old ritual feel personal, beautiful, and new again.
I’m only keeping it open for 7 days.
Reply here if you want to try it, or sign up at https://t.co/uNuhsokcAP. iOS only.
The real calculation is whether you keep supply idle and only serve high-LTV markets with a smaller active fleet, or open it up to lower-LTV markets like APAC and India.
The catch is APAC has massive DAU, so serving it fully may require even more GPUs. You either need a much bigger fleet, or you cap availability to a smaller percentage of users in those markets.
So the question is not just “high LTV vs low LTV.” It is “idle supply vs constrained global access.”
The frontier model bubble may have an uncomfortable realization coming.
Most applications don't need the newest trillion-parameter model. "Good enough" models will keep getting cheaper, while on-device AI eats away at inference demand from the other side.
When customers realize they can get 90–99% of the value at a fraction of the cost, pricing power gets tested.
2027 will be interesting.
Pulled the trigger today and switched 100% of Lindy traffic to DeepSeek v4, churning from Anthropic models.
Saves us millions of $ and we're actually seeing an *increase* in performance on many core use cases. Transformative for the business.
Apple blocked my fun AI palm reading app because the category was considered “spam.”
Which is funny because I built it for the least spammy reason possible: nostalgia.
Growing up in Sri Lanka, palm reading was always part of the background magic of life. An elder, astrologer, or family friend would take your hand, study the lines, and turn them into a story about who you were and where life might take you.
Then I saw what new vision models could do and went down a rabbit hole.
So I built Chirogram, a private TestFlight experiment to see if AI can make that old ritual feel personal, beautiful, and new again.
I’m only keeping it open for 7 days.
Reply here if you want to try it, or sign up at https://t.co/uNuhsokcAP. iOS only.
Key to winning:
Choose to be positive and grateful. Then, just keep at it. Time is the great compounder and will do the rest.
So many people just don’t have the discipline to stay positive and grateful. Then time compounds the bitterness instead.
I’m keeping the TestFlight open for 7 days because I want real feedback, not vanity installs.
Try one reading, ask one follow-up question, and tell me honestly: spam, magic, or somewhere in between?
Sign up at https://t.co/uNuhsokcAP and I’ll send you an invite. iOS only for now.
I don’t think the next wave of consumer AI apps wins by being better prompt wrappers.
They win by turning model capability into experiences people can feel.
Chirogram is my small experiment in that direction: an old ritual I grew up around in Sri Lanka, rebuilt with AI and wrapped in a bunch of product details, like a guided scan flow, structured first reading, generated follow-ups, history, guest mode, and little sparks and animations while the reading comes to life.
Here’s the onboarding: