So far this year I’ve planted:
- 2 x Asian persimmon
- 2 x peach
- 1 x pomegranate
- 1 x fig
Currently in a rush to obtain and plant:
- multiple American persimmon (host plant for Luna moth, regal moth)
- paw paws (host for zebra swallowtail butterfly)
- drawf figs
- all the berries
- passion fruit vines
The token cost to build a production feature is now lower than the meeting cost to discuss building that feature.
Let me rephrase.
It is literally cheaper to build the thing and see if it works than to have a 30 minute planning meeting about whether you should build it.
It’s wild when you think about it.
This completely inverts how you should run a software organization. The planning layer becomes the bottleneck because the building layer is essentially free. The cost of code has dropped to essentially 0.
The rational response is to eliminate planning for anything that can be tested empirically. Don’t debate whether a feature will work.
Just build it in 2 hours, measure it with a group of customers, and then decide to kill or keep it.
I saw a startup operating this way and their build velocity is up 20x. Decision quality is up because every decision is informed by a real prototype, not a slide deck and an expensive meeting.
We went from “move fast and break things” to “move fast and build everything.”
The planning industrial complex is dead.
Thank god.
something strange happened today - friend sent me an audio - but that was an audio that I sent to someone else many days ago.
once I played - that audio is not available on the sender or my phone.
not sure how that glitch happened
I’ve taken 4 auto rides this week. All between 15-25mins. Every auto driver has been on the phone non-stop for every single minute of every ride. It’s the same in stores, salons, shops,… Everyone is busy on their phones instead of doing work. We have a serious cellphone addiction problem in our country.
I said "what is the point of your son running a saas (@onetimefax, for those who don't know) if you can't use it for free?" and my mom was very offended.
Spring break 2023: Cancun
2024: Paris
2025: Puerto Rico
2026: ???
Was originally going to take the family to Mexico City, procrastinated, didn’t get tickets on time, and now… maybe I’ll wait a bit.
Direct flights to Ogdensburg are cheap, I should figure out what that is.
Just randomly thought of a really funny (to me) incident I helped with.
This particular service read things off of a Kafka topic and sent those out to an external service. Kafka is at-least-once delivery, for the uninformed. However the external service had zero tolerance for duplicates. You can pretend each API call to their system placed an order for pizza or something.
So this team designed a basic idempotence system. When they processed a Kafka message they checked Redis for the existence of the message id. If it already existed, the message had already been processed, so was discarded. If it didn't exist, they processed the message and added the id to Redis.
Fast forward to a couple months and anytime we have a ton of traffic we have duplicates at the third party that has zero tolerance for duplicates and they're pissed.
They start poring over the code trying to figure out where the bug is. They run a ton of tests on Stage. Cannot even reproduce the issue. Only happens on Prod.
I get summoned. I ask for a summary. Hmm, I see. Can I see your infrastructure?
Your Redis memory is full. When your cluster runs out of memory, it cannot store the idempotence identifier, so the next time, it assumes this message hasn't been sent yet. This only happens during high traffic because the TTL expiry hasn't hit yet.
Copilot is powerful—but it gets dramatically better when you give it the right context.
Here are 4 practical context-engineering techniques to improve Copilot’s accuracy 🧵👇