Static Images, AI Voice Clones and the 'From Scratch' Series Format

LinkedIn

The 17th-century inventor nobody credits for public transport

“In 1662, a mathematician invented something we use every day — and almost nobody knows his name for it.”

In 1662, a mathematician invented something we use every day — and almost nobody knows his name for it. Blaise Pascal — yes, the Pascal of Pascal's Triangle and Pascal's Wager — created the world's first fixed-route public transport system in Paris. Horse-drawn carriages on a set schedule, with a fixed fare, open to anyone who could pay. It ran for fifteen years before it collapsed after his death. The lesson I keep coming back to: the same mind that cracked probability theory also cracked urban mobility. Pascal did not silo his thinking. He asked 'what is the hardest coordination problem in this city right now?' and applied his best tool to it. The best founders I know do the same thing. What problem in your industry is waiting for someone to just apply a tool that already exists somewhere else? I am curious — drop your answer below.

Why it works: The famous-name-unexpected-achievement structure creates an immediate 'I never knew that' reaction that LinkedIn's algorithm reads as dwell time. The open question at the end is specific enough to generate substantive comments rather than one-word reactions.

YouTube

Switzerland at 1:1 scale — and what it reveals about city size

“Someone just built Switzerland block by block at actual real-world scale — and the comparisons are wild.”

Someone just built Switzerland block by block at actual real-world scale — and the comparisons are wild. The project at blocks.manifold3d.ch renders the entire country in 3D at a 1-to-1 ratio, so you can fly over Zurich and then immediately compare it to, say, Los Angeles. Here is what jumps out: the entire urban footprint of Switzerland — every city, every town — fits inside the greater LA metro area with room to spare. But Switzerland moves 60 percent of its population by public transit every day. LA moves about 3 percent. Same land area, completely different infrastructure logic. This is what I want to explore in this series: what does scale actually reveal about how cities work? Next episode I am looking at Tokyo versus Texas. If there is a comparison you want to see, put it in the comments — I will do the top-voted one in episode 3.

Why it works: The scale comparison makes an abstract urban-planning argument visceral and visual. Seeding the next episode in the comments section turns viewers into co-producers and signals high community engagement to the algorithm.

X

The anti-AI-slop design diff that actually matters

“I ran the same UI prompt twice. The only difference was a 200-line design language file. Here is the diff.”

I ran the same UI prompt twice. The only difference was a 200-line design language file. Here is the diff. [Image 1: raw AI output — generic card layout, default blue, Comic Sans energy] [Image 2: same prompt + Hallmark design language — consistent type scale, real spacing, actual hierarchy] [Image 3: the three lines of the prompt that did the most work] The file is called Hallmark. It lives at github.com/Nutlope/hallmark. It is free. It works with Claude Code, Cursor and Codex. The core idea: AI does not have taste by default. You have to give it a taste file. 2244 people starred this in the last 48 hours. That number tells you how many developers are quietly solving the same problem. Save this for the next time someone tells you AI-generated design is good enough.

Why it works: Three images do the work a thousand words cannot on X. The star count provides social proof without hype language. Ending with a reframe ('save this for next time') primes the save action, which is the highest-value engagement signal on X.