If everything on a slide moves, nothing on it feels important. Spent the last while building an AI Agent skill around that idea.
smalldigits.com/blog/narrati…
Every time I built a scroll-driven article, I had to re-explain the same effects to my coding agent from scratch.
Prompts don't persist. Skills do.
So I wrote it once as a SKILL.md. The demo page shows the structure, framework, and every technique with its purpose, where to use it, and a live example of each.
aakatheeri.github.io/narrati…
#ClaudeCode #AgentSkills #buildinpublic
I read many premium #articles that explain things through a story as you scroll. Most of them reuse the same handful of techniques.
Out of all the techniques they use, these 5 are the ones that kept pulling me down the page.
Here's each one, and when to use it 👇
✦ Before / After Comparison:
Two states in the same frame, with a handle the reader drags.
Why it hooks: change is hard to perceive from two images side by side, but trivial when they occupy the same pixels. The reader's hand does the analysis.
Use it when: the transformation is your argument: redesigns, cleanups, growth over time.
That's 5.
Each one answers a different attention problem: starting momentum, holding a model, feeling alive, tracking actors, seeing change.
#buildinpublic #scrollytelling #datavisualization #frontend #UX #design
Did Google just dethrone OpenAI and Anthropic? youtu.be/b2tjnxX8IHU?si=j9oF…
#Argon
Reusable AI agents don't just scale good habits. They scale bad ones too, faster.
This piece covers the guardrails worth building in before you turn a prompt into an agent.
Worth bookmarking:
smalldigits.com/blog/how-to-…
#BuildInPublic #IndieHackers #ProductManagement
Ahmed Alkatheeri retweeted
Industrial software doesn’t have to feel like industrial software
Manage warehouses like you’re playing a strategy game
Built with Opus 5.5
Ahmed Alkatheeri retweeted
Replying to @ramzcoder
By continuing to engineer software, and by realizing that engineering software now means the management of agents to produce that software.
Using agents will make you X times more productive. This does not mean that companies will require fewer programmers. It means that companies will get X times more done, and make X times more profit.
I expect the demand for software engineers, who know how to successfully drive agents, to go through the roof.
Most solo builders don’t fail for lack of ideas—they slow at the starting line.
The real blocker? No structure for launch tasks.
Here’s how I solve it: → 4 GenAI prompt templates for customer synthesis, benchmarking, experiments, and copywriting.
From blank page to launch-ready—fast.
Details here 👇
smalldigits.com/blog/four-re…
#GenAI #ProjectManagement #BuildInPublic #SoftwareQuality
Vibe coding works, and it's fun to use to make interesting, useful things. Right now it still costs tokens on strong models — or, if you go open-source, it costs hardware instead. The cost doesn't disappear; it just moves.
Most of the actual work now is reviewing what the agent built and deciding how the app should be structured. Ready-made agents can do a lot of this for you. But someone still has to make sure it's built on something solid.




