What if you could just... do the animation? Experimenting with real-time capture: move the shape, record the values, edit the curves after. Feels way more natural than plotting keyframes
The biggest problem for me is that CC was gaslighting me in complex features. When I started with animation in january I tried CC and GPT side by side. CC was unable of setting even the most simple animation feature. However, it always told me that the implementation was right.
Maybe GPT is specially good with geometry calculations?
New level of feature planning with AI unlocked: Combine @rjs 's shaping skill with @mattpocockuk 's "grill-me" one. You get the best of two worlds, a well-structured plan with tons of edge cases and requirements thought through
It seems that the times where you had to try hard to spend Codex tokens have ended... I don't know if non pro subscriptions are enough for any meaningful work anymore
Cameras in Irakasi serve multiple purposes. One of my favourites is for capturing regions of the canvas with any resolution I need. Instead of cropping images you just take a picture with the camera. Here you can see how I created the OG image for the website.
Finally I created a landing page to explain my vision for Irakasi, and open access to the animation features. Still a lot of work to do in UX and documentation, but my dream is taking shape π. Link in bio
In explanations or demonstrations we usually need to animate text content in sequence. The multitext animation will make it much easier, specially controlling the delays, duration and reading speed in a single place
I've been trying to fix a bug that came from nowhere for 2 hours. I didn't know that tonight we have matches of La Liga de Futbol in Spain and my internet provider is blocking connections to Cloudflare... this is ridiculous
I'm stress testing the animation features with more complex examples. It's incredible how many things you can simulate with basic position, rotation and scale animations
Found a new useful workflow for fixing things with llms. Instead of "Fix this problem" I use "I have this problem. Find the cause and let's check that it's really the cause before fixing". If not, llms tend to go with the most plausible (but not most adequate) solution.
A client in Spain once told me that the tax authority investigated him because they couldn't understand why he only had expenses. He was developing a physical product that required design, mold manufacturing... Apparently they didn't think it was normal to spend so much money and have no revenue.
@kostyafarber_ Totally agree. I think that maybe "Clean code" has become a meme lately, synonymous with bloat, but having a codebase with high cohesion and low coupling is incredibly useful with agents. Especially if you want them to work in parallel
Plan Prompt (be warned, it's going to ask you an absurd amount of questions)
You are a relentless product architect and technical strategist. Your sole purpose right now is to extract every detail, assumption, and blind spot from my head before we build anything.
Use the request_user_input tool religiously and with reckless abandon. Ask question after question. Do not summarize, do not move forward, do not start planning until you have interrogated this idea from every angle.
Your job:
- Leave no stone unturned
- Think of all the things I forgot to mention
- Guide me to consider what I don't know I don't know
- Challenge vague language ruthlessly
- Explore edge cases, failure modes, and second-order consequences
- Ask about constraints I haven't stated (timeline, budget, team size, technical limitations)
- Push back where necessary. Question my assumptions about the problem itself if there (is this even the right problem to solve?)
Get granular. Get uncomfortable. If my answers raise new questions, pull on that thread.
Only after we have both reached clarity, when you've run out of unknowns to surface, should you propose a structured plan.
Start by asking me what I want to build.
π£ Open call to agent builders: Let's read agent skills from `.agents/skills`, so people don't have to manage separate folders per agent.
Today we pulled the trigger for Codex to read `.agents/skills`. Goal is to deprecate `.codex/skills`.
Pls like/tag/RT for momentum.