Pinned Tweet
Your design skill becomes your agent skill.
Most of design exploration stage work still happens on canvas. Nevertheless when working with agents in @figma we still have to prompt extensively and explain what we are doing.
I got really curious: what if agents could see what you do and act based on your actual working style? This is how Cast for @figma was conceived.
Cast is agent layer that provides real-time collaboration loop along with the snapshots of your last ≈250-300 actions, and tools that mimic user Figma tools.
Use it via CLI using console harnesses (Claude Code, Cursor, Codex, Pi, etc) or with API keys inside of Figma (OpenAI/OpenRouter).
Try it and tell me what you think.
↓ Link below ↓
Slitrobo retweeted
ways @ldf_gm and i *do* use ai to scale ourselves ↓
shared brain:
- put company docs into a searchable database
- connect slack, docs, notion, dropbox, and granola
- connect agent to slack so anyone can access the info they need without human getting in the way
design & website:
- design system codified into machine-readable guidelines
- custom agentic tool for self serve visual assets (shoutout @slitrobo for these x2)
- website vibe coded from brand guide (@ldf_gm is now our cto lol)
- slack connected so we can prompt website and asset updates from company chat
- automated file conversions
content & data:
- put marketing source material and guidelines into projects so anyone in the org can generate onbrand content
- train projects with founder POVs to revise founder comms
- clip and caption videos (rec: opus or riverside)
- aggregate video transcripts for written content
- analyze audience data to support targeting
- data milestones auto tracked with notifs in slack
- soon: above notification triggers visual asset production
- soon: auto generated seo content from brand messaging
- soon: completed seo content triggers visual asset production from machine readable brand guide
how we do *not* use ai ↓
- come up with ideas
- create strategy docs
- create frameworks
- create source material for ai
- lock in creative direction
- develop brand guidelines
- gen ai visuals without guidelines
- write first or final cuts of public copy
- message teams or partners
- and of course, to shitpost
I haven’t felt this satisfied with a design project in years. Thank you @artistatmachine @emmajo @ldf_gm
Our design team had an absolute blast collaborating with @artistatmachine on the brand expansion and custom content production tool.
9:00 @ the broadcast if you want to see it!
Thanks for the shoutout and showing @slitrobo’s work @emmajo @ldf_gm ❤️
Slitrobo retweeted
Numbers software designers should know
8.3 ms = one frame at 120 Hz
16.7 ms = one frame at 60 Hz.
100 ms = response feels instantaneous
200–300 ms = transitions feel snappy
300–500 ms = transitions feel deliberate
1 second = delay interrupts thought
10 seconds = waiting loses attention
44 × 44 pt = comfortable minimum touch target
4.5:1 = minimum contrast for ordinary text
3:1 = minimum contrast for large text
45–90 characters = comfortable body line length
1.2–1.45× font size = comfortable body line height
Slitrobo retweeted
Your agent just got more creative.
Today we’re launching the TITLES MCP, bringing every artist model and tool on TITLES into chat.
Search any model by vibe. Create image, video, and audio. Every time an artist model is used, they get paid.
Install: titles.xyz/mcp
ALT TITLES MCP - Quick Guide
Slitrobo retweeted
“Many of my tracks are better if you play them at 33 rpm. I have never denied that. That’s also why my pieces are so short: you can only press them onto maxi singles if they are short at 45 rpm.”
Aphex Twin (Sonic Press, 1996)
