AI native creative strategy doesn’t need your performance data to figure out what to make next.
Humans barely do.
Replying to @tomdmeredith
Solve unsolvable math problems!
But please read this file before attempting to write meta ad copy.
There's gonna be a whole new skill set... "How to Market to Machines."
I'm gonna start writing our emails for the agents instead of the human. Maybe they can advocate on our behalf.
There's not enough "better design through AI" chat.
Most AI design gets labelled as slop...which the first pass usually is. But, just like a normal design process, you need to try things, give notes, and do a lot of revisions.
Design wins!
@lil_dill's talk is now my most viewed YT video of all time. Amazing.
It's insane... that at the end of 2026 you still can't copy-paste an image into a LinkedIn post.
- You still have to save an image to your computer.
- Click "Media"
- Find the image on your computer.
- And upload.
It's not just devs... designers too.
I'm hill climbing a rebrand and it's way better than anything Opus has done in the past.
Maybe it's a skill issue.
I'm being so real with you right now @theo. This video has progressed the space more than any other one you have made.
Personal software is turning into personal hardware.
Every member of your creative team should be on the $100/m Claude plan.
If they aren’t using 50% of it each week… replace them.
It’s a whole new world for creative strategists.
xitter.cf/jasonyimco/status/2115…
Nothing is more frustrating than calling my dot and having to wait 3 or more rings for it to pick up…
Anyone else burning through the "free" usage credits while we wait for the reset?
Maybe that was the plan all along.
It gets even worse on the Grok Heavy + Cursor Ultra.
You have to manage 4 usage pools...
1. Cursor model quota (Composer/Grok)
2. Cursor other model quota (Anthropic)
3. Grok quota (Build, Chat)
4. Grok Bot quota
The new Angel's Stadium is gonna cook.
Even if it ends up in Inglewood.
OpenAI launched a $230 Codex controller. I looked at my Stream Deck and thought... why don't I just make my own?
Now two keys show my Codex + Claude capacity and whether I'll run out before the refresh.
You can just do things now.
I have this theory I call the AI Hamburger…
Top Bun: Human 1 outsources work to AI
Meat: AI outsources work to human
Bottom Bun: Human 2 does the work
Replying to @tomdmeredith @ycombinator
For real. Paperwork is my kyrptonite so this is the only way it’s happening.
Yesterday agent booked a mobile notary to come to the office to get stuff done 😅
Has Claude quietly brought back "peak-hour rate limits"?
I get significantly less usage during 5 hr Sessions during US business hours.
Just me?
Replying to @claudeai
Effective today, we are:
1) Doubling Claude Code’s 5-hour rate limits for Pro, Max, and Team plans;
2) Removing the peak hours limit reduction on Claude Code for Pro and Max plans; and
3) Substantially raising our API rate limits for Opus models.




















