I asked GPT-5.6 Sol a simple question. It searched the web and told me GPT-5.6 wasn’t available in ChatGPT yet.
I was talking to it at the time.
Wrote up what that mistake reveals about model reliability, and the test I’m running next:
https://t.co/FFp6M0nAMr
A few patterns we frequently use with Fable 5:
Use Fable 5 as an "advisor."
An executor (Sonnet 5) calls Fable 5 for guidance.
Most tokens are billed at the lower executor rate.
The vault surfaces the evidence. The judgment is still mine. Blur that line and the writing goes generic, because the conclusions aren’t yours anymore.
I never changed how I take notes. Messy daily notes, same as always. I changed what happens to them afterwards. That one shift produced more drafts than any capture system I ever tried.
Every AI productivity guide starts with “first, build this routine.” That’s where it loses me. Everything that survived in my system attached to a routine that already existed.
30 days into my AI adoption log, the useful pattern is clear:
The best tools didn’t “do more.”
They removed repeated reconstruction.
Capture got easier.
Review got stricter.
Publishing got less brittle.
The real test isn’t capability.
Claude didn’t start my website redesign by asking what color I liked.
It started from the vault.
New piece on why AI gets more useful when it works from a real body of notes, not a blank prompt.
https://t.co/cpMnYhKsc1
When your personal knowledge system starts feeling like a side project, it’s already failing. The vault, the CSS, the plugins are NOT the work. Thinking is the work. The system should disappear into it.
#PKM#DigitalGarden
Most habits don’t fail from lack of willpower. They fail from too much friction.
If I want to read more, the book goes on the pillow. If I want to skip the scroll, the app comes off the home screen.
Design the path. Stop drafting the motivation.
#Habits#Productivity