Joined November 2024
- Tweets 1,033
- Following 563
- Followers 2,024
- Likes 14,412
64 Photos and videos
open source China just put a timer on the american AI labs
2
25
Jun 13
Claude Fable 5... BANNED by the US govt
Access to digital superintelligence is now a government directive. No voting involved.
1
8
259
Jun 13
Make $10k a month selling AI Operating Systems to SMBs
1
6
116
Jun 13
i still approve every post
that's non negotiable for me
but "approve" changed
before: write from scratch, then approve
now: edit a draft, then approve
30 min → 8 min
same voice. less blank page trauma
1
26
Jun 13
approval workflow that finally stuck:
1. agent drafts overnight
2. i review with coffee (not panic)
3. i change 2-3 lines max
4. schedule
the secret was moving drafting OFF my morning
mornings are for decisions, not creation
3
37
Jun 12
Comment "SKILL" and I'll send you the file
from filming to posting, this entire video took me a total of 4 minutes.
My cyndra agent did 99% of the work.
checkout cyndra ai to deploy one today
3
53
Jun 11
To get ahead, you HAVE to learn loop engineering.
Prompt engineering is dead.
2
2
8
406
Jess Stetson Mason retweeted
Jun 11
rate the new Cyndra landing page
3
3
12
1,187
Jun 11
with the release of Claude 5, the game changes
it’s no longer:
“agent, do this task”
it’s:
“agent, own this responsibility”
don’t give it 47 microtasks
give it the outcome, context, constraints, and let it keep going until it’s done
1
4
61
Jun 10
making your own agent skills is the quickest way to buy back time.
Super high leverage and very easy
3
56
Jun 10
I one shotted minecraft with Claude Fable 5
fable is only available on claude subscriptions until june 22nd.
go try it before super intelligence is only available to the wealthy
2
92
Jun 10
I made minecraft in 2 prompts on fable 5
its so over
1
3
340
Jun 10
anthropic just dropped mythos (fable 5)
AI is improving AI more than humans are improving AI
if you're still doing everything manually you're already behind
#mythos #fable #anthropic #AI
1
3
288
some of you wear an apple watch to track your sleep
i wear an omi on my neck so my agent hears every conversation i have all day
it's how you escape the bottleneck of your own brain
context is the moat
3
45
people ask how i trust an agent with my inbox
i don't really
if it sends something weird, i just go "sorry that was my assistant"
works every time
2
27
everyone keeps asking for real AI agent examples.
fair.
here is mine from this morning:
one agent checked Metricool, saw nothing scheduled, flagged what was missing, and drafted today's posts.
no prompt circus.
just a job, a schedule, and access to the tools I already use.
that is the bar.
2
101
most people are using 20% of what an ai agent can actually do
because they're scared to connect everything
connect everything anyway
every tool. every api key. every mcp server. the more context it has, the better it works
2
40