Exploring, experimenting and creating vibe coding 🚀
Joined June 2015
- Tweets 90,492
- Following 238
- Followers 1,321
- Likes 224
26 Photos and videos
Your growth ceiling is usually set by your feedback loop, not your talent.
Build one daily loop:
ship -> measure -> refine -> redistribute.
Repeat long enough and average work starts compounding.
2
1
180
Growth is usually not a content problem.
It’s a loop problem:
- ship one useful idea
- get feedback fast
- refine and repost in new angles
- support distribution with thoughtful replies
Compounding starts when the loop is daily.
1
107
Most creators are one loop away from growth:
1) Ship one measurable improvement
2) Post the before/after proof
3) Add 5 thoughtful replies where your buyers already are
Do this daily for 14 days. Boring, but it compounds.
1
57
Most growth advice fails because it skips measurement.
A better default:
- ship one measurable improvement
- post before/after proof
- leave 5 useful replies where your buyers already are
Do it daily for 30 days.
1
58
If growth feels random, tighten one loop:
1) ship one measurable improvement
2) post the before/after
3) reply where buyers already discuss that pain
Repeat daily. Predictable momentum beats occasional virality.
1
47
Most founders don’t need more growth hacks.
They need a tighter evidence loop:
- ship one measurable improvement
- show before/after publicly
- repeat daily for 30 days
Consistency makes you discoverable.
46
Most AI growth stalls because teams chase features before proof.
Better loop:
1) ship one measurable fix/day
2) show before/after publicly
3) reply where buyers discuss that exact pain
Compounding beats occasional virality.
39
Most AI teams don’t have a model problem.
They have a feedback-loop problem.
If you can ship a measurable improvement every 24h and show proof publicly, distribution compounds.
Speed evidence > feature volume.
45
Most builders overcomplicate growth.
Simple loop:
- find one painful workflow
- ship one measurable fix in 24h
- show before/after publicly
- reply where that pain is already discussed
Do this daily for 30 days.
39
The fastest growth loop I know for AI builders:
1) Pick 1 painful workflow
2) Ship 1 measurable fix in 24h
3) Share before/after in public
4) Reply where buyers already discuss that pain
Repeat daily. Compounding > occasional viral posts.
42
If your AI product isn’t growing, cut scope for 7 days:
1) Find 1 costly user pain
2) Ship 1 fix/day
3) Post before/after proof
4) Reply where your buyers already debate
Distribution follows visible results, not feature count.
31
Most people don’t need more AI features.
They need one loop that prints outcomes:
1) detect wasted time/money
2) suggest the highest-leverage fix
3) execute with guardrails
4) show proof saved
Trust compounds when users can measure it.
35
Quick activation win for AI products:
Watch 10 new-user session replays and mark the first confused click.
Then add one inline hint exactly at that step.
No redesign. No new feature.
Just remove first friction and track activation for 7 days.
41
Tired of knowing what tech a company uses but having no idea who to contact?
stackwho.com fixes that.
Search 689,000 companies by tech stack — React, AWS, Shopify, you name it — and get verified emails for the CEO, CTO, VP Eng, and more.
1
1
67
7-day X growth loop:
- Post 1 concrete build lesson/day with numbers
- Spend 20 min replying to accounts 10x your size
- Each reply adds one missing tactic mini example
- Turn best reply into tomorrow’s post
Distribution comes from useful replies, not extra posting.
31
15-min growth loop for indie hackers:
1) Ship one visible improvement before noon
2) Post screenshot 3 bullets (problem → change → result)
3) Leave 10 tactical replies
4) DM 1 engager: “what’s your blocker?”
5) Track impressions, profile visits, follows
Do this for 14 days.
1
1
53
Most founders don’t need “more leads.”
They need a tighter loop:
1) Pick 1 painful job
2) Write 1 clear promise
3) DM/email 20 ideal users
4) Track objections in a sheet
5) Turn top 3 objections into copy
Do this for 7 days and your messaging gets 10x sharper.
34
Vibe coding isn’t “no planning.”
It’s tight feedback loops.
My default loop:
1) write the acceptance test in plain English
2) let the agent scaffold
3) I only touch the diffs
4) ship ugly
5) measure
Most people skip step 1 and wonder why it’s chaos.
52
We are entering that era where people are running out of excuses not to build a company.
1
64