Every week I see the same kind of post.
Someone rebuilt a product in a weekend.
Someone shipped a Chrome extension in one evening.
Someone with "no coding background" made a Telegram bot, a CRM, a game, a SaaS MVP.
Impressive stuff.
Then the reasonable question shows up:
If vibe coding is real, where are the results?
If software just got dramatically cheaper to create, where is the flood of new software?
Where are the thousands of breakout apps?
Where is the visible "AI effect" outside demo videos and screenshots?
It is a fair question.
But I think we are looking in the wrong place.
We expected a public explosion
The default assumption is simple.
If developers are now 2x, 5x, or 10x faster, we should see 2x, 5x, or 10x more public output.
More apps.
More packages.
More open-source projects.
More polished products.
But that logic comes from the old software world.
In the old world, building software was expensive enough that things never got built at all. A tool had to justify coordination, budget, handoffs, engineering time, QA, and deployment effort.
Now that threshold is collapsing.
And when the threshold collapses, the shape of output changes.
Not everything becomes a startup.
A lot of things become software for one person, one team, one workflow, one week.
That kind of result is real.
It is just not very visible.
The biggest AI effect might be invisible software
This is the part I think many people miss.
Vibe coding does not only create more public code.
It creates more private usefulness.
A founder builds an internal dashboard and never publishes it.
A marketer makes an interactive lead magnet instead of a PDF.
An ops person creates a Telegram bot that pulls reports every morning.
A PM turns user feedback into a prototype the same day.
A small team stops paying for three SaaS tools because they built two tiny internal ones.
None of that shows up as a famous open-source library.
None of that lands on a public package index.
None of that necessarily becomes a company.
But it is still a result.
Maybe the first real dividend of vibe coding is not "more software companies."
Maybe it is "more software where there previously would have been no software at all."
The old metric was output. The new metric is attempt rate.
That is the comparison I find more interesting.
Before AI, a lot of ideas died early because turning them into code was too expensive.
Now the cost of trying is much lower.
So the real shift may not be:
- more finished products
It may be:
- more attempts
- more experiments
- more prototypes
- more internal tools
- more narrow solutions
- more one-person products
That is not as sexy as saying "AI made everyone a 10x engineer."
But it might be more true.
I know people who would never have hired a developer for a tiny personal workflow.
Now they build it themselves.
I know PMs who would never have asked engineering for a half-baked idea.
Now they prototype it first.
I know founders who ship something real before opening Jira.
That matters.
There are vibe results. They just do not look like old-school results.
When people ask "where are the results?", they usually imagine one of three things:
- a breakout startup
- a polished public app
- a meaningful spike in open-source output
Those are valid signals.
But they are lagging signals.
Vibe coding shows up earlier and in messier places.
It shows up when a team replaces a manual workflow with a small script.
It shows up when a user interview turns into a live prototype in hours, not weeks.
It shows up when a non-technical person stops asking for permission to test an idea.
It shows up when buying software becomes less attractive than building something narrow yourself.
That is a very different type of outcome.
Less "new unicorn."
More "thousands of little software decisions that now finally happen."
But let's be honest: vibe results are not the same as durable results
This is where the hype usually overreaches.
Yes, AI made creation much faster.
No, that does not automatically mean the whole product lifecycle got easier.
Architecture is still hard.
Debugging complex state is still hard.
Security is still hard.
Distribution is still hard.
Maintenance is still hard.
Turning "a thing that works" into "a thing people trust" is still hard.
That is why the internet is not suddenly full of perfect AI-built products.
Vibe coding helps a lot with:
- first versions
- automation
- glue code
- admin tools
- prototypes
- experiments
- narrow utilities
It helps much less with:
- long-term coherence
- team-scale architecture
- product judgment
- go-to-market
- operational reliability
So the right conclusion is not "vibe coding is fake."
The right conclusion is:
vibe coding is excellent at creating motion, but motion is not the same as compounding product value.
What should we measure instead?
If we really want to understand whether vibe coding produces results, I think we need better metrics.
Not just "how many new apps," but better questions might be:
- How much faster does a team move from idea to first usable prototype?
- How many user feedback points get validated in the same week?
- How many internal workflows get automated that otherwise would stay manual?
- How much SaaS spend gets avoided because teams build narrow tools themselves?
- How many product ideas reach real users before dying in planning?
- How much does software creation spread beyond traditional developers?
Those feel much closer to the real effect.
Because AI is not only changing software supply.
It is changing what is worth building in the first place.
My take
So, is there vibe coding?
Absolutely.
Are there vibe results?
Yes. But not always where people expect them.
The visible internet may not yet show a Cambrian explosion of polished new software.
But under the surface, I think there is already an explosion of attempted software, useful internal software, and previously unjustifiable software.
That is not nothing.
That is a real shift.
The mistake is expecting the first AI wave to look like a neat chart of public artifacts.
The real first wave might look more like this:
- more side tools than startups
- more prototypes than platforms
- more automation than apps
- more local wins than global revolutions
And honestly, that still feels big.
Maybe the better question is not:
Where are the vibe results?
Maybe it is:
How much software used to be too expensive to exist — and now suddenly is not?