There are many things in the world that can be paperweights.
A stone can do it.
A stapler can do it.
A coffee mug can do it, although that creates a different problem.
The funny thing about a paperweight is that it doesn't have to be especially good at being itself.
It only has to be heavy.
And perhaps that is why expensive paperweights often have names engraved on them.
When the function is ordinary, the name has to work a little harder.
Lately, I have been wondering whether the CTO role has become something like that in many companies.
An important-looking object.
Quite heavy.
Nicely placed.
Name engraved.
And nobody is entirely sure what would happen if you removed it.
Perhaps nothing.
That is the uncomfortable part.
There Are Really Only Two Jobs
If we remove the conferences, the executive offsites, the strategy decks, and the LinkedIn posts about “building high-performing cultures,” the CTO has two very simple jobs.
Simple does not mean easy.
It usually means the opposite.
When looking upward, the CTO translates.
The CEO does not really need to know why one distributed system behaves differently from another.
The board does not wake up in the morning thinking about database replication.
They need someone to say:
This is possible.
This is expensive.
This will probably break.
This can wait.
This cannot.
And if we rush this, here is the bill we will receive later.
Not a technical explanation.
A usable explanation.
Something another human being can understand well enough to make a decision.
Then there is the other direction.
When looking downward, the CTO decides.
Engineers do not always need one more person to say, “Interesting. Let's hear everyone's perspective.”
Sometimes they need someone to say:
We are doing this.
Not that.
This deadline moves.
That one stays.
We are accepting this technical debt.
I understand the consequences.
And if I am wrong, my name is on the mistake.
That last sentence is important.
Leadership is partly the strange profession of being wrong where everyone can see you.
And yet, after watching technology organizations for a long time, I think many CTOs do neither of these things very often.
Upward, they become vague.
There are many qualifications.
Many diagrams.
Many sentences containing the phrase “it depends.”
Eventually, the board understands almost nothing.
But everyone feels reassured because a technical person attended the meeting.
Downward, the CTO becomes a facilitator.
People align.
They socialize the proposal.
They create space.
They seek consensus.
And then, somehow, a decision appears without anyone quite making it.
Usually the person who ships first wins.
Or the person who talks the longest.
This is not necessarily leadership.
It may simply be gravity.
If you remove translation and decision-making from the CTO role, something remains.
Presence.
A chair in the executive meeting.
A box near the top of the organization chart.
Someone who is technically “accountable.”
In other words:
Weight.
Which brings us back to the paperweight.
AI Didn't Kill the CTO.
It May Have Removed the Paper.
For a long time, being the paperweight was not such a bad arrangement.
There was plenty of paper.
Engineering organizations were large.
Systems were complicated.
Boards were often uncomfortable with technology.
Simply having a credible senior technical person in the room had value.
The weight really did hold something down.
But now the desk is changing.
And the paper is disappearing from underneath the weight.
The Pragmatic Engineer recently described a remarkable number of CTOs, VPs of Engineering, and Heads of Engineering choosing to leave prestigious, well-paid jobs.
Often without another job waiting.
Six out of ten engineering leaders surveyed were either seriously thinking about leaving or already moving toward the exit.
People usually do not walk away from impressive titles because Tuesday was difficult.
Something deeper has changed.
I think there are several things happening at once.
1. The Founder Has Discovered a New Toy
One departing CTO described the problem as managing a founder's “AI psychosis.”
This sounds dramatic.
Until you imagine the following Monday morning.
The founder has spent the weekend using AI coding tools.
They are delighted.
Productivity has arrived.
The future is here.
And also, there is now a 60,000-line pull request.
The founder is smiling.
The CTO is not.
Now comes a difficult choice.
The CTO can explain that the company has just produced an impressive quantity of future suffering.
In which case the CTO becomes the person who hates innovation.
Or the CTO can say nothing.
In which case everyone learns a different lesson:
Standards are optional.
Architecture is optional.
Roadmaps are optional.
Agreements are optional.
And perhaps the CTO is optional too.
A paperweight is supposed to stop papers from moving.
But what happens when the owner of the desk keeps pulling the papers out from underneath it?
The paperweight remains.
Very dignified.
Very heavy.
Holding down absolutely nothing.
2. The CTO Has Become a Magic Object
Companies once hired CTOs to lead technology.
Now some companies seem to hire them to radiate transformation.
“Please make us AI-native.”
Wonderful.
What does that mean?
Nobody is entirely sure.
But it would be nice if engineering costs fell by 30 percent.
And delivery speed doubled.
And nobody new was hired.
And quality improved.
And morale remained high.
And perhaps the AI tooling budget could also somehow stay small.
This is a rather impressive wish list.
Claire Vo, the former Chief Product & Technology Officer at LaunchDarkly, has argued that many product and engineering organizations will never truly become AI-native because the change requires unusually good change management.
That seems important.
Technology is often the easy part.
Humans are the complicated software.
But organizations frequently give CTOs an enormous transformation target without the authority, time, or support needed to accomplish it.
So the CTO receives a new assignment:
Please sit here and emit AI transformation.
Like a small shrine.
And if transformation does not occur, everyone knows which shrine to blame.
Meanwhile, experienced technical leaders are discovering another unpleasant truth.
If they spend several years managing AI transformation without actually building with the tools themselves, their own skills may become less valuable.
Charity Majors, CTO of Honeycomb, has argued that lacking meaningful AI experience is becoming a serious career risk — serious enough that taking an individual-contributor role may sometimes be the smarter move.
That is an extraordinary inversion.
The person sitting on top of the paper stack begins thinking:
Perhaps I should become paper again.
3. The Desk Is Getting Smaller
There is also a simpler reason.
Engineering teams are becoming smaller.
In some AI-forward companies, one or two engineers can now handle work that once required a larger team.
Each engineer may also be coordinating several AI agents.
At some point, adding another human does not necessarily create more capacity.
Sometimes it creates another meeting.
Full-stack development reduced some specialization.
Cross-platform tools reduced more.
The management flattening that began several years ago did not stop.
And technical founders can remain directly involved in engineering much longer than before.
So the old question was:
“When do we need a VP of Engineering?”
The new question may be:
“Do we?”
A desk containing three sheets of paper does not require a two-kilogram object with an engraved brass plate.
It may require a small stone.
Perhaps even a mug.
4. The Paperweight Eventually Reads the Cap Table
Then there is money.
Money has an interesting ability to turn philosophical concerns into resignation letters.
CTO compensation often includes substantial equity.
Which means part of the bargain is:
Work very hard now.
Possibly become wealthy later.
The important word is “possibly.”
The Pragmatic Engineer article tells the story of a CTO holding an unusually generous 2% stake in common shares.
That sounds excellent.
Until you calculate what happens before the common shareholders get paid.
Investor liquidation preferences came first.
In this case, Series A investors had negotiated a 2x preference.
The result was that the company needed an exit north of roughly $200 million before the CTO's common equity became worth anything.
The CTO did the math.
Then did something perhaps even more important.
He believed the math.
He concluded that the exit was unlikely.
So the equity was effectively worth zero.
Nobody had written “ZERO” on the certificate.
But numbers do not become kinder because we refuse to introduce ourselves.
There is a small irony here.
The CTO is often one of the people in the company most capable of understanding exactly this situation.
Senior technology leaders are expected to understand business.
So the paperweight may be the first object on the desk to discover that the desk itself is not worth very much.
That is a different kind of weight.
Still, Weight Is Not Useless
After saying all this, I should say something that sounds contradictory.
I think the real CTO function may be more valuable now than before.
Not the title.
The function.
Because paper is flying everywhere.
AI enthusiasm.
AI panic.
Founder experiments.
Sloppy pull requests.
Quarterly pressure.
Board anxiety.
Cost cutting.
Consultants using the phrase “AI-native” with unusual confidence.
In that environment, something that genuinely holds things steady has tremendous value.
Someone who can tell the board what is actually happening.
Someone who can tell engineers what they are actually going to do.
Someone who can absorb uncertainty and return a decision.
Someone who can say:
“No, this is not magic.”
Or:
“Yes, this changes everything.”
And know the difference.
That person is not furniture.
That person may be one of the rarest people in the building.
The problem is that the letters C, T, and O stopped guaranteeing that person a long time ago.
Now the environment is testing every engraved nameplate.
Some are attached to leaders.
Some are attached to rocks.
Perhaps the Best Leaders Are Leaving First
One interesting part of the current exodus is that the people walking away are not necessarily incompetent.
Often the opposite.
They can see the arrangement clearly.
They understand when responsibility has become disconnected from authority.
They understand when equity has become imaginary.
They understand when they have been asked to perform transformation theater.
And they do not want to become part of the furniture.
The Pragmatic Engineer article gives a useful counterexample.
Matt Boyle, who became VP of Engineering at Gitpod, later Ona, did something unusual before taking the job.
He interviewed the company.
He wanted to know whether the organization genuinely intended to embrace the changes created by AI.
Not just talk about them.
Not just hire someone senior and hope transformation leaked out of their pockets.
He found a company where the answer was yes.
And apparently, it worked well.
There is something nice about that.
The paperweight chose the desk.
Two Questions
So perhaps the CTO title itself is not the important thing.
Perhaps there are only two questions worth asking.
When was the last time you made a decision that someone below you could immediately act on?
And:
When was the last time you explained something so clearly that someone above you could actually make a better decision?
If you can answer both questions quickly, good.
The current exodus may even be an opportunity.
There are still desks that need real weight.
But if the answers take a long time to find, perhaps the problem is not AI.
Perhaps it is not the founder.
Perhaps it is not the board.
Perhaps, somewhere along the way, the job quietly turned into an object with a title engraved on it.
And an engraved rock is still a rock.
Reference：
https://newsletter.pragmaticengineer.com/p/the-great-engineering-leader-career-break
Top comments (0)