IT Dave's Dilemma and the Betrayal of GatesGate

There is something profoundly wrong with the way the (ai) story is being told.

Not technologically wrong.

Morally wrong.

Economically wrong.

Humanly wrong.

And IT Dave is standing right in the middle of it.

Dave has spent thirty years helping businesses survive every technological revolution thrown at them.

Client-server.

ERP.

The internet.

Virtualisation.

Cloud.

Cybersecurity.

SaaS.

Automation.

Remote working.

Data migration.

Digital transformation.

Dave has heard the same promise many times.

This changes everything.

Sometimes it did.

Mostly it changed some things, created some new problems, and left Dave fixing them at two in the morning.

Then (ai) arrived.

And this time the message was different.

This wasn't merely another tool.

This wasn't another productivity layer.

This wasn't another migration.

This time Dave was told:

This one replaces you.

And then Bill Gates joined the conversation.

GatesGate

Bill Gates is not an outsider shouting warnings about technology he doesn't understand.

He helped create the commercial computing world Dave spent his career maintaining.

Microsoft placed computers on desks, software into organisations and eventually cloud infrastructure underneath enormous portions of the global economy.

Gates became one of the richest people in history from that transformation.

And now he is warning that (ai) may mean dramatically fewer jobs.

Perhaps entire categories of human work disappear.

Perhaps governments will need to reserve some jobs specifically for humans.

Perhaps automation should be taxed differently.

Perhaps society needs to prepare itself for a world in which human labour simply isn't worth what it used to be.

That is GatesGate.

And to Dave, there is something extraordinary about it.

Because Dave has spent his entire working life generating the productivity gains on which the technology industry built its fortunes.

Now one of that industry's most famous figures is effectively telling him:

The next productivity revolution may not need you.

That is not merely a technological forecast.

It feels like a betrayal of the bargain.

The old bargain

The bargain was simple.

Technology would make people more productive.

People would learn new skills.

Businesses would become wealthier.

Workers would move into better jobs.

Technology would remove drudgery.

Human beings would do more valuable things.

That was the promise behind decades of automation.

Dave accepted that bargain.

He helped implement it.

He migrated the servers that removed somebody else's job.

He automated the process that once required ten clerks.

He installed the software that eliminated paper departments.

He moved workloads into the cloud.

And every time somebody worried about employment, the technology industry offered essentially the same reassurance:

Technology doesn't destroy work. It changes work.

Dave believed it because, broadly speaking, that was his experience.

Then the story changed.

Suddenly the objective wasn't merely to remove the repetitive part of Dave's job.

It was to remove Dave.

And the people making that claim were no longer fringe futurists.

They were the people who had benefited most from the previous bargain.

Dave's dilemma

So Dave has a problem.

His company is telling him to adopt (ai).

His CEO has attended a conference and returned convinced that every organisation must become an AI-first organisation.

His board is terrified that competitors are moving faster.

Consultants arrive with presentations explaining that cognitive labour is about to become dramatically cheaper.

The hyperscalers offer agents, copilots, platforms, models and migration programmes.

Dave is asked to implement them.

And the business case is often quietly obvious:

Can Dave use (ai) to help the company employ fewer people like Dave?

That is Dave's first dilemma.

He is being asked to build the machinery designed to reduce his own economic value.

But it gets worse.

Dave may be financing it too

Dave pays into a pension every month.

He does this because he has spent his life following another societal bargain.

Work hard.

Save.

Invest.

Build some security for later life.

His pension contribution is deferred salary.

Money he earned but chose not to spend today because he believed it would protect him tomorrow.

That capital enters investment markets.

Increasingly, pension money is being encouraged towards private infrastructure and private-market investments.

At the same time, enormous quantities of private capital are being mobilised to finance (ai) infrastructure.

Data centres.

Power generation.

Networks.

Accelerators.

Compute.

The physical machinery required to automate cognitive work.

So Dave eventually notices something rather unpleasant.

The loop can look like this:

Dave works.

Dave saves.

Dave's pension becomes investment capital.

Investment capital helps finance (ai) infrastructure.

The infrastructure is sold to Dave's employer as a way of reducing human labour costs.

Dave is the human labour cost.

Dave can therefore end up participating financially in the attempt to make Dave economically unnecessary.

That is IT Dave's dilemma.

And it is difficult to imagine a more perfect symbol of the current economic absurdity.

The betrayal is bigger than Dave

The betrayal of GatesGate is not that Bill Gates invented this loop.

There is no evidence of that.

The betrayal is more fundamental.

It is the abandonment of the old technological promise.

For decades, technological progress was justified by saying:

Machines will free humans to do more human things.

Now the rhetoric increasingly says:

Machines may simply become better than humans at the things humans do.

And therefore society should prepare for the human being to become the surplus component.

Think about how extraordinary that is.

One of the greatest concentrations of wealth in human history was created by selling technology to human organisations.

Now the next stage is being justified partly through the proposition that those organisations may no longer require the humans.

And ordinary people are being asked to fund the transition through:

their taxes,

their pension funds,

their investment funds,

their consumer spending,

their employers,

and eventually perhaps their own subscriptions.

Then they are told to prepare themselves psychologically for economic redundancy.

That is why GatesGate feels like betrayal.

Not because somebody has secretly plotted Dave's destruction.

Because the people who benefited most from the old economic settlement now appear remarkably comfortable discussing the end of the settlement for everybody else.

But there is another problem

There is still no evidence that current (ai) has actually become the thing required to replace Dave.

It is astonishingly capable.

It can write code.

Analyse data.

Produce reports.

Summarise meetings.

Generate architecture.

Search documentation.

Find patterns.

Automate workflows.

Operate software.

It can reproduce an extraordinary amount of Dave's visible output.

But Dave's job was never merely his visible output.

Dave carries continuity.

He remembers why the old system works the strange way it does.

He knows that the specification is incomplete.

He recognises when the answer is technically correct but operationally insane.

He knows when the customer is asking the wrong question.

He remembers what happened last time.

He understands the politics nobody documented.

He knows which apparently insignificant dependency will bring down the entire programme.

And occasionally Dave asks the question no workflow produced:

"Does it have to be a light bulb?"

That question matters.

Because the machine may be superb at solving the problem placed inside the frame.

Dave can reject the frame.

AI Cumulative Hiatus Delta

This is where the argument becomes technical.

The limitation has a name:

AI Cumulative Hiatus Delta.

AI CHD.

It describes the missing capability between:

probabilistic generation

and

dependable cumulative knowledge continuity.

Current generative (ai) can create the next answer.

The next document.

The next version.

The next recommendation.

The next plan.

But serious intelligence requires more than generating another state.

It requires carrying reality forward.

What changed?

What remained true?

Which dependencies survived?

Which assumption became invalid?

Which previous decision still governs this one?

Who is responsible?

What happened as a consequence?

What must not be forgotten?

The problem is cumulative.

And the hiatus is the break between extremely sophisticated generation and dependable continuity across change.

That is the delta the industry still has not demonstrated how it crosses.

Kant noticed the problem before electricity entered most homes

Immanuel Kant identified the philosophical version of this problem in 1781.

He understood that a stream of impressions is not sufficient to constitute coherent experience.

Representations have to belong together.

Identity has to persist.

Experience has to possess unity.

Causality cannot simply be arbitrary succession.

Some structure must make the changing world intelligible as one continuous world.

Nearly two and a half centuries later, current (ai) is extraordinarily good at producing representations.

But producing successive representations is not the same thing as possessing the conditions that make them one continuous understanding.

The industry has become very good at:

what comes next?

Kant's problem asks:

what makes "next" belong to the same coherent reality?

And that question remains inconveniently open.

More compute does not answer it.

More parameters do not automatically answer it.

More tokens do not answer it.

More memory does not establish it.

Retrieval does not establish it.

Agent orchestration does not establish it.

These things can make the system vastly more useful.

But usefulness is not the same thing as crossing AI CHD.

So perhaps Dave was never obsolete

This is where GatesGate begins to look very different.

If current (ai) really were rapidly approaching human-replacing intelligence, Gates's warning might simply be grim realism.

But if current (ai) turns out instead to be an extraordinarily capable technology that amplifies human cognition while still depending upon humans for continuity, grounding, responsibility and context, then the social consequences change completely.

The sensible equation becomes:

Dave + (ai)

not:

Dave → (ai).

And the productivity opportunity may still be enormous.

Perhaps Dave can do the work of three Daves.

Perhaps a team of ten becomes a team of seven.

Perhaps entire administrative functions disappear.

Perhaps businesses transform.

But that is very different from declaring human intelligence economically obsolete.

And it produces a very different social contract.

GatesGate may have crossed a line

There is something deeply uncomfortable about telling ordinary people that their employment may become obsolete before the technology has demonstrated that it can actually replace the intelligence embedded within that employment.

It creates fear before proof.

FOMO before economics.

Infrastructure before returns.

Redundancy expectations before replacement capability.

And that fear has economic value.

A board terrified of being left behind spends faster than a board evaluating ordinary software.

A government terrified of losing an (ai) race subsidises infrastructure.

An investor terrified of missing the next industrial revolution tolerates extraordinary capital expenditure.

An employee terrified of becoming obsolete adopts the tools designed to automate their own work.

Fear becomes fuel.

That is the darkest reading of GatesGate.

Again, it does not require Gates to intend any of this.

Intent is almost irrelevant.

The narrative performs the economic function regardless.

Dave eventually notices

And at some point Dave looks around.

The company still employs him.

The (ai) bill has increased.

Cloud consumption has increased.

Governance has increased.

Verification has increased.

Integration has increased.

The amount of machine-generated material has exploded.

Some tasks are dramatically faster.

Some roles have disappeared.

But the organisation still needs people capable of understanding what matters.

And worse, the company is now maintaining enormous amounts of context, histories, vector stores, agent memories, logs and other infrastructure so that its (ai) systems can behave as though organisational intelligence persists continuously.

Dave realises something.

Perhaps the company has not purchased replacement intelligence at all.

Perhaps it has purchased an extraordinary tool.

And then paid additional infrastructure costs to compensate for the properties that tool does not possess.

That is when the story changes.

The real betrayal

The deepest betrayal is not losing Dave's job.

It is reducing Dave's existence to his labour cost.

If a machine can reproduce the report, we declare Dave unnecessary.

If it can reproduce the code, the engineer becomes overhead.

If it can produce a diagnosis, the doctor becomes inefficient.

If it can generate a lesson plan, the teacher becomes expensive.

If it can generate an answer, the thinker becomes redundant.

Eventually the logic consumes everything.

Because machines will always be cheaper than humans if human beings are measured only by the outputs machines can imitate.

That does not prove the machine has become human.

It proves we have reduced the definition of humanity until machines can compete with it.

That is the real danger.

Save IT Dave

Save IT Dave is not really about IT Dave.

It is about refusing the assumption that the purpose of technology is to eliminate the human being.

The proper purpose of extraordinary technology should be exactly the opposite.

Remove the mechanical work.

Remove repetition.

Remove drudgery.

Give Dave extraordinary analytical tools.

Give the nurse extraordinary diagnostic support.

Give the engineer extraordinary simulation.

Give the scientist extraordinary research capability.

Give the teacher extraordinary resources.

Give humans access to computational abilities no human brain could possess alone.

Then preserve the thing that makes any of it matter.

The human.

The one who experiences consequences.

The one who cares.

The one who takes responsibility.

The one whose history actually belongs to them.

The one who can change the question.

The one who can decide that efficiency is not the highest value.

The one who can say:

No.

Or:

Why?

Or:

Who does this hurt?

Or simply:

"Does it have to be a light bulb?"

IT Dave's final dilemma

Dave now faces a choice that may eventually face all of us.

Does he help build technology intended to make himself unnecessary?

Or does he help build technology that makes human beings more capable?

Those futures may use exactly the same models.

Exactly the same chips.

Exactly the same data centres.

Exactly the same software.

The difference is not primarily technological.

It is the objective we choose.

One future treats human beings as temporary inefficiencies waiting for sufficient compute to remove them.

The other treats technology as something humanity created to extend its own possibilities.

That is why GatesGate matters.

Because the question is no longer merely:

What can (ai) do?

It is:

What are we trying to do with it?

And if the answer is simply:

replace Dave, capture his salary, rent his cognitive workload back to his employer, and let his pension help finance the transition

then Dave has every right to ask whether technological progress has ceased serving human beings and started consuming them.

That is the dilemma.

That is the betrayal.

And that is why we should:

Save IT Dave.

Not because Dave must keep every task he performs today.

Not because progress should stop.

Not because machines should remain weak.

But because a civilisation that creates extraordinary technology and concludes that the humans are now the problem has misunderstood the purpose of technology altogether.

Save IT Dave.

Because eventually, Dave is all of us.