AI Revolution or Downfall?

For years we have been hearing promises about how AI will revolutionize software development. With Claude Opus 4.6, many of these promises are finally becoming reality.

Not faster, but deeper

The surprising truth is: you don’t get faster. At least not in the way most people expected. You don’t suddenly produce ten times more code in the same time.

What changes is depth.

You can explore architectures more thoroughly. You can question decisions you would have previously accepted out of time pressure. You can refactor with a safety net that understands your intentions. You can dive into unfamiliar territories – a new language, a new framework, a new paradigm – without the usual fear of getting lost.

The AI doesn’t replace thinking, it doesn’t replace knowledge or experience – you still need to know what to ask for, and you still need to judge the answers. After all, these models are stochastic algorithms. They predict the most probable next token – be it a word, a line of code, or a musical note. The results can feel like magic, and sometimes they are genuinely breathtaking. In truth, they are more like the Oracle of Delphi than like wise scientists – priestesses in a trance, uttering words that sound profound but are driven by fumes, not by understanding. And just like those ancient oracles, AI hallucinates with great confidence. Probability is not wisdom, though. But it removes the friction – the syntax lookups, the boilerplate, the exact formulation – that used to stand between you and the actual problem.

The loneliness of the developer

There is a side to this that nobody talks about.

When your most productive conversations happen with an AI, something shifts. The casual chat with a colleague about a tricky bug. The whiteboard session where someone draws a terrible diagram and suddenly everything clicks. The lunch break where you accidentally solve a problem by talking about something completely different.

These moments are disappearing.

Not because AI forbids them, but because they become unnecessary for the task at hand. And what is unnecessary tends to fade away.

We are not just changing how we work. We are changing how we connect.

The bigger picture

But the real question is not about software developers. At least not those who are more than just programmers – those who seek to understand the domain, who can think in terms of solutions rather than syntax, and who can communicate those solutions to others. They will be fine – for now.

The real question is about the millions of people whose work is being automated away. Customer support, translation, copywriting, data entry, bookkeeping, basic legal work – the list grows every month.

And here is the uncomfortable thought: Who will buy the products that the machines produce, when fewer and fewer people earn a living?

A capitalist economy needs consumers. Consumers need income. If the machines produce everything but the people earn nothing, the whole system collapses. Not with a bang, but with a slow, quiet decline in demand.

Universal basic income: social project or capitalist life raft?

This is where universal basic income enters the conversation again. Perhaps not in the way its original advocates imagined.

The idea was born as a social project. A way to give people dignity, freedom, and security.

But what if it turns out to be something else entirely? What if universal basic income becomes the rescue anchor for capitalism itself?

The logic is simple: If companies automate most of the work, they save enormously on labour costs. Yet they still need customers. A universal basic income, funded by taxes on automated production, would keep the cycle of consumption alive.

It would not be a gift to the people. It would be a life support system for the market.

Taxing productivity, not labour

There is another idea, perhaps even more fundamental: What if we stopped taxing labour altogether and started taxing productivity instead?

Today, most tax systems are built around human work – income tax, payroll tax, social security contributions. But if machines do the work, there is no income to tax. The tax base erodes while the profits concentrate.

Taxing automated productivity – the output of machines, algorithms, and AI – would redistribute the gains at the source. It would keep the system running without depending on the goodwill of corporations to share their wealth.

Here is the catch, though: In a globalized world, no single country can do this alone. If one nation taxes automation, companies simply move their production elsewhere. The race to the bottom continues.

This is not a national project. It requires international coordination – agreements between governments, not just between companies. And that may be the hardest part of all.

The question we should be asking

The question is not whether AI will change the world. It already has.

The question is: who benefits?

There is something we tend to overlook: every large language model, every music AI, every image generator was trained on the works of the many. Our texts, our music, our images, our conversations – the collective output of human culture – were taken to build products owned by a few. The community provides the raw material. A single corporation shapes the product and captures the profit.

Is this just?

It is hard not to see a pattern here. For centuries, capitalism has extracted from nature without giving back – forests, oceans, atmosphere – treating shared resources as free inputs for private gain. Now the same logic is applied to human knowledge and creativity. The commons are mined, again.

Perhaps this is the deeper lesson: After decades of individualism, of “everyone is the architect of their own fortune”, we need to rediscover what it means to be a community.

If the productivity gains flow only to those who own the machines, we are heading towards a society that is richer than ever – and more unequal than ever.

If we find a way to distribute the gains, we might be heading towards something genuinely new. A society where work is optional, where creativity is valued over productivity, where the question “what do you do?” no longer defines who you are.

I don’t know which way it will go. But I do know that the answer will not come from the machines. It will come from us.