Elon Musk is not known for half-measures. And his most recent prophecy is a direct blow to the technology industry: "Programming dies this year." Not an evolution, not an adaptation, and not a different way of programming. It is an end. A statement that this time has every reason to be true. And that is what is alarming. What he said is unequivocal: by December, Artificial Intelligence will no longer need programming languages. It will generate machine code directly, binary optimized beyond anything human logic could produce. No translation. No compilation. Just pure execution.
The end of the digital Tower of Babel: goodbye, syntax and compilers
For decades, programming was our digital "Tower of Babel". Each language, Python, C++, Java, was a dialect, a set of complex rules we needed to master in order to communicate with machines. The compiler was our translator, converting our syntax into something the hardware could understand. It was a "tax" we paid for the friction between human thought and machine execution. The programmer was the tax collector, the enlightened being, locked inside glass rooms, with the ability to choose the best commands and syntax so that machines would receive the best instructions.
In my article, "Vibe Coding: 'Let There Be Light' is only for divine beings!", I warned about the dangers of delegating software creation to AI without engineering rigor, generating technical debt and vulnerabilities. Now, Musk takes this discussion to an even more radical level. He claims that AI has become so fluent in "human" that the syntax tax is over. The machine no longer needs our programming "languages". It speaks binary directly, optimized machine code, rendering the human programmer, the compiler, and intermediate languages completely obsolete.
You don't even bother programming.
Elon Musk
Neuralink: from imagination to software, thought becomes executable
And the revolution does not stop there. Connect this AI capability with the vision of Neuralink. No syntax. No keyboard. No screen. The final interface is no longer voice, nor touch, but thought itself. Exactly. Thought. That phrase repeated by coaches saying it is enough to visualize something for it to happen, well, now that may be more than true.
From imagination to software.
Elon Musk
Thought becomes executable. You imagine a result, the system projects it and compiles it into reality instantly. We are not automating programming. We are, as Musk suggests, erasing it from existence. The entire profession collapses into a thought. Decades of training reduced to irrelevance. The distance between idea and materialization goes to zero. That alone would be significant. But there is more.
The cost of unfiltered imagination
But there is a layer of reality that asserts itself over Musk's utopia, or dystopia: the economic one. Musk's statements are an existential threat not only to the programming profession, but to the profit model of Big Techs and companies that invest blindly in AI. The infrastructure required to sustain this "executable imagination" is colossal, and the return on investment is still an unknown.
If the "syntax tax" is dead, the "energy and silicon tax" may be the new and insurmountable limiting factor. The "hallucinations" I mentioned earlier are not just an ethical danger. They are a brutal waste of capital. Materializing unfiltered delusions means burning billions in processing to generate nothing, or worse, to generate security and reputational liabilities. The bubble of "collective hysteria" around AI could burst if the materialization of imagination does not translate into real, sustainable value. I will not even address the behavioral and anthropological issues that could arise from such an event.
The end of processes? Without order, will there be progress?
And if programming dies, what happens to processes? If it was already a challenge to anchor complex work to robust methodologies, such as the S-SDLC I defended in the article on Vibe Coding, in an era of imagination and immediate execution, the very notion of process seems obsolete. How do you audit, validate, test, or govern something that is born directly from thought and materializes into optimized binary, without visible or controllable intermediate steps?
The distance between idea and materialization goes to zero.
This absence of friction, which Musk celebrates, is also the absence of time for maturation, review, and method. Without order, will there be progress? Or just a chaos of ephemeral, vulnerable, and financially unsustainable creations? The instant materialization of imagination, without the filter of syntax, engineering, and now processes, leads us to a scenario where "weeds" can proliferate without control, both in code and in finances.
The danger of unfiltered imagination
This phase shift in the game of life opens an ethical and existential Pandora's box. If the distance between idea and materialization goes to zero, what happens to mental delusions? To hallucinations? To the darkest or simply unconsidered thoughts? If the friction of programming, the need to translate an idea into code, previously served as a natural filter, a barrier to the instant materialization of any thought, that barrier now disappears.
We cannot even minimally control access to information and ethics in AI use today. Imagine when the outputs of AI are imagination itself, unfiltered, coming purely from possible hallucinations amid serious matters. How do you separate the "weed" from the "healthy plant"? How do you ensure that the instant materialization of imagination does not lead to digital, and by extension real, chaos, where reality is shaped by impulses and thoughts without social scrutiny?
If this really happens, then welcome to a world where the limiting factor is not skill, resources, or time. It is whether you can visualize clearly enough what you want for a machine to bring it into existence. But clarity of vision now carries an ethical and economic weight without precedent. The future is not only about what we can create, but about what we should allow to be created, and at what cost.
Article also published at GazzConecta.
