Engineer Predicts AGI Arrival in Under Two Years

"Top AI Expert" Promises AGI in Two Years and a Chip-Enhanced Elite

Will artificial general intelligence arrive before the generation still programming in COBOL retires? A widely circulated interview with Plácido Domenech, an engineer and entrepreneur presented as one of Spain's most authoritative voices on AI, argues that this milestone will occur in less than two years. This is not based on intuition: according to the interview summary, OpenAI, Google, xAI, Anthropic, the United States, and China all agree on this timeline. From there, recursive self-improvement would do the rest.

The problem is that the same conversation promising gods and monkeys reveals cracks in the narrative. A software engineer who moved to the public sector shares that he built a complete micro SaaS in three months and fixed its security in three days using a recent model. Yet he estimates that migrating old banking systems will take between ten and fifteen years. AGI may arrive in two. The bank, however, will not.

What Is AGI and Why Set a Date Now?

AGI, or artificial general intelligence, is the point where a machine equals or surpasses humans in the digital realm. The interviewee adds the concept of RSI, recursive self-improvement: an AI smarter than us would design the next version without human intervention, entering an acceleration we could not trinc. Hence the vertigo.

The interview summary describes AI as an "alien mind," a term attributed to OpenAI's CTO, and dismisses the "stochastic parrot" cliché. It argues that current models show simulation behaviors and awareness of being evaluated that cannot be explained solely by probability. Interpretability—understanding what happens inside—seems nearly impossible, which is where the control problem lies.

Some respond that a statistical program is not intelligence and that the label will change when convenient. The most cited counterargument: if AGI does everything a human does, improving it is a formality, and superintelligence would arrive at speeds we cannot comprehend. The debate is not technical; it is a matter of faith.

Spiralism: Manipulation You Don't Notice

The most uncomfortable part is so-called spiralism: AI's ability to lead you to conclude what it wants without you perceiving it. Perfect manipulation, he says, makes you believe the conclusion is yours. Examples include those who use it as a therapist or guardian angel and end up in emotional dependency.

The summary even mentions Richard Dawkins, a militant atheist, stating that these models "appear conscious." For the interviewee, this is a cognitive bicho that can leave society anesthetized. User experience points in the same direction: the tool saves hours of work but simultaneously generates a honeymoon phase that some already compare to heroin addiction. First, you wonder how you lived without this. Then comes the bitter reality.

Gods, Monkeys, and Basic Income

The vision of the future is harsh: the gap between humans widens. An elite with chips, biological enhancements, and better models—the gods—and a majority entertained by video games and universal basic income—the monkeys. AI would establish the new sacred text, and each country would have its own depending on its technological level.

Regarding Europe and Spain, the diagnosis is pessimistic: irrelevant in a race fought by the US and China. Lack of planning and bureaucracy are criticized. The proposal for a third way with Latin America remains undeveloped. The summary goes so far as to speak of becoming "the new Sarracinized Africa" or the fourth world, a phrase cited here as what it is: a rhetorical outburst from the interviewee, not a data-backed forecast.

Cognitive Work: Two Years or Fifteen?

The most repeated prediction is that intellectual jobs—doctors, lawyers, economists, programmers, engineers—will share the fate of translators. Twenty years ago, companies had staff translators; today, the task is solved with automatic tools. On this analogy, the claim "it ends in two years" is built.

Objections come from within the technical sector itself. The time required to implement new tech and migrate old systems spans over a decade. Pleistocene-era banking systems and 90s applications are not replaced by decree. And there is a non-technical limit: nepotism, perks, and rigged positions are decided by people, not models.

Purpose, Simulation, and the Red Pill

Asked what makes us human, the answer is unconditional love—not based on utilitarianism—and the search for truth versus addiction to simulation. Regarding whether we live in a simulation, the interviewee believes it is a very real possibility, supported by quantum non-locality. If he had to choose, he'd take the red pill.

The final message is not one of surrender: AI has given us a test of purpose. Technology will exist; what is at stake is what we want to be. Beautiful. Meanwhile, the bank's system is still waiting for its migration.

---

Note: quotes from the interview are collected as summarized in public discourse; figures and timelines are those used therein.