AI and the End of Drudgery: The Repeated Five-Year Warning

Five Years Until the End of Drudgery, Again

How many times has the end of work as we know it been announced? The figure repeats with the punctuality of a broken clock: five years. The same timeframe used to promise electric cars in every garage, and to announce that teachers, lawyers, and programmers are about to disappear. The debate, which starts with an apocalyptic warning about artificial intelligence, ends up becoming a reckoning between those who already use it daily and those who suspect it’s all smoke.

The starting point is a known warning: AI is not an era of change, but a change of era. The combination of technologies, exponential acceleration, and society's inability to digest that pace—political culture, media, education, administration, unions, parties—paints a complicated decade. There are no solid plans, no pedagogy, no leadership. The phrase that sums up the mood: "It’s not just a labor issue."

Which Jobs Disappear According to Optimists

The circulating list is extensive: teachers out, lawyers out, programmers out—in the end, only architects remain, but they’ll fall too—copywriters out, marketing out, customer service out. The most aggressive forecast claims humanity will advance more in the next 18 months than in the last 500 years. And it adds: chickens and vegetable gardens will still need someone to take them into the sun and water them.

Against this, the everyday experience of those who use it for work. A specific case: asking for help with a couple of work tasks and getting unusable results. "If it’s not for copy/paste and you have to come up with something damn different, AI is useless," summarizes a skeptic. The contrast with another case: teaching technology to a friend’s daughter and seeing how AI solved and broke down problems perfectly just by taking a photo of the question. If that doesn’t replace teachers within five years, something is missing.

The Protein Folding Example and Medicine

The strongest argument in favor doesn’t come from programming, but from science. AI has completed deciphering the 3D folding of all proteins, a problem stuck for decades. Those who have used it for complex medical consultations argue it doesn’t spit out text translated from a database: it relates everything. And they add a technical nuance: language models handle billions of parameters and we don’t fully know how they work inside, just like looking at a piece of brain doesn’t show thought, only neurons and electrochemical changes.

The response comes from the other side: the definition of reasoning belongs to the previous paradigm. AI doesn’t reason; it does something else. We’ll have to invent a new word to understand each other. That’s the underlying discussion: if thinking is just emergent complexity from simplicity, then current models are already on that path.

The 1890 Car and the Five-Year Trap

The most repeated historical comparison: in 1890, a motorized carriage was invented that broke down after 50 kilometers, shattered gears, and stalled. People said it would never replace the horse. Give it time. AI is here to stay, and what we see is the first car of 1890.

The problem is that the five-year deadline has been stated for a year and a half, and will probably continue to be stated in five years. The parallel with the electric car is inevitable: two million vehicles were promised in Spain in five years, and nearly twenty have passed. The figure becomes a rhetorical device that never expires because it is always renewed.

Paid Version Changes Perception

A recurring turning point: paying the €20 monthly subscription. Those who have done so describe a major shift. The advanced reasoning version, not available in the free tier, uses techniques similar to human ones. And the kicker: this is just the generalist and cheap version. What will an AI trained specifically for a task with a €10,000 annual subscription do?

The gap between free users and professional users becomes the axis of disagreement. Those who have only tried the free version conclude the tool is mediocre; those who pay discover it’s a different product. Perception of potential depends on the monthly bill.

Military Impact and the Undisputed Warning

There is an uncomfortable consensus: the worst of AI won’t be its effect on the labor market, but on the military and geostrategic plane. And Europe is last in all this. Drones are the future of military aviation; 90% of pilots will be redundant, leaving at most 10%. The repeated recommendation: stay away from screens, stay away from cities, boost physical presence, relax digitally, and arm yourself.

On the economic front, the forecast is more gloomy than revolutionary. The abundance created by AI will be enjoyed by elites. Society won’t maintain an army of idle people doing nothing, among other reasons because that would generate social conflicts. Money is nothing more than a metaphor for work. And the process has already begun.

Administration Ignoring the Internet

The diagnosis regarding the public sector is unanimous in contempt: administration will ignore AI as it ignored the internet. It only uses it sparingly, and each agency has its own authentication system. Civil servants, according to the recurring joke, will be eliminated by AI. Reality is less funny: security guards don’t disappear, and promised handouts never arrive.

Meanwhile, the warning about the buying and selling process: automatic assistants will end up managing emails, CRMs, and negotiations. Humans will only approve. And the third derivative: being opaque on social media will make you less vulnerable to strategies managed by AI.

The most honest summary of the thread is signed by someone who expects nothing anymore: "In five years, you’ll have to wait five years."