Forum Debate Suggests AGI Arrives This Year

AGI Arrival This Year: The Job End Debate Dividing Experts

General Artificial Intelligence (AGI) is imminent. Or not. It depends on who you ask. While some argue that OpenAI's latest agents have accelerated the timeline, making superintelligence inevitable within months, others point out that current models are performing worse than previous versions. This discussion is far from academic; it has immediate consequences: if AGI arrives this year, millions of office jobs disappear. If it doesn't, the hype will have served to justify layoffs and restructurings already underway.

The Announcement That Changed Everything

The starting point is a prediction circulating for months: OpenAI's agents moved from reasoners to agents in just four months. These agents perform automated tasks, causing an exponential increase in research. According to this thesis, rumors suggest they have reached Level 4, AI that learns and researches. This researcher AI would arrive in two or three months, allowing the model to read thousands of research papers, apply that knowledge, and self-improve. Consequently, Level 5 would be reached in weeks. From there, everything would explode: AI would improve exponentially, reaching AGI in days and superintelligence in hours.

This scenario, presented with almost religious enthusiasm, has its flip side. An article in Nature argues that we have already achieved AGI, stating that current LLMs pass the Turing Test easily and function as experts in many domains. We lack only a superhuman intelligence that masters everything, but that is not necessarily the definition of AGI. Many arguments against LLM intelligence no longer hold: they do not merely repeat what they learned but reason, handle physical concepts, process various data types, and solve new problems. The lack of body, own goals, or permanent memory does not automatically disqualify them.

Why Do Some See Imminent AGI While Others Place It 20 Years Away?

The discrepancy is significant. Some claim Gemini 3 is superior to most humans and that AGI was achieved in September 2024 with reasoners. At the opposite extreme, a more skeptical analysis suggests functional AI may arrive, but surpassing human cognitive levels is doubtful until quantum computers exist. They can do many things quickly and outperform 90% of the population, but the smartest 10%, those who provide ingenious ideas, need capabilities beyond reading and pointing at data. They conclude this group needs another 20 years.

Between these extremes lies an uncomfortable observation: users testing the latest GPT and Claude versions notice worse results than previous ones. "So far, GPT-5 gives me worse results than its previous version, and I also notice Claude weakening. I don't know what they did, but for me, they went backwards," notes one comment. Primary use cases include programming and research. If models worsen in key tasks, the prediction of exponential improvement wobbles.

Impact on Employment: From Office to Field

The practical consequence concerns employment. Most office jobs, those 'warm chairs,' are doomed, summarizes one intervention. He hopes robots won't appear in ten years to do manual work too. Those worst off are those claiming their job security is assured. The word 'job' appears repeatedly: before the year ends, working won't be necessary.

The impact isn't uniform. Some physical-creative jobs won't be replaced, points out someone. Everyone back to the fields. Another adds that in Spain, implementation will take years: tell a bricklayer who started a small business because he had a cousin in an Ibex company about AGI and AI. Such a worker distrusts all of it. The gap between technology and emotion is brutal, says a third. I don't doubt tech readiness, but human assimilation takes time.

Warning From Those Seeing It In Motion

Not all is speculation. Concrete data exists. Spotify revealed its best engineers haven't written a line of code since December thanks to an internal AI system called Honk, based on Claude. The company launched over 50 new antiestéticatures in 2025 alone, with AI enabling real-time bug fixes and deployments. This anecdote, spread on social media, exemplifies how cognitive task replacement is happening in leading firms.

Meanwhile, AI integration advances via other paths: CPU farms for parallel processes, simplification, shorter times, personalized products. AGI isn't arriving this year; specialized agents are, costing like workers but operating 24/7. Theory says companies replacing staff gain competitiveness. Practice remains to be seen.

---

The exact point where analysis stalls is the same as the beginning: nobody knows if AGI arrives in months or decades. Proponents rely on recent estimulante ilegal. Skeptics cite worsening models and argue general intelligence requires more than data processing. Meanwhile, companies reorganize staff using AI as an excuse, regardless of AGI arrival. And IT vocational students look to the horizon with curiosity and vertigo.