China retired 12,000 degrees of the past. Brazil still issues them.

Between 2021 and 2025, Chinese universities revoked or suspended roughly twelve thousand two hundred undergraduate programmes and created more than ten thousand new ones in their place. The move touched over thirty per cent of the country's entire higher education offering. The cuts concentrated in fields the Chinese market came to treat as saturated or outdated, and the creations concentrated in areas tied to the so-called industries of the future, including new programmes in embodied intelligence, the frontier between AI and robotics. It was not a marginal adjustment. It was a state decision about which world the country will educate its next generation for.

In the same period, China took artificial intelligence into basic education. In 2025, the Ministry of Education published guidelines to structure AI teaching systematically across the whole of schooling, from primary to secondary, moving from local pilots to national implementation. Cities such as Beijing and Hangzhou made AI classes mandatory in public schools. The design is progressive. In primary school, the aim is to spark interest and build basic understanding. In middle school, technical principles and application. In secondary school, systems thinking and the capacity to innovate. The child does not learn to use a fashionable tool. They learn to reason in a world where the machine thinks alongside them.

One detail usually goes unnoticed, and it contradicts the caricature. That adoption was not done without brakes. The guidelines forbid primary students from freely using content-generation tools on their own, and prevent teachers from outsourcing the central act of teaching to AI. In other words, China did not put AI in schools on a wave of hype. It did so with rules, with limits and with pedagogical criteria. It adopted the future by governing the future. Hold on to that idea, because it applies far beyond education.

Now turn to Brazil. The most cited projections indicate that around sixteen million jobs may disappear in the country by 2030 because of automation, roughly fourteen per cent of the current workforce. And the signs of mismatch are visible well before that. Only forty per cent of technical course graduates in Brazil actually work in the field they trained for. Read that again. Most of the people the country trained for a specific occupation ended up somewhere else, which means training and the market point in different directions. We keep spending years and public and private money carefully preparing people for shrinking roles, while the roles that are growing go unstaffed.

The contrast is not about talent. The young Brazilian is no less capable than the young Chinese. It is about decision. One country looked at the labour market that is arriving and rewrote what it teaches to fit it, with the hardness of retiring degrees that no longer stand up. The other, to a large extent, still administers education like someone maintaining an inherited catalogue, updating slowly and by exception, training for a world that is leaving while the world that is arriving knocks unanswered. The difference between the two is not in the starting point. It is in the willingness to decide what to retire.

Intellectual honesty is needed here, and it does not weaken the argument, it strengthens it. The Chinese model has real and legitimate grounds for criticism. Cutting humanities, letters and arts programmes so aggressively carries the risk of over-instrumentalising education, of treating training only as economic input and of impoverishing a generation's critical repertoire. Human thought, judgement and sensibility are not luxuries in a world of machines, they are precisely what remains irreplaceable. Copying the Chinese gesture line by line would be a mistake. But ignoring the underlying lesson would be a bigger one. The lesson is not to cut humanities. It is to have the courage to ask, frankly, which world we are training people for, and to act on the answer instead of postponing the question for another decade.

And this is exactly where the story stops being about education policy and becomes about any organisation. The same question that separates China from Brazil separates the companies that will thrive from the ones that will wither. Training people for jobs that are leaving the market is the social version of a mistake companies make every day, when they insist on processes, tools and competencies built for a world that no longer exists. It is the same reflex as keeping reports that explain the past when the present already demands a decision. The past becomes a report. The future demands that someone decide what to retire today.

There is also the more direct parallel, that of governance. China did not adopt AI in schools carelessly, it adopted it with rules and with limits. That is precisely the thesis separating mature AI adoption from reckless adoption, inside or outside the classroom. Intelligence without governance is risk. Intelligence with governance is advantage. It holds for a country putting AI into children's education and it holds for a company putting AI at the heart of its operation. Adopting the future without control is opening a door you may not know how to close. Adopting it with governance is what turns novelty into an asset.

I will end on a note of optimism, because Brazil is not condemned to the role of spectator, and Mars exists precisely because we believe that. We are a Brazilian artificial intelligence company building frontier technology from here, with global ambition. If a country can rewrite in four years what it teaches an entire generation, a company can rewrite in far less time how it decides, how it operates and which future it prepares for. The question China answered with a state reform is the same one every Brazilian leader can answer inside their own organisation, starting tomorrow. Which world are you preparing for, what is leaving, and what do you have the courage to retire today. From complexity to clarity begins with that decision.