Bill Gates - can we still mention him? - has previously said, “We always overestimate the change that will occur in the next two years and underestimate the change that will occur in the next 10.” Well, here’s my strong forecasts for the next 10 years:

  • Donald Trump will not be US president;
  • The world will be hotter by 0.2 - 05°C, and will probably have passed 1.5°C of heating;
  • Humanoid robots will be much more common, and will be becoming commoditised;
  • There will be a lot more pressure on the global cement industry to decarbonise;
  • There will have been a major recession, partly caused by the implosion of the AI bubble, with knock-on effects on equities, crypto, gold, bonds and derivatives (Warren Buffet called derivatives ‘financial weapons of mass destruction’ - let’s see);
  • Energy prices will be lower, led by development of renewables, and lower than fossil fuels;
  • Battery technology will be significantly better than now;
  • Fusion power will still be 10 years away (but, this time, really 10 years);
  • People’s lifestyle expectations will have increased: It will be ‘more expensive’ to have kids;
  • China’s population will continue to fall, faster and faster, while populations in Japan, Greece and Russia are also set to fall;
  • By the mid-2030s, those aged 80 and over will globally outnumber infants;
  • The Sun will rise and set, and governments will still demand your taxes.

Just as the dot com bubble didn’t end the internet, implosion of the AI bubble will not end AI. There will be a Darwinian winnowing of companies, and only the strongest will survive. Again, as Warren Buffet said, “Only when the tide goes out do you discover who’s been swimming naked.” AI development will continue, just with more realistic company valuations, (and certainly slower building of data centres and so less cement demand for those projects).

In January 2026, Gates suggested that AI improvements are on an exponential trend, and he suggested that by 2036 there will be an Era of ‘Free Intelligence’: top-tier expertise - such as great medical advice and personalised tutoring - will become ‘free and commonplace’ through AI, removing bottlenecks in healthcare and education.

Gates predicted that AI will automate manufacturing, agriculture, and routine office tasks to such an extent that humans may only need to work two or three days per week. To which I add the rejoinder - “for 100% of the money?” That’s not going to work, is it?

He suggested that humans will no longer be needed for ‘most things,’ as machines take over physical labour and specialised knowledge work. In the cement industry, if you have 200 workers in a 1M/yr cement plant now, then perhaps in 10 years’ time, with fully-integrated AI, you will only need 100 - or 50. Those workers will all be using AI - having helped to train it. Retirees will not be replaced - by humans at least.

Given that most governments make large parts of their tax revenues through taxes on working income, then they are going to have to start taxing AIs and robots - possibly by putting up the taxes on company profits - in order to make up the shortfall (same with taxes on EVs versus fuel taxes). AI will have an effect on local taxes on cement plants - which currently help to sustain local communities.

AI or AIs? Will there be one ‘general’ AI that looks at every aspect of a cement plant and decides on everything, or multiple ‘narrow’ AIs that examine specific problems, and pass that ‘upstairs.’ Probably both - like an octopus with a brain in each leg, giving each leg some independence, but sending signals up to a central AI brain to give overall control. This is already developing, with local AIs already active in cement plants for mill control, predictive maintenance, diagnostics, strength forecasting, and process control. A system for bringing them all together (in a cement plant) is analogous to making a general AI from the various AI talents that already exist (in the tech world).

The question perhaps should be - are there professions in a cement plant (and the wider cement industry) that will not be taken over by AI? Aspects of every job will be taken over soon, if they haven’t been already. If you think that your job could be done by an AI in the next 10 years, then it might be time to develop your professional networks, be open to change and innovation, be open to working with AI (and robotics), and be willing to retrain if required. A recent report1 suggests that the professions that are least likely to be affected by AI are dredge operators, bridge and lock tenders, and water treatment plant operators...

1 https://fortune.com/article/what-are-the-jobs-most-exposed-to-ai-microsoft-research/