Growth is over globally except India and Africa due to demographics [1]. Developed countries and unions will compete for the last of the world’s young, prime age workers over the next century. Financial stability will be a function of who manages this situation to the best of their circumstances. Can you attract and retain these workers and leverage that for economic success? If so, that’s where investment will flow and capital will remain invested. My thesis is Europe is best positioned to succeed in the near term, based on the above.
Sibling comment mentions Europe needs the US for energy; Canada has known fossil gas reserves of ~200 years and LNG export capabilities, and Europe is scheduled to end Russian fossil gas consumption in 2027 [2]. The world is deploying 1GW of solar every 15 hours; like the rest of the world, everyone will arrive at energy independence/sovereignty eventually through cheap renewables (solar primarily, but also wind) and battery storage (LFP and sodium most likely, as of this comment) exported by China to the world. China is also selling inexpensive EVs to as many global consumers as they can find (while internal sales of battery electric and hybrids is already at ~50% this year). This leads me to believe the future of US oil and gas is an internal petrostate similar to Russia, not an energy exporter of relevance far into the future.
Having billions of low-skill "prime age workers" is not a good thing in the AI era. Africa's population surge is a liability as much as it is a benefit, especially considering they already have to import a large portion of their food supplies. Africa is 20% of world population but just 2% of world GDP... Maybe they will turn into China at some point, but I wouldn't bet on it yet.
> Africa's population surge is a liability as much as it is a benefit, especially considering they already have to import a large portion of their food supplies.
They also need to import phones, dishwashers, and cars. But how many cars do you really use? How many phones?
Africa's population growth is projected to drive steady demand for consumer products, while the West and parts of Asia are expected to see a decline.
They have abundant natural resources, and they’re likely to become the cheapest labor force after countries like Vietnam and Thailand see rising wages and living standards.
We’re already seeing this in China, where wage growth is pushing some manufacturing to neighboring, lower cost countries.
Demand is ultimately limited by the number of people. You can produce as many goods as you want using AI powered factories but without demand, they’ll just end up in the garbage.
AI era remains to be proven to not be bullshit. If it turns on to be something of value and not pets.com and webvan 2.0, prime age workers should be provided training and hiring pipelines for jobs AI cannot do: healthcare, construction, infrastructure, housing, agriculture scaling, etc. Knowledge/white collar jobs are most at risk with LLMs, not work building society up and operating it in the physical world.
If you think an LLM is a thinking machine, I don’t know what to tell you. They are great at predicting tokens, but there is no evidence they are thinking in the human sense. They rely on patterns and probabilities, not understanding. Powerful search engines, certainly, but there is much work ahead to see how hard the limits being bumped up against are.
> As someone that has sold a bunch of LLM enabled software over the past 6 months, I don’t really buy the AI capex turning into huge productivity gains. Everything to date are just chatbots with RAG and API calls. None of them are going to do my laundry or file my taxes.
Without getting hung up on the definition of the word "think", the fact that anyone can feed an LLM a couple of bullet points and generate a pile of convincing sounding marketing copy, is worth something if you're to someone. We can disagree on exactly how much and to whom, but if you're not willing to see that as fact then there's no common ground to base a conversation on.
It's under appreciated that some jobs being trivially done by LLMs doesn't mean that all jobs are trivially done by LLMs.
Low-value-add marketing copy, for example, being automated changes... what? Now LLM-generated copy is the new minimal-cost baseline, and everyone adapts to the new normal.
The real killer feature will be autonomous business planning, but we're a long way from there.
Sibling comment mentions Europe needs the US for energy; Canada has known fossil gas reserves of ~200 years and LNG export capabilities, and Europe is scheduled to end Russian fossil gas consumption in 2027 [2]. The world is deploying 1GW of solar every 15 hours; like the rest of the world, everyone will arrive at energy independence/sovereignty eventually through cheap renewables (solar primarily, but also wind) and battery storage (LFP and sodium most likely, as of this comment) exported by China to the world. China is also selling inexpensive EVs to as many global consumers as they can find (while internal sales of battery electric and hybrids is already at ~50% this year). This leads me to believe the future of US oil and gas is an internal petrostate similar to Russia, not an energy exporter of relevance far into the future.
[1] https://www.sas.upenn.edu/~jesusfv/Slides_London.pdf
[2] https://news.ycombinator.com/item?id=43506589
(not investing advice)