Article by Tim Treadgold, courtesy of Business News
11.08.2025
There appears to be a connection between failed hydrogen power projects and artificial intelligence, and it should serve as a warning to all Australian governments and businesses.
At its simplest, the lesson of hydrogen is ‘beware yes-men’. Fortescue, Woodside Energy, and BP are the latest companies to admit that hydrogen is a more difficult energy source to tame than management acknowledged when committing millions of dollars to bold, but failed, attempts to commercialise the gas.
When entering the hydrogen race, the three big companies were cheered on by environmentalists and governments which, in some cases, helped with funding. Some people warned that it was a pointless exercise, but they were drowned out by yes-men.
Recrimination time has arrived with governments asking for their money back; a demand some shareholders might also consider because hydrogen was always a lost cause for industrial-scale power production using known technology.
What happened with hydrogen can be explained by a belief that radical, unproved, solutions are needed to slow climate change.
Amusingly, many of today’s hydrogen advocates were yesterday’s opponents of nuclear power, only to now acknowledge that nuclear is part of the solution not part of the problem.
The physics of hydrogen explains why it cannot work as promised, as most students who paid attention in science class are aware, given hydrogen lacks ‘volumetric’ energy density.
In other words, hydrogen contains a lot of energy but takes up a lot of space, which makes it hard to store and transport.
Like the missed opportunity with nuclear, the false appeal of hydrogen led to a vast loss of money by everyone who believed they could master a gas that has always been a challenge.
Why hydrogen gained such a strong following is a question worth close consideration, because the same psychological force that drove hydrogen investments appears to be at work in AI software, which produces a similar confirmation bias.
Gautam Mukunda, a Harvard Business School research fellow and Bloomberg columnist, warned last month that “AI is the ultimate yes-man” because it often tells people what they want to hear rather than what’s correct.
The reason for AI behaving like that is because many of its models are rooted in “reinforcement learning from human feedback”, Mr Mukunda says, a description once known as GIGO (garbage in, garbage out).
In other words, some AI programs are trained to please the questioner with flattering or agreeable answers. They are, in effect, sycophantic.
Roll back time to the early years of the now-failed hydrogen revolution and the same situation arose, with a cheer squad of hydrogen enthusiasts demanding that companies and government invest in their favourite fuel.
Hydrogen believers won the ear of government while sceptics were dismissed as ignorant.
Even Tesla founder and serial technology developer Elon Musk was ignored when he dismissed “hydrogen cars as mind-bogglingly stupid”; a description that can now be extended to the wider hydrogen economy (until someone conquers the gas).
More importantly, the admissions by Fortescue, Woodside, and BP that hydrogen is not yet a commercially viable energy source does not seem to have made an impression on the Australian government, which is determined to press on with hydrogen.
Perhaps no-one in government understands the meaning of a statement associated with Albert Einstein, that insanity is doing the same thing over and over and expecting different results.
Profit centre
ON the issue of lost causes, there is something naive about calls for Rio Tinto to shift its head office from London to Perth because it earns most of its profits here and now has a Western Australian, Simon Trott, as its chief executive.
The reason Rio Tinto will not shift head office is as simple as asking who owns the company rather than where it works, because most shareholders are British or live in Europe or the US.
If the place a company earns most of its money is considered the test for head office location, then Wesfarmers should shift to Sydney or Melbourne where its biggest business units – Bunnings, Kmart, Officeworks, and Wesfarmers Health – earn most of their profits.
A list to miss
FIVE WA businesses have been named among the top 10 on a recent ASX list.
Unfortunately, the list that features Paladin Energy, Boss Energy, Pilbara, Mineral Resources and Liontown Resources contains the most heavily short-sold companies, with investors believing their share prices are more likely to fall than to rise.