You mean the evils of OpenAI in particular and the ongoing large-scale LLM hype in general?

Well, data centers outputting exorbitant amounts of heat is one part, yes. That's the most literal reading of "burning our planet".

Communities suffering water shortage because a data center diverts half the municipality's water to quench that heat is another.

A more poetic reading is the future they sabotage:

Companies are fucking over workers, firing existing ones and no longer hiring new ones for a short-term profit, causing long-term damage to the job market. By the time they figure out that the people they lay off can't actually be replaced by bots, there are a bunch of people looking for job: the laid-off workers that couldn't find a new job because other companies likewise stopped hiring, as well as the new graduates that are looking to enter the job market.

Between that increased supply of workers and the financial damage companies suffer from that mistake, wages may well get pushed down, probably without the cost of living following suit. This means more financial pressure, a greater power imbalance between employers and employees and overall a shift in power towards the rich that don't suffer to nearly the same extent.

Speaking of the rich, there is one particular group of people that will profit from this: those selling AI.

People who have become reliant on their tools will be reluctant to return to working without them, or even find it difficult to do things they haven't had to do on their own in a long while. So a lot of them will pay. Businesses in particular will be weighing the cost of dropping those tools versus paying for licenses. There is also a human factor of admitting you were wrong and throwing good money after bad one (sunk cost fallacy, if you've heard the term).

All of this means they'll have a good angle to start squeezing for money. Add on the usual bullshit of hidden fees and predatory pricing models and you've probably got people paying out the nose. They can and most likely will sell data gathered from the interactions people have with their tool.

By the time the mistake becomes evident, they'll have made off with the money everyone else has been giving them to fund their own self-destruction.

OpenAI specifically is aiming to go public, making their Initial Public Offering this year, meaning they will eventually be obliged to start generating profits for the shareholders. They've worked hard to get people hooked on ChatGPT and its siblings, to convince everyone that AI (specifically: LLMs) are the future, done nothing to remedy the misconceptions about the actual nature of LLMs, still haven't installed functional safeguards against people asking their text generator for advice on how to kill themselves or others. They were the vanguard spearheading this hype with ChatGPT and they'll keep pushing it for as long as there's cash to squeeze from rubes.

I'm not even starting on the social and psychological cost. They're harder to pin down, but that doesn't mean they don't exist. But just the ecological and economical damage this company (among others) continues to cause is immense.

In that context, their finding a solution to a mathematical problem is a faint achievement, like a killer helping an old lady across the street on the way to continue their slaughter on the other side: nice in isolation, but not enough to weigh up their evils.

(And if I may be particularly cynical, if this is used to show that AI can render yet another job obsolete, it would be closer to pretending to help, then pushing her over.)