AI to help planes avoid climate-warming 'sky graffiti' | Hacker News Reader

AI to help planes avoid climate-warming 'sky graffiti'

bbc.com

Yeah, let's use a computationally heavy task that contributes to burning coal and gas releasing CO2 into the atmosphere to reduce a slight fraction in emission so we could pat ourselves on the back.

It is not a “slight” reduction in emissions, it’s absolutely huge! About half of the aviation industry’s total impact on global warming is due to contrails! This is 1-2% of all human caused warming across all industries.

(Though slight caveat for any nit-pickers. The warming isn’t related to emissions. The contrails essentially create a blanket that traps heat in the earth’s atmosphere - the BBC article linked mentions this)

Source: https://youtu.be/QoOVqQ5sa08

(Or check the sources in the linked video)

If you shift a route to avoid making contrails and end up on a longer or more turbulent route what is the net benefit to the climate? For example if a plane ends up using more fuel on a longer route to avoid a cold area where contrails could be made or ends up on a route with more headwinds or turbulence.

Is this a bit of green washing?

From TFA:

> Rather than dramatically changing the route of a flight, the trial will mostly involve changing an aircraft's altitude by around 2,000ft.

What if there are more headwinds at that height which require increased fuel use?

If you have any experience with the majority of airlines, you can be sure they are not planning to spend any significant amount of extra fuel on this problem. I am not a pilot, but I assume once you are at cruising altitude, small altitude changes are a rounding error

I mean what if there’s a mountain at that height? Will they just fly into it?

It's an optimization problem which could be solved if impact of contrails could be estimated. Making a long detour unlikely is good choice but if a small route / altitude changes could help - why not?

That’s not deep insight stuff. I promise you these people are more keenly aware of what constitutes an undesirable solution than that.

Or we could fly less.

This exactly.

Yes.

I read somewhere that 90% of human never took a plane.

Ok, say we somehow cut air travel by 50% overnight. Seems unlikely, but we did it.

Would you advocate to not pursue this approach to mitigating the remaining 50%?

Check out other comment. It could actually end up causing more warming in the long term depending on the net change in fuel use or extra maintenance on plane etc.

The required changes are typically minimal. A little bit of altitude up or down, a minimal detour. Also mentioned in other comments (and in the article).

Isn’t that the purpose of running a study? To establish that a theory works and could be beneficial at scale?

This is not what I was envisioning for major AI breakthroughs

It can hack into companies and exploit anything, but beyond that, forget the medical breakthroughs and real climate solutions, we’re tackling “sky graffiti”. Here I was, an ignorant dolt, unawares that we even had a major “sky graffiti” problem.

Reply on news.ycombinator.com