Energy storage. We could already produce all the power we need and more using solar power, the problem is that we can’t store it in an efficient energy dense form. The word efficient there is doing some heavy lifting. It needs be comparable or better than our methods today in terms of cost, safety, energy density, climate impact. If we could solve energy storage, it’d change society and technology dramatically.
Cultured Meat. Without relying on any major breakthroughs, a price competitive with “traditional” meat is feasible with a few rather reasonable and conservative assumptions and developments. Dropping cows as meat source globally alone might be sufficient to slow down further climate change significantly.
The Internet Archive - long term storage of as many books/films/music/journals/games as possible.
Love this answer! I wish the EU could move it in as an EU agency.
Nuclear fusion, right? That’s got to be the big one.
How do we stop it killing the global economy over night ?
Why would we want to do that? I hate the global economy.
The obvious point is the global economy that you hate will kill any free and limitless fuel before it is born .
I’m sure you’ll get a bunch of respectable answers, but keep in mind that secretly, everyone is actually thinking sex robots.
Which is an application that generative AI is important for, ironically.
And if they were not, they should have been.
Climate change
I would specifically pick Fusion research.
Kind of a two-fer right there. Without “AI” sucking up so much power, we’d already be better off climate-wise.
True, but still only marginally.
I think concrete is still king.
Trains.
I like trains!
Tech Debt. So many of these companies investing in AI are doing so at the expense of out dated, broken shit that AI will never address.
I like the answer, but that’s not a single endeavor.
I can’t speak for other companies, but in mine we do have a single line item for Tech Debt. It gets IGNORED most of the time, but we do have it! ;)
What would be the first 3 tasks to be done under that line item?
Largely software updates. You would not believe how old some of this stuff is.
The problem becomes each update is tied to multiple systems, so testing has to be done across the board for each update.
So it’s easy to go “Why are we running xxx from 2016?” but then you have test it everywhere and that’s why it’s almost 10 years out of date.