Ben Matthews

I find this analysis is a useful starting point for discussion, although there are plenty of details one might adjust.
Personally I’m using (inter alia) claude to help me refine an interactive climate model (example here - although that’s last year’s version pre-ai-help ). So I care about these things.

As my own life also has an energy cost - even just sitting at a desk with computers and some heat light and food. I reckoned by my own crude calculations that my ‘human’ energy cost per hour was considerably higher than that of my AI assistant, which certainly helps me progress faster, so the net effect was less energy per ‘task done’, meanwhile we don’t have infinite time to solve such problems. I’m only using claude within the limits of a pro subscription, and achieve that with tough claude_md instructions - not to go digging rabbit holes without consulting me. Sometimes it analyses and fixes autonomously and efficiently, but you have to keep alert - sometimes I interrupt and say no there’s a simpler way, and draft better algorithms / structures myself. Also I use scala whose sophisticated (non-ai) tooling constrains mistakes and its mcp/lsp makes searching and refactoring across a large codebase much more efficient than claude’s normal grok by subagents. Combine tools carefully, not brute force.

Evidently a big unclarified issue is the energy cost of training these things. But we don’t need so much more training - for my purposes they are already good enough. The frequent new releases are about scary headlines to pump the IPOs. If this race could slow down, we could just learn to use what we’ve got more efficiently. In the general public discussion, I’d also appreciate clarification about how much of ‘AI’ energy-use is going into creating images and videos, rather than text and code, my hunch is it’s much worse for videos most of which are about trivial stuff. Also loads of datacenter energy is wasted transmitting talking-head videos around the world - that’s really inefficient. So well designed code, part-aided by ai, might help find more efficient ways to run needed global dialogue.

Indeed - why so low in most of Poland, compared to neighbouring regions with similar climate/crops, also I wonder if tibetan plateau is really so great for Fungi? Or are these artefacts from the machine-learning algorithm?

“…at a rate of roughly 0.05 percent per day … would take a very long time” … but by my quick calculation 0.9995^3650 is 84% per decade, which is not long. Almost instantaneous on a geological timescale - and think how much the world changed when fungi learned how to digest lignin in wood - ending the era of coal-forming swamps.

Sure, but this is also a real game we need to win (well, maybe not <1C in that timeframe) , and we only get one chance to play. This example helps people learn, but there are things to adjust.
Another (I didn’t mention above) is that construction (including new energy, ‘green’ cities etc.) takes massive time, energy, materials - it’s not clear that’s sufficiently taken into account, and likewise not by real “socialist” planners.

You can get those ‘accelerationists’ within the coalition by funding lots of research, just don’t expect it all to work, don’t even need to apply it. Actually I think that ‘bias’ is realistic. Problem is rather political groups that are missing - religious for example.

OK, so I tried this, able to win on the second round. :-)
First time you risk to do some things too early, others you must do early, but I won’t spoil the challenge by giving details.
Good emphasis on land-use limitations.

Concept is nicer than ‘fate of the world’ which was rather similar (and even fotw told me their idea was partly inspired by an idea on my website about 23 years ago). Both this and fotw based on ‘cards’, while prefer to adjust levers gradually, and see graphs move in real-time.
(btw going back even further, does anybody remember ‘lincity’ )?

Some things confusing - e.g. you adjust percentages not totals, but totals change, which hits limits in not-obvious ways. No mention of space-heating challenge eg heat-pumps (suggests made in tropics?), no modal-shift in transport (except inside cities). I’d like to see whether the numbers reflect current emissions of China, and Arabia (I doubt it, doesn’t fit the ‘south is good’ narrative). Overall I suspect that the calculations are too optimistic, but can’t say more without detailed plots of changes over time, or a view of the engine code.

But biggest unrealities:

• We don’t have such a scenario - there is no global planner - “god games” are too easy concept.
• The fraction of contrarians is larger (than the 3 groups I couldn’t satisfy in this game), maybe increasing (?).
• Some of the ‘mandates’ are far too easily implemented.

I ponder how to design a game which is more realistic in these respects.
Having said that, I think the ‘magic card’ has some merits, if everybody would play, maybe that helps tip the balance.