Very good question! These models are trained on ~40 years of ERA5 data –you can think of past forecasts from numerical models integrated with real observational data to have a continuous distribution of weather parameters (temperature, wind etc.) Therefore model resolution is 0.25 degrees (28km x 28km at the equator).
The way accuracy is measured is through picking targets (say temperature at 2 meters, at x,y lat & lon and forecasted 24h ago) and comparing them on RMSE and ACC (anomaly correlation co-efficient). For instance, in Google Graphcast paper they pick 1380 targets and the model out performs NWPs in 90% of them.
OK so feels like this needs some explanation. It is in an entirely unrelated field, just like Palantir (also directly from LOTR). Not to distract the discussion, but Skyrim trademark is registered at USPTO in the field of computer game related anything.
Old school advice but worked really well in my previous startup –(i) conferences and (ii) snow-ball interviewing.
(i) In your case you need to somehow find this high density potential early user in-person gatherings. You can for instance hang out in a non-generic co-working space where you know it is the choice for digital nomads etc. in your city.
(ii) This is somewhat under-rated but you really can crawl the network of people you know ad infinitum. You only need few people you know who are productivity obsessed to start with. And then always ask them "who should I chat for 15 mins. to get feedback on this etc. (I would suggest crafting this ask really well so that it doesn't come off as demanding. Try to genuinely give value at every interaction, that's how I enjoyed the process as a hardcore introvert.). I spoke to about 5-10 people every week and ended up with a decent community of early users in <3 months.