One thing I’d note is that Guesstimate took quite a bit of time (maybe around 1.5 years engineering/PM time.) I’d guess is that if you were to try to pay for someone to make something similar, it could cost quite a bit (>$200k.)
It may be possible to start with simple Python library with some visualizations or similar.
Do you think that Guesstimate has not yet made $200,000 worth of value for the world? I’m legitimately unsure about this point but my priors say its’ at least possible that it’s added that much value in time saved and better estimates. I think that systems modelling could have similar impacts.
I would like to think it’s worth it, and wasn’t suggesting that this would make the project not-worth-it, but just wanted to make it more clear what the costs would be.
Given that the costs are significant (though possibly justifiable), one easier approach would be with small solution to start.
Impact assessments. I think our ability to do impact assesments are bounded by our tools (for instance, they were on average much worse before guesstimate). If EAs started regularly modelling complex feedback loops because there was a readily available tool for it, I think the quality of thinking and estimates would go up by quite a bit.
A tool that makes systems modelling (with sinks, flows, and feedback and feedforward loop) as easy as Monte Carlo modelling was made with Guesstimate.
One thing I’d note is that Guesstimate took quite a bit of time (maybe around 1.5 years engineering/PM time.) I’d guess is that if you were to try to pay for someone to make something similar, it could cost quite a bit (>$200k.)
It may be possible to start with simple Python library with some visualizations or similar.
Do you think that Guesstimate has not yet made $200,000 worth of value for the world? I’m legitimately unsure about this point but my priors say its’ at least possible that it’s added that much value in time saved and better estimates. I think that systems modelling could have similar impacts.
I would like to think it’s worth it, and wasn’t suggesting that this would make the project not-worth-it, but just wanted to make it more clear what the costs would be.
Given that the costs are significant (though possibly justifiable), one easier approach would be with small solution to start.
What would you use it for?
Impact assessments. I think our ability to do impact assesments are bounded by our tools (for instance, they were on average much worse before guesstimate). If EAs started regularly modelling complex feedback loops because there was a readily available tool for it, I think the quality of thinking and estimates would go up by quite a bit.