Comment by olivierlacan | Hacker News Reader

Comment by olivierlacan | Hacker News Reader

Parent Full thread

Sure. But this same argument goes for every open source project ever built. At some point some convergence is in our best interest one some things, or we all burn tokens recreating the same thing until the planet is mostly data centers.

You'll note I'm not reinventing things, most of this project is connecting dots between open science and already amazing projects like iNaturalist. But if you're familiar with it, you might also be frustrated with its extremely simplistic approach to identifying natives.

The reason I suspect is why your last point exists, except I've spent an inordinate amount of time and effort ensuring that Indigene doesn't claim a plant is native to a region (vs. a whole continent) when it's not. That makes it less comprehensive than other resources, but unless you can point to specific issues it doesn't make it less reliable.

I can't make you care, but if you might this is where I'd look if I were skeptical:

https://indigene.app/sources

https://indigene.app/confidence

https://github.com/olivierlacan/indigene/blob/main/DATA_SOUR...

https://indigene.app/plants/penstemon-digitalis (example of a status conflict)

https://github.com/olivierlacan/indigene/blob/main/app/scrip... (which is what does this status checek)

And if you care to call me out on any GIS mistakes (given my own lack of expertise in that area) this where you're likely to find mistakes: https://github.com/olivierlacan/indigene/blob/main/app/scrip...

First, really cool project and love what you're doing.

I'm working on something VERY similar, but focused on a particular country, so I'm in the weeds (hah) on this and here's my feedback:

* I absolutely do not trust the data provenance here, and reading through all your links doubly so. You cannot trust claude to do analysis. Not only does it misattribute often, but fake sources and poor interpretations seep in. Without url sources tied to each claim it's inevitable that it rots. You'll have no way to keep track of claims across plants, and over time things slip.

I did see your data structs in github, but a lot of it is just LLM summaries of what it read. This isn't bad (and at some point necessary), but Claude in particularly is notorious on tasks like this, that don't pass human evaluations.

HOWEVER, that said, it's probably okay because it's not a life or death subject. Claims like "full sun" aren't very usable as is, and these can just serve as pointers for people to go research plants on their own.

----

Excited to see where you get to!

Thank you for the feedback, I'd love to see your thing and what angle you're focusing on. I see what you mean, but I've never trusted Claude to do analysis. And I'm still only trusting it to help me build acceptance criteria that I can validate against a multitude of region-specific sources for different data points:

- host plants (e.g. European caterpillar counts via Gaytán et al: https://onlinelibrary.wiley.com/doi/10.1002/ece3.73004)

- food sources

- pollen interactions

- broadly documented wildlife & plant dependencies

- etc

What is your skepticism based on? If you've identified a gap I'd be very obviously interested to close it. But if you think I'm allowing Claude to editorialize without verification, then I can assure you I'm not. I've definitely caught it trying to skip steps and "leave out" verification steps, especially for new region coverage where we have insufficient data for a clear cut boolean answer.

And like everything, I keep expanding the rings of what I cross-reference for each region. Hopefully the more keen eyes on this project will mean folks can spot fallacies and holes. But to be frank, I'm in this for the long run so I very mulch welcome it :-)

PS: https://github.com/olivierlacan/indigene/blob/main/DATA_SOUR... might be a good one to focus on if you hadn't picked it up yet

Reply on news.ycombinator.com