An online interactive field guide - where the printed book can only say - also sp A with smaller leaves, and sp B on limestone, and sp C gone extinct (let us know if you find one!)
Printed is also a snapshot of taxonomy and knowledge at the time of editing. Immediately out of date in places.
Very nice guide!
I haven’t seen one done as a forum post.
A lot of published journal posts have links to reference materials and notes, so I started one myself.
I don’t recommend anyone else use the forum - mine was an experiment and showed the forum is ill-suited to the task, being optimized for editing much less complicated posts, and it’s a weird place for it. That said, mine turned out well enough despite the annoyances, and if finding a place to get started is what’s stopping you (as it was for me) then just do it - the content can easily be copied elsewhere later, and it’s much better to have guides right now than at some nebulous time in the future.
One good thing about the forum: a forum topic can be edited by others, if you check the appropriate checkbox. Mine has only had two other people make edits so far, but each edit has been an improvement, especially @chuuuuung’s addition of a couple of species.
So I do want to point something out to everything that I don’t think anyone expressed (I apologize if I missed it.)
Anyone can edit Wikipedia, and the about pages for taxa pull from Wikipedia.
There are many taxa pages that are functionally nonexistant; I doubt there would be much objection if people started going in and writing descriptions for taxa that are underserved, especially if things are properly cited (find the description paper and translate it to layman’s speak, more or less.)
https://www.inaturalist.org/taxa/418443-Panaeolus-cinctulus here is good example where the text of the Wikipedia page has clearly been written with identification in mind. Here is another one https://www.inaturalist.org/taxa/206899-Psilocybe-caerulipes
Yes, all of my examples are active mushrooms, but the community for that group of species is quite a bit larger and less overworked than the rest of us so they have a bit more time to go in and do wikiedia edits.
https://www.inaturalist.org/taxa/1550317-Macrolepiota-macilenta https://www.inaturalist.org/taxa/1550316-Macrolepiota-pallida But here are two I may go in and write up a brief on, if I can find time in my schedule. The paper is open source and came out recently, so it is accessible, and it would be trivially easy to get the information rewritten in a way to make it wiki-appropriate https://fuse-journal.org/images/Issues/Vol14Art9.pdf
In fact, I’m going to make a note to myself to remember to do this….
I just had a discussion about this with a bunch of entomologists yesterday. One thing I learned while being on iNaturalist is, that when you are out there taking picture, the most important thing (besides quality pictures) is to know what parts of an animal or plant to take detailed photos of (if you can).
I have a basic idea for a lot of groups by now, so I have a routine of which picture I take. But occasionally some subgroups/species require something special and I might just learn about it after being back home.
My vision would be that when you are out taking pictures. The app will recognize what you might be looking at (as it already does in the new version) and then give you hints about which additional perspectives might be needed for identification.
For example: “Seems like a Bombus species. For identification detailed photos of the face and the striped body from the top are needed”
One issue with this when the required view is something like ventral (belly) view that can only be photographed with the organism in hand. For vertebrates, this could create…problems.
This is true. However, there was this recent topic: Can species identification tips be added to Wikipedia?
While there was a lot of support from us iNatters to include identification guides, there appears to be opposition to such “how to” guides within the Wikipedia community.
That’s fair.
Though broadly, what I’m suggesting is just updating the species pages so they have details about an organism.
If this was done, keys and ID guides could link to the wikipedia page, but not necessarily be on the wiki.
As nice as it is to see who supports what, the question remains: what about that grant from Google? As I understand it, the reason the GenAI thing is still being argued about is because iNaturalist has received a grant to build it. Not to build a wiki.
Every thread about this becomes contentious. Until this thread, the “News and Updates” forum almost never came up to the top of the list; the news and updates just didn’t garner much discussion. Now “News and Updates” looks like one of the most active forums, not beause there are suddenly more interesting news and updates generally, but because there are multiple new posts a day in this one thread. What resolution do we realistically expect, given the backstory as revealed in several earlier threads?
Definetly agree. There will be cases where recommending certain actions would go against the idea of iNaturalist oftrying to not disturb the animals too much or which could be dangerous. For example I wouldn’t recommend someone stumbling across a bat to check its teeth to distinguish them.
From the example provided in the original post, these cases could probably be address by the community by downvoting or flagging recommendations.
Will the ‘‘Exploring New Ways to Learn from iNaturalist’s Community Expertise’’ thing be available for apps in a way or just web browser?
Many reticences are in line with the (justified) concern of information being summarised or compiled inaccurately by the LLM. But I think many of you should give us (non-experts willing to learn) with a bit of benefit of doubt. We know that the information is to be taken with a grain of salt (and a disclaimer wouldn’t hurt). In the same way that I know that the CV proposed by iNaturalist on each observation is also not the same than an expert option. But it’s a hint to something that might lead us to the right direction and sometimes to the wrong one. And leading to the wrong one is not bad, as it gives you a genus to compare and realise that the CV was wrong and which resources to check with. Most of my learnings about Nature started being wrong and that is not an issue, it’s a process.
This could be a great tool to enhance learning, not the bible of ID, or a scientific backing, but nonetheless a first step ahead into generating the next generation of IDers.
As we wrote in the blog post, it will be like these existing demos: https://www.inaturalist.org/vision_language_demo and https://www.inaturalist.org/computer_vision_demo. So web only.
Will the demo include an example (or multiple) of how this new feature deals with a taxon that usually cannot be IDed?
The llm discouraging overconfident IDs was the aspect of this I am most hopeful for, so I’d be curious to see it.
I use a text expander to drop ID tips into comments so I have many copies of identical tips. I would expect the llm to grab one and ignore the rest as duplicates. If it were to grab them all…
That’s an amazing idea, I’d like to second this as a request. Probably should add it to the Feature Requests category for consideration to add to Seek or iNat Next, etc. This would really add a lot of value for not much effort, especially for new and infrequent users.
Photography tips and on-boarding of some type are sorely needed, and a feature that provides real-time tips on what to photograph would go a long way towards that goal.
Hi everyone, thanks again for your helpful feedback and questions about our forthcoming ID Summaries demo. Your input has been extremely valuable in shaping this work.
We’ll be launching the demo for everyone to explore on Wednesday, December 3. That day, we’re hosting a webinar — please register and join us if you’re available! We’ll share how the demo works and hear directly from some of the experienced identifiers whose knowledge it draws from. This demo is just one approach we’re exploring to highlight our community’s amazing identification expertise, and we’re excited to keep working toward building scalable tools to help people learn about nature from one another.
We’ll share the webinar recording afterward, both via email with all registrants and on the iNat blog. Stay tuned for more updates on Wednesday!
Thanks, but I’d like more information on whether this will change as it is expanded.
Just finished the webinar, live. This tool looks amazing!
My favourite moment was when @nathantaylor basically said (my phrasing),
“Imagine if you were hosting a Masterclass (‘Michelangelo teaches painting’, or whatever). Now imagine that everyone in the world was interested in your taxon. Would your comment be more helpful? Would it be more clear? Because that is reality now, thanks to ai tools. Instead of giving a man a fish (instead of doing mass IDs in your taxon), you can teach the whole world to fish!”
I also know that there were MANY comments that didn’t make it into the shared Q&A chat. I personally had at least three.
Please share your comments here!