After almost 18 years, I left iNaturalist, the product and organization I helped create. I left because I don’t believe the current Leadership team is pointing the product in the right direction, and I don’t think they are managing their talented staff in an empathetic or effective way. If you’d like me to continue working on natural history software, support me on Patreon.
This post is an announcement for those who were unaware, an explanation for those who are confused, and a record so I don’t forget.
Some History
I wanted to build something like iNat shortly after I moved to the San Francisco Bay Area in 2003. In 2007 I attended the UC Berkeley School of Information and built it along with my fellow students and co-founders Nate Agrin and Jess Kline. Nate and I worked on it a bit in our spare time after graduating, and I started collaborating with Scott in 2009, still in our spare time. Scott was instrumental in recruitment, funding, and collaboration, and we formed an LLC together as a way to have a bank account to accept funds to build out functionality like the first iPhone and Android apps. In 2014 we joined the California Academy of Sciences (CAS) in the hope of gaining access to more resources to hire more staff, which we did and thus survived almost a decade of growth in usage. In 2023 we left CAS after arduous negotiations and formed an independent non-profit.
Up to this point, iNat functioned as an unstructured anarchy. Scott and I were titular “co-directors” but we did not provide a lot of direction and most of the big moves and features were driven largely by individual initiative. We never found a great way to collaborate. We struggled to set collective goals that might override individual ones, we struggled to keep longer-term goals in mind amid day-to-day firefighting, and we remained merely the sum of our parts, if that. Toward the end of our time at CAS we experimented with sociocracy as a way to organize without hierarchy and coercion, but despite my enthusiasm for the form, we didn’t start with universal buy-in or understanding from the whole team, we didn’t fully adopt its structures, and, like many democracies before us, we ultimately voted to abolish our own democracy when we formed the “leadership circle” and created a hierarchy.
The leadership circle was me, Scott, and Carrie. We didn’t have clearly-defined roles, but our mandate was navigating the separation from CAS while forming a new organization, which meant lots of talking with lawyers, understanding the requirements of tax-exempt status, forming a new board and writing its bylaws. I had very little interest in any of it and to my discredit, I let Carrie and Scott do most of the work. When we did need to make decisions, I was generally the minority, e.g. in how much to disclose in our negotiations with CAS (IMO everything) or whether board members should be required to donate money (IMO no, plutocracy is bad at all times and at all levels). In addition to constantly feeling like the losing vote, the “leadership circle” was sliding into becoming a Leadership team of department heads, and being a dis-empowered leader of people doing the actual work wasn’t what I wanted, so I stepped down and was officially just an engineer, work I’d been doing all along anyway.
The new mobile app (which we were calling “iNat Next” and is now simply “iNaturalist,” the older app becoming “iNaturalist Classic”… still following? I’m going to refer to them as “iNat Next” and “iNat Classic” for clarity) was proving to be more work than the team dedicated to it could handle, so I put all of my time into that. We hired two new staff to work on it, and did our best to reconcile what little guidance we got from Leadership with an internal process focused on discussion and consent. By late 2024, the app wasn’t what anyone wanted it to be, but we had soft-launched on iPhone and were iteratively improving. At this point Leadership started establishing arbitrary goals like a hard launch with promotion in time for City Nature Challenge 2025 with the hopes of getting featured in the Apple App Store. They insisted the app needed to be simpler, to cater first to incidental users who wanted a quick answer, to be a friction-less path to a feeling of contribution. I don’t believe that’s possible while also serving existing users who value (don’t laugh) the power and nuance of iNat, including, among many other things, the way it doesn’t give you a quick answer, forcing you to consider options when making an identification. At this point it was clear to me Leadership wanted the app to be something I had no interest in using and that I didn’t believe would serve people like me, and that they were totally uninterested in hearing feedback from the team developing it, so I left and joined the web team.
Over that spring, Leadership changed direction on iNat Next again and again, shifting the baseline for what constituted a viable feature set for release and driving the team toward a deadline for getting featured in the App Store that ultimately proved fruitless when the app wasn’t featured after all. The team felt dis-empowered and Leadership seemed unwilling to listen to their complaints. Some of the team asked if I could do anything, so I tried to do what seemed like the only thing I could do in a hierarchy and communicate the problems up to the Head of Engineering and Head of Engagement (early March 2025), but there was likewise no change in behavior from Leadership in the way they managed the mobile team. I ultimately proposed restructuring product Leadership under a new Head of Product role with the explicit responsibility of consulting with workers and with users about their needs and capabilities, working closely with design and engineering staff to ensure that no one was blindsided by sudden changes in direction but also ensuring no team was left without product direction (17 April 2025). I proposed that this role have independent control over product decisions to resist impulses from the rest of Leadership to override existing priorities before work could be tested by use. I also proposed that I could do the job.
A day later, Leadership informed me they had no intention of adopting my proposal (“We’re not planning to change the Head of Product structure at this time,” 18 April 2025, referring the three-member sub-group of the Leadership team that was making product decisions at the time). A few days after that Scott announced that he was taking on an additional role to his Executive Director responsibilities: Head of Product (circa 21 April 2025). On 5 May 2025, the Leadership team summoned everyone who worked on iNat Next into an agenda-less meeting titled “post launch Strategic alignment” and announced that they would offer us half a year’s pay to quit. Offended and incensed that they would rather so many staff leave than listening to and addressing their concerns, I decided iNaturalist was no longer an organization I wanted to work for and I told them to write up the offer. I consulted with friends and advisors, who all suggested that I take longer to decide, perhaps on a sabbatical. Leadership agreed to that, so I tried to spend the last few months thinking about whether I could remain in an organization led by people in whom I have lost so much faith. In the interim, most of the recipients of that buyout offer left the organization, as well as another engineer, and the Head of Engineering, resulting in a 30% attrition in staff. The org is trying to fill that gap with the three new engineers that have been hired for the mobile team, former board member Dan Rademacher joining staff as Head of Product (though not, apparently, with the independence I proposed, so responsibility without the necessary power), and hiring two more engineers for the web/ops team (in progress, to my knowledge).
As I was leaving, the Google gen AI debacle happened, a fiasco big enough to merit description in Scientific American. This was an own goal. iNat’s Engagement team predicted the backlash but the Leadership team chose not to listen to their warning. I was not involved in this grant, its announcement, or its fulfillment, but while it didn’t directly lead to my own choice to leave, it is symptomatic of the problems that led to that choice, and I think it played a role convincing more staff to quit. In my discussions with staff since, almost everyone recalled being blindsided by the announcement of the grant and confused about what the money was going to be used for.
Product
My fundamental difference from Leadership in terms of the organization’s products is that I think different products should meet different needs, while the Leadership team believes a single product can meet all the needs of the organization’s potential users (I have many others differences, but this might be the least reconcilable). At the beginning of iNat, I thought one product could meet all those needs too. In grad school, we described personas of potential users and quickly identified the divide in needs between enthusiast naturalists and people with a more occasional interest in nature, and we tried to design to meet the needs of both groups of people. Again and again over the years, I saw that we just couldn’t do it. Enthusiasts need power and complexity that is immediately confusing and intimidating to more casual users, and that extends from the use of scientific names through the kinds of filters you can use to explore the data. Casual users need structure and guidance to invite them into the amazing but complex and often unfamiliar world of biodiversity, but that structure can be a hindrance or even infantilizing to people who just want to get things done. iNaturalist the product is fundamentally complicated, and I have watched many, many people bounce off that wall of complexity over the years, even as I’ve seen so many people enrich their lives after they climb over it.
The last nail in the coffin on this subject for me was Seek, an idea from Alex and Joelle (later rebuilt by Amanda and Abhas) to build a version of iNat specifically for that casual crowd, focused on quick answers with a gamified structure. Seek was wildly successful, eventually equaling and occasionally exceeding the usage of the iNaturalist apps. Two critical anecdotes: I remember looking for rare plants in a burn zone and overhearing two young botanists one hill over trying to figure out a plant and eventually saying, “well let’s just see what Seek says.” I was alarmed that these professionals were using such an unprofessional tool, but also reminded that in that moment, they didn’t need iNat-level complexity. They just needed a Seek-sized nudge in their identification process. One summer a year or two later I went home for my annual family vacation and my dad told me he’d just discovered the marvels of Seek and was puttering around everywhere seeing what Seek thought of the plants in the yard. Dad has been a meticulous recorder of information his entire life, and has been a loyal iNat contributor for years, but he has never had much interest in computers and absolutely couldn’t give a toss about connecting with other people on the Internet, so Seek really worked for him in a way iNat never did.
iNat’s history and Seek’s success prove to me that my initial belief in a single app to meet all needs was wrong: iNat the product should serve the enthusiast users, and Seek should serve the casual users. Arguments about Seek eating into iNat’s potential usage are absurd: the mission of the organization is to connect people to nature through technology, and if two products are doing that better than one product, that’s success. If Seek users aren’t contributing to a global data set, that’s also fine. Data generation has always been a byproduct of building connection. Seek could be better about encouraging data contribution, but it doesn’t need to.
iNat’s current Leadership does not share this belief. To them, Seek is an off-brand liability that they don’t intend to improve. They think iNaturalist the product can serve those Seek users while also serving existing core iNat contributors to the detriment of neither.
Management
Leadership made many mistakes over the last year, including
- Failing to listen to the concerns of the mobile team in the development of iNat Next and overriding their decisions
- Abandoning product direction on the web team and for Seek
- Trying to address discord by jettisoning staff instead of addressing their concerns
- Accepting a grant without any consultation with staff about how its obligations might be met
- Ignoring staff warnings about how the announcement of that grant would not go over well with users
I lay the departure of 30% of staff at their feet, primarily because of the buyout offer. I think that showed the entire organization that their bosses are the kind of people who deal with criticism with a firm shove out the door. Yes, the people who left may have had other opportunities, but they primarily left because they were not being consulted when they had feedback, not getting help when they needed it, and not granted the level of independence they needed to get things done. The reputational damage from the Google AI grant was similarly due to Leadership’s inability to hear and digest criticism. If they had listened to the warnings of the Engagement team, it might have gone down very differently.
Since the exodus, Leadership has improved on some fronts. Within the Engagement and Engineering teams they are consulting with staff more than they have in the past. Within Engagement that consultation seems to be translated into action more than it has in the past. They hired three new engineers for the mobile team that seem both experienced and enthusiastic. Bringing Dan into a Head of Product role will hopefully provide more empathy and clarity to the engineering teams.
But fundamentally for me, even if Leadership has learned something from their mistakes, they have not learned to admit them, which proves to me they haven’t really learned to accept criticism. There has been no all-staff discussion about stumbling through the release of iNat Next, of the Google gen AI debacle, or, most importantly, why 30% of staff left the organization in three months. The mistakes of the Leadership team are partially or totally to blame for these events, and if they do not own their responsibility before staff and before the community, I don’t think they deserve their power. They certainly don’t deserve my trust.
What’s Next
First of all, I quit. I can’t work with a Leadership team that has such a different vision for where iNat the product needs to go and how iNat the organization should be managed. Or if I can, I can’t find satisfaction while doing so.
I’d like to keep working on natural history software, and if you’d like me to as well, support me on Patreon. In a perfect world, I’d be employed by the people who use the software I make, so I figured I’d try it. Right now I’m working on a way to back up your iNat observations and an app for viewing geologic maps. Both could use a lot of work. In the future I’d like to experiment with a decentralized version of iNat. In all likelihood I’ll need to get a “real” job so let me know if you’d think you’d like to hire me.
I am, of course, heartbroken. I genuinely like almost everyone on staff and I will miss working with them. It would be hard to find a more talented, bright, and perceptive group of people, and they genuinely care about iNat and are trying to do right by it, which makes it all the more tragic that the organization can’t seem to do right by them.
To those who are mystified by a co-founder’s choice to leave the thing they co-founded, I don’t really care about the fact that I helped create iNat (it was not a unique idea), but I very much care about it existing in something like its present form because I use it every day, interact with people like me on it every day, and it hurts to step away from an active role in improving and maintaining it.
Obviously, I think the Leadership team should change, and maybe they will. Hiring Dan is a good step, but I think they will need more fundamental changes to their approach to right the ship, starting with learning how to admit when they’re wrong.
I also think the board should change, because they are either complicit in Leadership’s mistakes or failing in their oversight function. I think the board should develop lines of communication with staff and users outside of the Leadership team. We set up iNat’s board to be largely non-interventionist, but I think that was a mistake. Now they only get information from a Leadership team that cannot admit fault. If the board is really to perform its oversight function, it should be hearing from staff and users more directly.
I also think the board should follow Wikimedia’s lead and commit to a majority of community-selected members, instead of selecting their own membership from people like them. Users should have more power in the organization. Right now they’re barely even informed about what staff are doing.
In case anyone is alarmed about the state of iNat after reading this, I do not advise stopping your use of iNat, and I definitely don’t advise deleting your account. I’m certainly not doing either. I was appalled at the number of people who went nuclear and deleted their accounts during the gen AI debacle. Temporarily withdrawing your data would be a reasonable form of protest if it was possible, but destroying existing data that can’t be retrieved hurts everyone. If you think iNat has problems, either the ones I’ve described or others, I think you should keep contributing observations, identifications, and if you can, donations, but also organize. Users don’t currently have any effective control over staff or the board, but you do control the life blood of iNat: participation. A large enough cessation of data or money would exert pressure, but it would take organization.
If you’re concerned, I’d also advise exploring or creating alternatives. While I don’t think iNat has reached Doctorow-level enshittification, it does suffer from centralization. If you don’t like the way your data is used to train AI models, you can’t currently move your data to a non-AI service while still contributing to and using iNat data, even if such a service existed. But you should be able to do that. It also suffers from stagnation. If you want features like client-side geofencing, video support, sonograms, sound ID, etc, you might need to build them. iNat still maintains the web 2.0 virtue of adversarial interoperability in the form of its API, so some of these alternatives can be built in non-competitive ways (and if the API goes away, we’ll know things have gotten really bad), but some may require competition. Those of us who benefit from tools like iNat should be looking seriously to the decentralized models being developed by the likes of Bluesky and Mastodon, because we can’t rely on any single organization to provide that benefit forever.