[Incubator] GeoLens community project application

Tue Sep 29 13:32:30 PDT 2026

FWIW I’m fine with AI agents and use them myself and agree AGENTS.md and CONTRIBUTING.md is something we should flesh out.  I think about a year ago I would have been squarely on Even’s side.

Now not so much just because I’ve seen vast improvements of the quality coming out of AI.

I still very much appreciate Even’s take on things and am very glad that we are not completely aligned. His “AI makes you stupid and leads to brain rot”  (I’m paraphrasing a bit) does give me pause every time I am tempted to relegate my thinking to AI bots.

What has changed about me – well I’ve seen PostgreSQL project use it a lot, we in PostGIS project use it a lot.  I’ve been able to contribute to projects in python and go I never would have been able to before because of it. E.g. I find a bug, I figure out how to fix it with AI, test what AI produced, try to understand what AI produced and do a pull request. So now I can give a bug report and a solution.

But Yes in some ways AI kills community – I see projects such as GDAL contend with garbage AI submissions.  This past run of Google Summer of Code (GSOC) developers I was disheartened to see many new coders relegate all their thinking to AI and not even know how to use Git anymore. Just recently we got a garbage pull request in pgrouting from someone hoping to be a GSOC 2027 contributor.  It’s been harder to pick out good GSOC proposals from the numerous AI crap we get. So if that direction continues, then AI will have no knowledge except past works of art and the infiltration of reckless and sometimes evil people.

I’ve seen new people who have used my projects e.g. a new person contributing to postgis tiger geocoder starting to contribute and even though his contribution is clearly AI generated and admittedly so, it looks like care was taken to test speeds and produce it.

But what I am still concerned with is the balance of things, and it’s not clear to me yet whether AI code contributions is a net positive or negative.

Yes I still want Even waving the guillotine to keep our group balanced. The worst thing possible is for us to all get lulled into AI as a solution to all development cause then really bad things will happen.

I consider myself a prag-optimist. I think AI is here to stay and even if I hate it, it is a lost battle, so the question is how can we make it not suck?

I think that with enough instruction, we can guide AI use so it is a net big positive.  So my focus now is how to give instruction and how to mitigate against the bad.

Try to understand what I do and don’t like about it, what gives me an uneasy feeling, and how to change that.

I was watching a video recently that said “AI is communication between people”.  I think that is very much the case.  AI imitates humanity and if all it has to imitate is bad code and bad behavior, then we’ll have very bad consequences.  However it is in our power to amplify the good and to destroy the bad, the question in my mind is how? We must not let innovators be drowned by AI or the youth relegate their thinking to AI, cause then all we’ll be left with is AI cow pox.

Thanks,

Regina

From: Incubator <incubator-bounces at lists.osgeo.org> On Behalf Of Even Rouault via Incubator
Sent: Tuesday, September 29, 2026 12:23 PM
To: Ian Shiland <ishiland at gmail.com>
Cc: Jody Garnett via Incubator <incubator at lists.osgeo.org>
Subject: Re: [Incubator] GeoLens community project application

I'll be frank and very old school: human accountability is a joke when you delegate code writing and reviewing to bots. AI is destroying open source and everything else. Fuck it

Le 29/09/2026 à 18:08, Ian Shiland a écrit :

Hi Nick/Even -  

One of the main reasons for applying for the OSGeo Community Project status is to help with product direction and attract other contributors, using AI agents or otherwise. 

I agree adding a section in CONTRIBUTING.md (human accountability, etc) makes sense; I also think having some standards for AGENTS.md is probably something to consider (testing, scope, and security expectations). I'm glad to do any of these if that is the direction the community aligns with. 

I appreciate the discussion regardless of the outcome for GeoLens here. 

Ian 

On Tue, Sep 29, 2026 at 3:45 AM Even Rouault <even.rouault at spatialys.com <mailto:even.rouault at spatialys.com> > wrote:

Looking at https://github.com/geolens-io/geolens/graphs/contributors?from=27%2F06%2F2026 , you can see that there were 4 other committers than Ian (excluding dependabot), contributing a grand total of 7 commits compared to Ian's 1095, and looking at their commits, it is clear (from the verbosity of the commit message) that 3 of them also used an agent.  So the collaboration space is limited to people using agents.

Le 29/09/2026 à 08:35, Nick Bearman via Incubator a écrit :

Hello,

Here are my thoughts. 

The board said they are not banning or enforcing the use if AI, it is up to individual projects. They didn't say about use of AI in Communtiy Projects. 

I suggest we apply the rest of the criteria - which, if memory serves incudes that the project shows collaboration:

<https://wiki.osgeo.org/wiki/Incubation_Committee#Step_2:_Join_OSGeo_Community_Projects_Initiative>

We ask that the project demonstrate collaboration, perhaps with a history of bug report or pull requests.

So I think this project qualifies. I've not reviewed everything else but assuming it does then I think it is ok. 

Please do share more thoughts everyone!

My personal opinion is that it would be worth making clear how you used AI coding agents Ian. Perhaps this is something to add to the README or CONTRINUTING  files?

Best wishes
Nick. 

On 28 September 2026 23:03:19 BST, Jody Garnett via Incubator  <mailto:incubator at lists.osgeo.org> <incubator at lists.osgeo.org> wrote:

The board has not really provided guidance on vibe coding. 

My own personal opinion is that open source requires collaboration - the test our incubation process has is to have input from several organizations (not just several developers) in order to offer users some assurance that a project will be sustained. With that in mind, while I am okay with AI being used for assistance / reference, it has some weakness from the sustainability angle.

The above is my own take only; there was some osgeo discussion over the summer but I did not understand the result.

- -

Jody Garnett

On Sep 28, 2026 at 2:55:04 PM, Ian Shiland <ishiland at gmail.com <mailto:ishiland at gmail.com> > wrote:

Hi Even,   

Yes. AI coding agents, mainly Claude Code with Codex as a reviewer, write most of GeoLens with guidance from me. 

If this is grounds for ineligibility, I'll withdraw the page submission. 

Ian

On Mon, Sep 28, 2026 at 5:27 PM Even Rouault <even.rouault at spatialys.com <mailto:even.rouault at spatialys.com> > wrote:

Hi,

@Ian looking at the commit history and pull request history, this is obviously vibe coded. Can you confirm ?

@Jody  Is this OK for vibe coded projects to be community projects ?

Even

Le 28/09/2026 à 23:15, Ian Shiland via Incubator a écrit :

Hi Jody,

Thanks for the access. The GeoLens project page is now submitted for review. I left the project type as None for now and will switch it to Community Project once the committee confirms. Let me know if anything on the page needs changing.

Ian

On Mon, Sep 28, 2026 at 12:33 PM Jody Garnett <jody.garnett at gmail.com <mailto:jody.garnett at gmail.com> > wrote:

Hi Ian,

I added you as project author role, so you should be able to create a page now!

- -

Jody Garnett

On Sep 20, 2026 at 7:42:07 AM, Ian Shiland <ishiland at gmail.com <mailto:ishiland at gmail.com> > wrote:

Hi Jody,  

I’ve confirmed that my OSGeo account, ishiland, still cannot access the Projects dashboard - it returns a permission error. Could you enable project-author access and confirm the current status of GeoLens’s Community
Project application? The listing content and logo are ready. 

Thanks, 

Ian

On Sun, Aug 23, 2026 at 1:42 PM Ian Shiland <ishiland at gmail.com <mailto:ishiland at gmail.com> > wrote:

Hi Jody,

I have an OSGeo user id now: ishiland. Regina sent a mantra through sysadmin, so registration is done and a profile has been created. 

Thanks,
Ian

On Sat, Aug 22, 2026 at 1:57 PM Ian Shiland <ishiland at gmail.com <mailto:ishiland at gmail.com> > wrote:

Hi Jody, I've requested a mantra from mantra-request at osgeo.org <mailto:mantra-request at osgeo.org>  so I can create the account. I'll send the user id as soon as it comes through.

Thanks,
Ian

On Fri, Aug 21, 2026 at 1:33 PM Jody Garnett <jody.garnett at gmail.com <mailto:jody.garnett at gmail.com> > wrote:

Hi Ian,

Please provide your osgeo user id, and sign into the osgeo.org <http://osgeo.org>  website to fill in your profile information. Then I can grant the profile permission to make a new project page.

For more information see: https://www.osgeo.org/community/getting-started-osgeo/add-a-project/

- -

Jody Garnett

On Aug 21, 2026 at 6:33:34 AM, Ian Shiland via Incubator <incubator at lists.osgeo.org <mailto:incubator at lists.osgeo.org> > wrote:

Hello,

I'd like to apply for OSGeo Community Project status for GeoLens, a self-hosted geospatial data catalog and map builder built on PostGIS. It ingests common vector and raster formats, serves them through OGC API - Features, OGC API - Records, STAC and DCAT, and includes a web map builder with vector tiles.

*	Repository: https://github.com/geolens-io/geolens (Apache-2.0)
*	User documentation: https://docs.getgeolens.com <https://docs.getgeolens.com/> 
*	Live demo: https://demo.getgeolens.com <https://demo.getgeolens.com/> 
*	Contribution policy: .github/CONTRIBUTING.md (DCO)
*	Code of conduct: CODE_OF_CONDUCT.md (Contributor Covenant)
*	Issue tracker and discussions: public on GitHub

The first public release was June 2026 and the project has shipped regular tagged releases since (currently v1.14.1). The repository has issues and pull requests from outside contributors, and GeoLibre (an OpenGeos project) ships a built-in GeoLens integration that I've contributed fixes back to.

If a project listing on osgeo.org <http://osgeo.org/>  needs to exist before this application, point me at the right process and I'll take care of it.

Happy to answer any questions.

Ian Shiland

_______________________________________________
Incubator mailing list
Incubator at lists.osgeo.org <mailto:Incubator at lists.osgeo.org> 
https://lists.osgeo.org/mailman/listinfo/incubator

_______________________________________________
Incubator mailing list
Incubator at lists.osgeo.org <mailto:Incubator at lists.osgeo.org> 
https://lists.osgeo.org/mailman/listinfo/incubator

-- 
http://www.spatialys.com
My software is free, but my time generally not.
LLMs contribute to global warming and brain rot.
Let's guillotine them! "Ah ! ça ira, ça ira, ça ira !"

_______________________________________________
Incubator mailing list
Incubator at lists.osgeo.org <mailto:Incubator at lists.osgeo.org> 
https://lists.osgeo.org/mailman/listinfo/incubator

-- 
http://www.spatialys.com
My software is free, but my time generally not.
LLMs contribute to global warming and brain rot.
Let's guillotine them! "Ah ! ça ira, ça ira, ça ira !"

-- 
http://www.spatialys.com
My software is free, but my time generally not.
LLMs contribute to global warming and brain rot.
Let's guillotine them! "Ah ! ça ira, ça ira, ça ira !"
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/incubator/attachments/20260929/ad2c8a84/attachment-0001.htm>