On September 21, Dan Greenheck, a well-known developer in the Three.js community, posted a long tweet. The gist was that he is considering hanging up his Three.js hat: he wants to finish Water Pro v4 first, and then think about what to do next. He said this was not a pity post, and that he had not fully figured things out himself. I think this is quite relevant for those working in WebGIS, digital twins, and 3D visualization, so today I will talk about it with you.
About Dan Greenheck
Dan Greenheck has long created commercial Three.js assets and tutorials. His representative work is Three.js Water Pro, a paid plugin for realistic water effects, currently priced at USD 199 on the official website.
Image above: Dan Greenheck's post
Reasons He Is Considering Giving Up
In the tweet, he listed several points. I have roughly organized them as follows:
- After Astra launched, AI produced effects close to or even surpassing Water Pro in visual quality.
- Someone recorded his demo video and fed it to Astra, reproducing it with about 90 percent accuracy.
- Customers are becoming fewer and fewer.
- A Three.js developer he deeply respects is preparing to leave the industry because of AI disruption and declining interest.
- Sales of courses and plugins have been hit by AI.
- Three.js content on social platforms is gradually shifting from works by senior developers to a large number of mini-games and demos created through vibe coding.
When things can be easily replicated, it becomes hard to see the value of this work. He likes tackling hard problems; after outsourcing the difficult parts to agents, the fun is gone too. Having nothing left but dispatching agents every day does not excite him. He also said that after not writing code by hand for a long time, his critical thinking, focus, and comprehension have clearly declined.
GIS-Related Reflections
First, I have to say that with AI, technical value has been compressed. A simple example is today's job market. Previously, salaries for frontend, backend, and WebGIS roles were at least decent, and there were many openings. Now an obvious phenomenon in China is that positions are decreasing and salaries are falling. The little red dots on Boss Zhipin are becoming fewer and fewer.
Of course, from another perspective, lower technical barriers are not necessarily all bad news for GIS. In the past, to get decent water surfaces, fog effects, and procedural terrain in a 3D scene, someone often had to specialize in shaders. Now, with AI tools such as Astra, even small teams can first set up a scene and leave more energy for data governance and business rules, the parts of GIS that are truly hard to crack.
In addition, all kinds of flashy demos on social platforms do not mean that someone can deliver maintainable products over the long term. A single effect is easy to replicate, but business data, performance tuning, multi-platform compatibility, project delivery, and long-term maintenance are still parts of GIS work that are not easy to replicate. (But with the development of AI, who knows? Writing this, I feel a bit uneasy myself.)
Summary
Dan's tweet exposed the brutal impact on ordinary individuals behind the prosperity of AI. Even a big name at his level is about to give up, so what choices do the millions of practitioners in China have?
Finally, if you are also working in 3D visualization, digital twins, or WebGIS, have you felt similar anxiety recently? Feel free to share your observations in the comments.
References
- Dan Greenheck's original tweet (I think it's time to hang up my Three.js hat.): https://x.com/dangreenheck/status/2102137910921240807
- Dan Greenheck's LinkedIn profile: https://www.linkedin.com/in/danielgreenheck
- Building games with Astra (OpenAI): https://developers.openai.com/blog/how-to-build-games-with-astra