I've recently gotten a bunch of questions for an "async" interview. Some of them triggered more in me that I expected, as such, I'm pasting them below:
When did your interest in programming and interoperability begin?
(Did it emerge alongside your work in architecture, or was it driven by a specific need for better collaboration in projects?)
It started much earlier than that - in high school. I was part of a specialised class doing 8hrs/week of C++ programming for 4 years, and I was very lucky to have an inspirational teacher that encouraged me in my hobbies: that’s how I started using ActionScript and Processing for interactive animations, presentations and installations (even a theater production). It’s also where I learned how to normalise a database. For my 12th grade dissertation I wrote a naive 3D engine in ActionScript, which was fun, but was maxing out at around 10FPS. Still, I learned about quaternions and projection systems!
I ended up studying architecture as I perceived it as a blend of technical and creative thinking. I kept on coding throughout university and early on started doing workshops for RhinoScript and ExplicitHistory (which soon thereafter became Grasshopper). My interests during this time were much more focused on generative geometry and fabrication, and I paid little attention to interoperability in its classical sense. I was focused on designing better buildings with the aid of code. Data, and transferring data back and forth, was just a means to an end: as an architect, you need to identify the correct abstraction (3D, 2D, sections, VR, AR, physical models, etc.) in which your intent is best communicated and you continuously transfer “data” from one to another.
My interest in interoperability, and how data is central to everything we do, was shaped largely by experience working in the industry - a later moment in my life. I’ll get to that in the next answer!
How and when did the idea for Speckle come about?
(Did it start as a personal solution to the frustration with data silos, or was it envisioned from the beginning as an open and scalable platform?)
I was working in Brussels at a boutique office. At one point I ended up doing an interactive application that took an economic model (excel) and linked it up with a masterplan model for Brussels Nord. After my initial experiments with Grasshopper, I realized that that was not enough, so I ported everything over to Processing and built it as an Android and IPad application - this was then used by 10x more stakeholders to engage with our design process and thinking. It linked us, architects and urban planners, with the economists, the developers and the mayor.
That was the first spark of an idea behind Speckle: we produce designs and data, but if others cannot engage with them, it’s all for nothing. This was a core frustration of mine: so much effort and thinking that we, as architects do, but so little of that transpires outwards. How can we make our designs (and associated data) accessible to more people? How can we make the compromises that we weigh at every step visible and understandable? Why is BIM (and even Grasshopper!) not helping us here at all, rather the contrary, hindering us, locking us in complex tools that others can’t use, or proprietary file formats?
After 3 years there, I ended up at UCL, London doing a research project exactly on that - and the real Speckle was born as a tool to enable fast data exchange - initially between technical stakeholders (architects, engineers, etc.) but soon thereafter also with wider audiences. Speckle was the byproduct of a lot of thinking, reading and writing (see my PhD thesis https://dimitrie.org/thesis/ if you’re bored!), and it was the key vehicle I used to define the problem.
In short, the core thesis at the heart of Speckle is that communication is not just a technical problem (like data formats), but also a social one. Current tools often focus too much on structure and standardization, while real-world collaboration is messy, human, and context-dependent. Interoperability is not about transferring perfect data, but about enabling meaningful conversations (with humans and machines).
Do you still see yourself as a Citizen Developer?
(As someone who, without formal training in computer science, has created powerful tools from deep domain knowledge in AEC workflows.)
Absolutely! I think being a citizen developer means also focusing on your responsibilities as a citizen: giving back to your community, which - in my case - I see as AEC. That’s why Speckle is open source, and all our code is available for others to get inspired on, copy paste, or build on. There’s a lot of innovation happening behind the closed door of various offices, which is a trend I dislike - it’s bringing the entire industry down. Why not collaborate and share when it comes to digital processes and tools, and stop reinventing the wheel? Speckle - in part - answers this by providing an open source base on which to build on and channel that creativity and energy.
It has not been straightforward to maintain this stance once Speckle became a company. Five years ago, the overall understanding of what open source is was contained to more niche people; nevertheless other brave people, behind other open source projects (such as the amazing BonsaiBIM) kept advocating for it, and, together, I like to believe we moved the needle.
Are you currently exploring or planning to integrate AI into Speckle or your roadmap?
(For example, using LLMs for data visualization, code generation, or intelligent interoperability agents.)
Always, and there have been some compelling community demos around this topic (as well as some enthusiastic internal hackathons). Right now, AI is in a bit of a goldrush moment: everyone’s searching for a meaningful application. The way we position Speckle, in the context of this gold rush, is as infrastructure: we’re selling shovels. To have productive AI applications, you need to have access to your data, and that data needs to be correct. This is what we are focusing on right now: providing a fundamental layer of object-based storage for BIM and geometrical data.
Practical applications that truly transform the industry have yet to materialize. There are some small pragmatic ones that we are working towards, but nothing really that would replace an architect.
What would you say to someone in the AEC industry who wants to start building their own tools but feels they don’t have a “technical” background?
(I’d love to hear your thoughts here, especially to inspire aspiring Citizen Developers who are just getting started.)
I can’t have enough words of encouragement - go for it! Right now, there are many AI (and not AI) based coding tools - programming skills is a much less important requirement to get going, but, nevertheless, don’t forget to dive into the fundamentals when you get to it; it will pay off.
I’d also want to encourage aspiring citizen developers to share with the wider world their progress - regardless of the level of polish or perfection. Speckle’s success and outreach is partly due to me making it open source from day one; it’s how I met my co-founder, Matteo, and this later led to us establishing the company that accelerates our vision.
As a citizen, you have the responsibility to participate in your community. I can’t wait to see you all there!