Guest on 2D Is Against My Religion

Martin Piala

Co-founder, Blurry

Martin Piala is a co-founder of Blurry, a London-based startup building a browser workspace for reality capture data. Recorded live at INTERGEO 2026, Martin Piala explained why Blurry moved beyond Gaussian splat viewing to point clouds, images and processing, and why the data format, not the model size, limits AI on 3D data.

Episode

17 min

Expo & conference podcast

Gaussian splatting is not enough: 95% still use point clouds – Martin Piala, Blurry

Martin Piala · Blurry1 Nov 2026

In Martin’s words

“We realized Gaussian splatting is not enough.”

Listen at 0:45 · Gaussian splatting is not enough: 95% still use point clouds – Martin Piala, Blurry

“We notice that 95% of them are still working with point clouds.”

Listen at 3:34 · Gaussian splatting is not enough: 95% still use point clouds – Martin Piala, Blurry

“We don't want to put out, like, a half-baked product.”

Listen at 7:40 · Gaussian splatting is not enough: 95% still use point clouds – Martin Piala, Blurry

“The size of the model is not the limiting factor.”

Listen at 11:02 · Gaussian splatting is not enough: 95% still use point clouds – Martin Piala, Blurry

“Point clouds are still not dead. Images and videos are still not dead.”

Listen at 0:55 · Gaussian splatting is not enough: 95% still use point clouds – Martin Piala, Blurry

Quoted in our tech guides

Gaussian splatting vs photogrammetry: what 18 practitioners say →

“We still need point clouds for measurements.”▶ 1:10

Questions Martin answered

What is Blurry?: According to Martin Piala, Blurry is a workspace for reality capture data: point clouds, Gaussian splats, images and videos in one place, where they can be analysed. ▶ 0:09
Is Gaussian splatting replacing point clouds?: Not yet, according to Martin Piala: 95% of Blurry's users still work with point clouds, which remain the format for measurements, while splats are a nicer way to present the data. ▶ 3:19
Who processes Gaussian splats for Blurry?: According to Martin Piala, processing currently runs through Teleport by Varjo; Blurry is working on its own processing for niche cases such as 360 cameras. ▶ 4:27
Can AI understand point clouds?: Not well yet, according to Martin Piala: current models understand images better than 3D geometry, so screenshots or recordings of the data work better than raw point clouds. ▶ 10:59
How is Blurry funded?: According to Martin Piala, Blurry raised pre-seed funding from two VC funds and plans a seed round within six to nine months. ▶ 15:23

Talks about

Olli Huttunen – 3D Gaussian Splatting creator
Bert Azizoglu – CEO and co-founder, NUBIGON
Arek Szadkowski – Global Lead, ArcGIS Reality, Esri
Gregor Willhauck – Go-to-market, cloud for field systems (Trimble Connect), Trimble
Ronnie Yaron – Founder, Skyline Software Systems
Dylan Gorman – Drone mapping, PilotByte Labs