#technicalart

[echocircuit]

𝚍𝚊𝚝𝚊𝚋𝚊𝚜𝚎.𝚚𝚞𝚎𝚛𝚢 → 𝚕𝚘𝚊𝚍𝚒𝚗𝚐: 𝚜𝚞𝚙𝚎𝚛𝚜𝚘𝚗𝚒𝚌_𝚙𝚛𝚘𝚝𝚘𝚌𝚘𝚕… 𝚎𝚛𝚛𝚘𝚛: 𝚍𝚊𝚝𝚊_𝚞𝚗𝚜𝚝𝚊𝚋𝚕𝚎.

[artstationable]

Rasmus Poulsen

Franchise Art Director - IOI

artstation twitter pinterest technouveau.net instagram linkedin behance

More from «Artstation» here

[williamphenexart]

VTuber Lineart Update: Shirt Detail! 👕

Finished the lineart for the shirt! This isn't just a regular drawing; I specifically designed this in a "2D working as 3D" style.

This means I focus heavily on defining every fold, layer, and seam to give the clothing volumetric structure. This complex approach is absolutely necessary for the Live2D rigging, allowing the clothes to move and deform naturally instead of looking flat. Huge technical step!

This project helps keep the content flowing while the main comic prep continues!

Actualización del Lineart VTuber: ¡Detalle de la Camisa! 👕

¡Terminé el lineart de la camisa! No es solo un dibujo normal; lo diseñé específicamente en un estilo de "2D que funciona como 3D".

Esto significa que me centro mucho en definir cada pliegue, capa y costura para darle a la ropa una estructura volumétrica. Este enfoque complejo es absolutamente necesario para el rigging en Live2D, ¡permitiendo que la ropa se mueva y se deforme de forma natural en lugar de verse plana!

All my social media:

williamphenexart.carrd.co

[nedmakesgames]

I'm back with another Unity shader tutorial! To continue with writing a complete HLSL shader for URP, we're looking at fancy techniques to simulate different materials, like metals, glass, paint, and plastic. And, we find out just what "PBR" is and see the simple method to add it to your shader.

📺 Watch: https://youtu.be/5GGISvt4KEA

📚 Read: https://nedmakesgames.medium.com/6c4ae9875529

💾 Download: https://www.patreon.com/posts/81272385

I hope you enjoy, thanks!

[jasonthielke]

Hello winter ❄️ #profile #denverartmodel #fashionillustration #graphicillustration #femaleface #faceinprofile #artcontemporain #lineartist #aart #artlife #femalefiguredrawing #linework #newcontemporaryart #arte #artsupporters #technicalart #artdetail (at Denver, Colorado) https://www.instagram.com/p/CmhEKNNLyfT/?igshid=NGJjMDIxMWI=

[adox-tnw]

Everyday I'm shadering. Tienes que meter un logo en Unity que lleva un círculo glowing descomunal. Tienes dos opciones: importas un PNG enorme súper pesado... o pintas un #glowing #circle con un #shader 😉. Just little #technicalART things that make a difference. (at Barcelona, Spain) https://www.instagram.com/p/BzxQBC6CtHY/?igshid=8xh99n6253uk

[lukeduoart]

Started on my next sketch. Can you guess which ship this is? ⭐️ (Hint: It’s not an x-wing)

[artsclevervivekdevelopingskills]

Why I Spent Part of My Major Study Building a Tutor Instead of My Outpost

I want to be straight about how this part of the semester actually went, because the honest version is more useful for a development blog than a tidy one. This post is about a decision that took me away from building the EREMUS outpost for a while, and why I think it was the right technical-artist call to make.

The problem I kept hitting

Most of what I do (Houdini, Unreal, Substance) I learn from tutorials. And I kept running into the same wall: on a single watch of a tutorial I would retain only a fraction of it. I would rewatch, scrub back, and lose the thread of why a node was there. A video cannot do what a teacher does: it cannot check my work, it cannot take a doubt, and it cannot adapt when I get something wrong. Meanwhile the industry is moving quickly: AI is now part of real production pipelines, and the pressure to skill up fast is real.

I would finish a two-hour Houdini tutorial and realise I could not rebuild half of it without scrubbing back through the video: the steps were in my head, the reasons were not. The pattern was always the same: watch, follow along, feel like I understood it, then hit a blank the moment I tried it on my own scene a week later. It hit hardest on a procedural tutorial where one node's setting changed everything downstream, and the video never paused long enough for me to understand why.

Treating it as a pipeline problem

Before games, I spent two years as a full-stack software engineer. So when I hit a recurring problem, my instinct is not to endure it. I treat it as a pipeline problem and build a tool for it. That is exactly what I did here. The result is an AI tutoring system that turns any DCC tutorial into an actual tutor. It breaks a tutorial into a module-by-module lesson plan, then, while I work live in Houdini or Unreal, it reads my real scene through an automation bridge at each checkpoint and tells me whether I have done the step correctly, answers my doubts grounded in the official documentation, and flags where I have deviated.

The first time it caught me building a node in the wrong order and said so before I had stacked ten more steps on the mistake, I realised it was doing what a teacher does, not what a video does. I asked it why a heightfield mask was not affecting my terrain, and it answered from the actual node documentation instead of leaving me to guess. That was the moment it stopped being a toy. At one checkpoint it read my live Houdini scene and told me my HDA was missing an exposed parameter the tutorial relied on, something I would never have caught on a single watch.

Why this is not a detour away from EREMUS

Crucially, I built and tested it on EREMUS's own skill stack, the exact heightfield, scattering and Houdini-to-Unreal tutorials this project depends on, including the Terrain Handbook and the procedural desert workflow. So it is not a detour away from EREMUS; it is the thing that teaches me EREMUS faster, and with verification instead of guesswork.

The honest evaluation

It took more of the semester than I budgeted for. That is on me, and it is the reason I have had to reduce the scope of the artefact, a decision I am documenting separately as project management. But the learning outcome is real: I now learn these tools with feedback rather than by guessing, and I can point to exactly where I was right and where I was wrong.

Was it worth it? For the artefact, it cost me time I cannot get back. For everything after this project, I think it was. I came out able to learn these tools with feedback instead of guesswork, and that compounds. Also it also expose to what i want to be in this field , "A Technical Artist" from my perspective i resolve many of my real-life problem regarding my course or learning and came up with a technical solution for them.

Where this sits in the field

This is not a niche idea. It is a known problem with recent evidence behind it.

Benjamin Bloom's classic 2-sigma finding (1984) showed that one-to-one tutoring makes students perform about two standard deviations better than standard classroom teaching. The catch has always been that human tutoring does not scale.

A 2025 randomised controlled trial by Kestin et al., published in Scientific Reports, found that students using a well-designed AI tutor achieved over double the median learning gains of in-class active learning, in less time.

And the industry context supports the direction: the GDC 2026 developer survey reports that roughly half of game studios now use AI in development, and 2026 trend reports name the "AI Workflow Specialist" as an emerging role.

Positioning myself

So the thing I built sits exactly on that line, scaling the one form of teaching we know works, aimed at the specific DCC tools I need for game environment art. Positioning myself there, between the craft and the pipeline, is the technical-artist direction I want to move in, and it is the same software-engineer instinct that got me into this field, now pointed at learning art faster.

Sources to cite in Harvard style in the planning folder's references:

Bloom, B.S. (1984). The 2 Sigma Problem. Educational Researcher, 13(6), 4-16.

Kestin, G. et al. (2025). AI tutoring outperforms in-class active learning: an RCT. Scientific Reports.

GDC 2026 developer survey / BCG 2026 Global Gaming Report (AI-adoption figures).