Sponsored by
Have you ever tried to learn 3D design?
Every 3D artist starts with the same thing: Andrew Price's Blender Guru Donut Tutorial.
It's a 14-part YouTube series that has become the universal rite of passage for anyone learning 3D.
The reason people start with a donut is because making a good one forces you to touch almost every core skill in 3D, like modeling, lighting, materials, texturing, particle systems for the sprinkles, and rendering.
If you can make a donut that looks photorealistic, you understand the fundamentals.
Learning Blender properly takes weeks.
Last week, a tool called Omma launched and made a bold claim: describe anything and it builds it in seconds.
The team behind it is Spline, the 3D design platform used by 3 million+ designers. Their COO called it "as close to magic as it's ever felt to build digital experiences."
So I ran the Blender Donut test on it. But before that, some catchup on AI this week:
NEWS NEWS NEWS
TOOLS of the Week
1. Noon: A dual-canvas design tool where what you design is literally the production code, without any handoff or translation layer. Just raised $44M from First Round, Elevation Capital, and design leads at Stripe, OpenAI, and Apple.
2. Roger AI: An AI screen guide that watches what you're working on and walks you through tasks step by step works across Figma, Excel, Photoshop, or any software you're learning. Think of it as a live tutorial that follows your actual screen instead of a YouTube video you have to keep pausing.
3. Atomic: An open-source knowledge graph that connects your notes, articles, and research with semantic search, auto-tagging, and AI-generated wiki summaries with inline citations. Runs on Mac, iOS, self-hosted server, and integrates directly with Claude and Cursor via MCP.
Test 1: The Donut
The prompt:
A photorealistic glazed donut with chocolate icing dripping over the edges. Pink sprinkles scattered across the top. Sitting on a white ceramic plate on a wooden table. Warm cafe lighting from the top-left. Render it as a 3D interactive scene.
The icing is a problem. It's not sitting on top of the donut and is instead in the gap, where it shouldn't be.
Blender Guru would not approve.
For reference, this is the benchmark:
Then Omma suggested its own follow-up: add a coffee cup and some steam animation.
That made it worse.
Donut test: failed, spectacularly.
Test 2: The Teacup
The prompt:
A delicate white porcelain teacup with a thin rim and curved handle, sitting on a matching saucer. Earl grey tea inside, slightly translucent. Rising steam. Soft daylight from a window on the right. Render as an interactive 3D scene.
Omma built something with ambition here. It titled the output "Earl Grey · Morning", added hover tooltips which hover over the steam and says "Rising Steam — Water vapour at ~85°C, gentle convection."
It also put a button at the bottom: "Trigger Pour Animation."
The idea is great and the experience layer is the right instinct entirely.
The 3D execution however is not good. The cup clips through the table.
When you trigger the pour animation, the liquid pours into empty space and doesn't actually appear.
The steam particles are all over the place and the physics don't hold.
Great concept, bad output.
Test 3: The Website and the Deck
This is where the tone of this piece changes.
Website prompt:
Build an interactive landing page for a 3D animation studio called "Render & Co." Dark background, neon accent colours. Floating 3D geometric shapes in the hero. Three service tiles that animate on hover. Ready to publish.
Dark background, floating chrome and neon 3D geometric shapes, animated and clean nav with "Let's Talk" CTA.
This looks like a real studio website.
Deck prompt:
Build an interactive presentation for a 3D design tool called Forma. Five slides. Title slide with a 3D geometric hero. The problem — designers switching between too many tools. The solution — describe anything, one canvas. Three example creations. Call to action. Dark theme, smooth slide transitions.
Love the gradient typography (white fading into pink and purple) with a 3D geometric object floating in the hero.
It also built the slide navigation dots at the bottom (1/5) with badge labels.
This looks like a real product pitch you can click through.
Both of these came back from a single prompt and are publishable.
My Take
Omma has good instincts and bad execution.
The teacup moment for example: It named itself "Earl Grey · Morning", added hover tooltips explaining steam physics, put a pour animation button at the bottom.
The instinct to build an experience rather than just render an object, that's good thinking.
But then the cup clipped through the table and the liquid poured into empty space.
The 3D failures aren't minor bugs I'm willing to overlook, it is literally the pitch.
The website and the deck looked great. But Framer already builds great websites, and so does Bolt.
Omma isn't really competing with either of them. It's trying to be something different, a tool where 3D and UI live in the same canvas.
That's a solid category and it pays to be early. But they’re not there yet and I'd check back in six months.
Try it this weekend and reply with what you got. I might feature the best results next week.
Until next time,
Vaibhav 🤝
If you read till here, you might find this interesting
#AD 1
How Jennifer Aniston’s LolaVie brand grew sales 40% with CTV ads
The DTC beauty category is crowded. To break through, Jennifer Aniston’s brand LolaVie, worked with Roku Ads Manager to easily set up, test, and optimize CTV ad creatives. The campaign helped drive a big lift in sales and customer growth, helping LolaVie break through in the crowded beauty category.
#AD 2
Looking for a free tech newsletter trusted by the industry’s biggest names? Subscribe to The Current, a free daily tech newsletter written by Kim Komando to help you understand AI, keep up with tech news, and learn useful tips in just 5 minutes a day.
