dicky

Award-Winning Writer-Director | Screen • Stage • Commercials • AI | Raindance/Austin | Taipei • London • Tokyo | EN/中文/日本語 | Portfolio ⬇️

[dicky @dickychalmers]

From Steenbecks to Spaghetti Nodes. 🍝

I started my directing career cutting 16mm film on flatbeds. Today, I’m wrangling nodes in ComfyUI.

I’ve spent the past year going hybrid, merging 20 years of commercial directing with GenAI.

[dicky @dickychalmers]

Reframe Workflow

[dicky @dickychalmers]

Weapon of choice.

[dicky @dickychalmers]

Stop using GenAI to copy reality. Use it to break it. ⚡️

Made using Wan 2.2 Animate in #ComfyUI, and a custom node to reshape the pose data.

#wan22 #VFX #VFXBreakdown

[dicky @dickychalmers]

Wan 2.2 Animate is powerful, but it hates "weird" proportions.

I vibe-coded a custom #ComfyUI node to warp pose data, allowing me to transform a standard performance capture into a cinematic Slenderman.

#wan22 #VFX #VFXBreakdown #Slenderman

[dicky @dickychalmers]

@ComfyUI @Alibaba_Wan @GeminiApp @suno @Adobe For the Extended Director’s Cut and a deeper look at the world of Jon and Nathan, head over to YouTube: https://t.co/baLpzFBbMZ

[dicky @dickychalmers]

From Script to Screen (via Latent Space).

My new concept trailer for HALF-DREAM STATE is live.

We wrote a high-concept thriller about dual identities in Manchester.

Instead of just a deck, we built the world using GenAI.

This is how we’re pitching TV in 2026.

#AIFilmmaking

[dicky @dickychalmers]

Tech Stack:⚡️@ComfyUI / @Alibaba_Wan / @GeminiApp / @Suno / @Adobe

Full Project: https://t.co/QCxXNWIX5i

[dicky @dickychalmers]

Not to replace the craft, but to expand the physics of what’s possible.

Tech Stack: ⚡️ ComfyUI ⚡️ Wan / Flux / Qwen ⚡️Midjourney / Veo⚡️ LoRA / 3DGS

Full Portfolio: https://t.co/2WK2kdAgoa

#AIFilmmaking #ComfyUI #GenAI #VFX #Wan #Flux #Veo #Director #Filmmaking #Midjourney

[dicky @dickychalmers]

From Steenbecks to Spaghetti Nodes. 🍝

I started my directing career cutting 16mm film on flatbeds. Today, I’m wrangling nodes in ComfyUI.

I’ve spent the past year going hybrid, merging 20 years of commercial directing with GenAI.

[dicky @dickychalmers]

@hi_Rawmantic The detail is incredible. Is this using GSplats? Or traditional photogrammetry?

[dicky @dickychalmers]

@apalahstudio @gmi_cloud @D_studioproject Looking good. The sound quality seems to have improved a lot since WAN 2.5.

[dicky @dickychalmers]

@ChetiArt This is awesome, a great combination of SCAIL and Uni3c!

[dicky @dickychalmers]

Finally got round to testing Wan 2.2 Animate.

The most impressive thing is how aware the model is of the environmental lighting.

[dicky @dickychalmers]

Scream Queens

[dicky @dickychalmers]

In a half-dream state

[dicky @dickychalmers]

Recorded myself performing a Hamlet soliloquy → turned still images into moving characters with AI.
We’re not out of the uncanny valley yet, but the expressiveness is already impressive.

Images: Midjourney | Video: Wan2.1 | Lip-sync: InfiniteTalk

[dicky @dickychalmers]

Hamlet, lip-synced by InfiniteTalk.

[dicky @dickychalmers]

Framegrabs and BTS from the StarLux Airlines commercial. Shooting Taipei for Phoenix.

[dicky @dickychalmers]

New commercial I directed for StarLux Airlines to promote their Taipei-Phoenix route. Ad dressed up as a fake road trip movie trailer.

[dicky @dickychalmers]

These aren’t drone shots.
Just still photos taken from hotel windows in Dubai, and animated using GenAI.

I2V via Wan 2.1 in ComfyUI.

The fact it even animates traffic is kind of mind-blowing.