Get Comfy with LAIA
Preview Our First AI for Creative Leaders Session:
Let Your Curiosity Get The Better Of You!
Hey {{first_name | Friend}} 👋
Adam at your service.
I need to be honest here: I didn’t know generative AI could go this far.
Every week at Lighthouse AI Academy, I’m reminded: this goes waaay beyond faster outputs; it’s about changing how we think, learn, and design.
That’s what excites me about our newest program, AI for Creative Leaders (Cohort 3).
It’s not a technical bootcamp. It’s a place where filmmakers, architects, designers, and creative entrepreneurs gather to rethink their workflows and together, uncover how AI helps us move from chaos to clarity.
Last week, we kicked things off with our first session.
Below, I’ll share the highlights: what our Course Director, Arminas (Armin) learned building viral projects, how our co-founder, Nejc, reframed the role of AI, and how the cohort itself is already shaping conversations about creativity, law, and workflow design.
Ready for a tantalizing taster of what we do?
Let’s go!
🗽 Inside the Cohort: AI for Creative Leaders Session 1 Highlights
And so it begins…
Step 1 — Student Introductions
Anastasia — Design lead at a global ride-hailing company: building AI pipelines for her design team.
Giedre — Photographer (Vilnius): shifting into AI-driven visuals.
Juan — Solution Architect Director at Epic Games (Unreal Engine): experimental short on beekeeping in Colombia and GenAI ↔ Unreal integration.
John — Freelance VFX supervisor (Oregon): AI spec watch commercial with consistent character design.
Vishal — Designer (India): supernatural/horror film; deep dive into ComfyUI.
Johan — Creative director at one of Sweden’s largest retailers: scalable fashion/product content with higher execution quality.
Franz — Filmmaker (Guatemala/Mexico): animated shorts + AR for events.
Jonny — Ex-Nickelodeon veteran (NYC), founder of an AI animation startup: AI rap music video.
Leandro — VFX supervisor (San Francisco): AI in production pipelines.
Fritte — Cinematographer/animator/producer: AI ads for a brewery; one-person creative lab.
Dennis — VFX lighting/look-dev: deciding between a render-speed toolkit and a short cinematic.
Henri — Senior creative specialist (Singapore): branding/marketing workflows with AI.
Tanja — Director/DP (Berlin): short film with a consistent character.
Kevin — Virtual Production Supervisor at Epic Games: AI + Unreal + personal photo archive workflows.
Giacomo — Freelance motion designer (Italy): strong in ComfyUI; expanding into LLMs and audio AI.
A Few Of The Amazing Crew Of Students That Form Our 3rd Cohort:
Henri
Giedre
Jonny
Anastasia
Step 2 — Setting the Stage
Nejc (“Nates”), Lighthouse co-founder, welcomed ~15 participants and framed a 12-week journey that’s about workflow design, collaboration, and mindset, above and beyond tools.
Armin (course director; 12+ years in advertising/animation; pivoted to AI 3 years ago) introduced himself, too.
While early lessons may feel basic, they connect into a complete system.
You can’t integrate AI tools overnight. The best strategy is to learn daily, connect the dots over time, and keep moving forward.”
Step 3 — How the Course Works
Discord is the main hub (projects, questions, feedback).
Journaling each week (experiments, obstacles, learnings).
Guest mentors: deep dives across LLMs, video, and training.
Make a thing: every student defines a project (film, ad, workflow, startup idea).
This is a practical course. Pick a project—short story, workflow, or product—and let us help you build it.
Step 4 — Armin’s Mindset Framework
Iteration is the engine: generate fast, learn fast.
All good things take time
Consistency beats talent: daily reps compound.
Keep showing up
Show your work: visibility creates opportunities (War of Art; Show Your Work).
Positivity & community: collaboration accelerates growth.
The Dip (Seth Godin): push through low-motivation phases.
Overcoming the doubts and dips
Traditional craft still matters: pre-prod → prod → post; AI enhances, doesn’t replace.
Step 5 — Armin’s Turning Point
Armin left client work, got scammed by a “startup,” and was down to ~$1,000. He invested in a Runway unlimited plan, generated obsessively, built a GenAI portfolio in 3 weeks, and posted Nike/Porsche/Gucci-inspired videos, leading to recognition and inbound offers.
Ship fast, generate a lot, and don’t judge the first results too harshly. Creativity grows in volume.
Step 6 — Image-to-Video vs Text-to-Video (Porsche Example)
Prompted by Giacomo’s question, Armin explained he used both. Image-to-video felt too static; text-to-video delivered dynamic drone-style shots (e.g., the couple shot). He combined both. At the time, there were no good tutorials, so trial and error was essential.
Advice: Pick one main platform with unlimited credits so you can explore without worrying about costs.
Plug & Play
Step 7 — Legal & Copyright
Nates: Law is catching up (active lawsuits exist). Think like Photoshop: spec is one thing; pro work requires clearance. You’ll help set internal policies and precedents with clients.
Johnny: Many clients think AI is “just typing.” Run user agreements through an LLM to flag ownership/IP and model-training clauses. Example raised: some services allow commercial use but still train on outputs.
Armin: Avoided using brand assets as inputs (e.g., Gucci); MidJourney for text-to-image only. Be extra careful when inputting real brand materials.
Step 8 — Creative Principles Reinforced
Armin’s mantra: Create → get feedback → iterate → repeat. Don’t overthink first prompts, generate volume. He once burned 100k Freepik credits in a week and was glad he did: the learning was worth it.
Just keep swimming
Inspiration & community: LinkedIn, Escape Cinema; save strong pieces and study frame-by-frame. Follow creators/mentors like Mauricio Tonon, Luca Tisler (Academy co-founder), Benjamin Bertram, and note the Dor Brothers (satire raises legal questions).
Step 9 — Overwhelm & Model Choices
Fritte’s concern: New models weekly, fear of falling behind.
Armin’s answer: Commit to one platform (Freepik, Krea, Weavy). New models get integrated anyway. Better to master one with an unlimited plan than scatter across five basics. Example: a $125 VO3 month enabled a Dior spec ad → agency offers.
Step 10 — Workflow & Team Integration
Johan’s challenge: scaling for a large team.
Armin: Treat it case-by-case; run a defined R&D phase, present concrete results, and build trust. Don’t hand teams too many tools — that’s chaos.
Nejc: You’ll adopt a developer mindset. Separate production (stable, repeatable) from development (experimental, breakable). Always budget R&D.
Step 11 — Nejc’s Core AI Principles
AI is a language transformer: text, image, video, audio — even Wi-Fi → 3D and MRI → image.
Write it into existence
Success = communication and problem formulation (break big asks into small steps).
Director’s mindset: understand the whole pipeline; think laterally.
Taking charge
Have fun: experiment; mistakes are data.
AI to enhance, not replace
Build skills in communication, critical thinking, curation, and workflow management.
Keep growing
Not a magic box: outcomes depend on inputs.
No substitute for the human touch
Step 12 — Technical Primer (fast)
Models are trained on datasets (often billions of items).
Inner workings
Tokens are basic units (e.g., “ice”, “cream”).
Parameters shape behavior (rules/weights).
Training often uses “teacher” models guiding “student” models.
Hardware matters (GPUs, power, cooling).
Ongoing challenges: bias, dataset gaps, copyright/privacy.
Step 13 — Next Steps
Your mind will fundamentally change over 12 weeks.
Define your project: film/ad/workflow/product.
Break it down: script → visuals → production → post; mark where AI fits.
Use the Q&A each week (tech, legal, process, ethics).
Discord:
General — share work.
Q&A questions — post in advance.
Personal channels — project threads & feedback.
Student section — alumni/community.
Platform: dashboard for sessions, materials, and journaling.
Respect to everyone for staying until the end. Let’s keep energy high.
🧠 Homework: Keeping the Momentum
Define your project on one page. Title, goal, audience, and success criteria.
Break it down (“problem formulation”). Script/plot → visuals/moodboard → production steps → post. Be explicit about where AI fits.
Pick your lane (for now). Choose one image tool (e.g., Freepik) and one workflow tool (e.g., Weavy or ComfyUI). Don’t over-optimize—stability first, novelty later.
Create → feedback → iterate. Post 3–5 WIPs in Discord. Note what improved each pass.
Journal. What worked? What bottleneck or question should we bring to the Wednesday Q&A?
Iteration = mastery.
We should run an open-door policy. If you have something to share, whether work, developments, or experiments, we’ll highlight it across our channels: LinkedIn, articles, and newsletter.
🔦 Industry Spotlight: ComfyUI’s $17M Leap into the Cloud
Comfy announced a $17M round (Pace Capital, Chemistry, Abstract Ventures, others) to double down on open-source ComfyUI, stabilizing the custom-node ecosystem, refining the interface, supporting the newest models, and building the best Cloud service.
A heavenly OS
Their philosophy is blunt and refreshing: “Open source must win.” They’re not building a walled garden; they’re aiming for the OS of creative AI.
Comfy Cloud private beta: A managed ComfyUI in the cloud:
Powerful server GPUs, popular models pre-loaded.
Growing library of custom nodes/extensions.
Feature parity with open-source ComfyUI; local remains free forever.
Future pricing will be based on GPU usage (not your idle workflow-building time).
Exploring revenue share for custom-node authors.
ComfyUI will always be free to run locally… Cloud is how we sustain long-term development.
Why it matters for us: This sits perfectly with our course’s hybrid philosophy. Keep local for control and R&D; use cloud when you need speed, scale, or team access, without fighting dependencies.
If you’d like to reach for the sky and be among the first to get your foot through the proverbial pearly gates for Comfy Cloud, sign up here.
This is a huge step for AI-kind: making professional-grade AI workflows stable, accessible, and community-driven.
🌄 Closing Sentiments
This first week's session was a stark reminder that:
AI doesn’t replace your craft. It extends it.
And with the right mindset—iteration, journaling, sharing—you turn scattered experiments into systems.
That’s the genuine skill we’re building here, and it’s only just beginning.
If you’re serious about shaping your creative future with AI, our cohorts are designed to walk that path with you. 🔥
That’s it for now: thanks for reading and for building this new era with us.
AI is moving fast (every single week), but that means there’s constant space to create, too.
At Lighthouse, we’re here to make sure you don’t just keep up, but actually lead.
Step by step, node by node — and away we go!
Keep creating and always remember to have fun.
— Adam & the Lighthouse AI Academy Team ☀️
Small Team, Big Dreams