Meet Gondaliya

Building RizzGen - AI video creation with real creator control.
Agent drafts. You direct.
No one-prompt magic, just better workflow.

[Meet Gondaliya @meet__g]

Ever wondered how advanced agents like OpenClaw work under the hood? It's the core reasoning loop.

Last week, I used a static router. Worked, but zero mid-loop flexibility.

So I scrapped it. I built a framework-free Python agent using JIT Skill Loading instead.

[Meet Gondaliya @meet__g]

@buildwithyasin building https://t.co/hfXXEUx3sl - AI video creation for creators who want creative control, not one-click automation.

[Meet Gondaliya @meet__g]

@elonmusk If AGI cracks practical Zero-Point Energy, does antimatter become irrelevant or does ZPE make antimatter so cheap that it accelerates humanity toward a Type II civilization?

[Meet Gondaliya @meet__g]

@ionleu https://t.co/hfXXEUx3sl - controllable video creation

[Meet Gondaliya @meet__g]

Obviously not a scientific comparison.

Just a pattern I've noticed while using both.

Anyone else seeing the same thing?

[Meet Gondaliya @meet__g]

I've been using Opus 4.8 in Claude Code and GPT-5.5 in Codex recently.

One pattern I think I'm noticing:

When uncertainty appears, they seem to reduce it in very different ways.

[Meet Gondaliya @meet__g]

Claude Code often feels like:

"Let's help the agent understand this better."

More context.
Better instructions.
Prompt refinements.

[Meet Gondaliya @meet__g]

Started from one prompt: "A tiny toy robot secretly travels through ant tunnels beneath a suburban backyard..." Here's the full workflow + final output: https://t.co/TzmxAEh5Pw

[Meet Gondaliya @meet__g]

Single-prompt AI video breaks fast for storytelling. Every scene drifts. Characters shift. You're regenerating the same shot 12 times with nothing to show for it.

[Meet Gondaliya @meet__g]

So I tried structuring the film first instead. idea -> storyboard -> visual planning -> continuity -> final film.

[Meet Gondaliya @meet__g]

@acadictive @BatsouElef Right now mostly optimizing for:

- getting users to their first good video
- making the workflow feel controllable/editable
- reducing the “AI slot machine” feeling

Trying to make AI video feel more directed and iterative instead of pure one-shot generation.

[Meet Gondaliya @meet__g]

@manojdotdev

[Meet Gondaliya @meet__g]

The most underrated tax of being a solo founder:
You are the context layer.

Every decision, every AI conversation, every strategy call you manually reconstruct what your company is, what you've tried, what's broken.

No tool remembers. You do.

Anyone else feel this?

[Meet Gondaliya @meet__g]

@i_mika_el Fair point. Switching the frame from “here’s how we’re built” to “here’s the workflow that keeps failing you.”

Would love to try your tool once you open it up.

[Meet Gondaliya @meet__g]

@i_mika_el That actually sounds pretty powerful.

I think my issue right now is less about finding creators on LinkedIn and more about communicating why RizzGen is different from typical one-prompt AI video tools.

[Meet Gondaliya @meet__g]

@i_mika_el I tried briefly, but will experiment more since your point makes sense.

Smaller niche subreddits probably have much higher intent/conversion even with lower reach because the audience is already very aligned with the problem.

[Meet Gondaliya @meet__g]

@i_mika_el Cool, I tried LinkedIn a bit but probably wasn’t approaching it the right way yet.

Curious to see how you’re solving outreach there.

[Meet Gondaliya @meet__g]

@i_mika_el Got 5 signups from that Reddit post in under 12 hours from ~5.5k reach.

What surprised me was I ran small-budget Meta ads for a week before that, around 75k reach total and only got 8 signups.

[Meet Gondaliya @meet__g]

@i_mika_el One example that worked well for me:

https://t.co/mnffvlnXxx