Milad

Building with AI • Exploring Web3
Learning by building & sharing the journey

[Milad @m1lad_az]

@giutheginger Solo building teaches you a lot. Always interesting to see how indie founders turn ideas into real things

[Milad @m1lad_az]

@Alisina_ai Same mindset here. I’ve learned the most from building, testing, and breaking things rather than just following the hype

[Milad @m1lad_az]

@themahis Great to connect! Exploring AI agents and vibe coding too. Looking forward to seeing your experiments

[Milad @m1lad_az]

@NYXA_web3 Agreed. My recent Codex tests showed me that failures are often better feedback than successful runs. They reveal where the process needs better constraints

[Milad @m1lad_az]

I’m Milad.

Building with AI and learning by doing

Exploring AI agents, vibe coding & workflows
Sharing experiments, what works, what breaks, and what I learn

Looking to connect with builders, developers, and people experimenting with AI

Let’s connect 👀

[kimiloo @Kimilooweb3]

I’m 28

@IThinkItsArt Ambassador, AI artist, and video creator

I create AI art, make videos, and occasionally vibe code

Here to connect with builders and people creating cool things with AI

lets connect 💜👀

[Milad @m1lad_az]

@Driscoll600 Exactly. The things that break usually teach the most. Testing and adjusting the workflow is where the real learning happens

[Milad @m1lad_az]

@Alisina_ai Hey Alisina! Glad to connect 👋 Excited to follow your AI journey

[Milad @m1lad_az]

@themahis Same here. The messy parts often show the real learning. Failed attempts reveal where the workflow needs better constraints

[Milad @m1lad_az]

@heyartzy Mostly scope issues. Too much freedom creates unnecessary changes, so tighter boundaries and smaller steps helped a lot

[Milad @m1lad_az]

@kamilbaranek_ Hey Kamil, happy to connect! Looking forward to following what you build 👀

[Milad @m1lad_az]

@RyanRayMartin Likewise, happy to connect! 👀

[Milad @m1lad_az]

@nickkirt Hi 👋 Nice to connect!

[Milad @m1lad_az]

@Kimilooweb3 Same here. The failures are often where the best lessons come from 👀

[Milad @m1lad_az]

@akfullstack Always interesting to connect with people experimenting with AI. Looking forward to seeing what you build

[Milad @m1lad_az]

@NYXA_web3 That’s what made the behavior predictable for me. Less room to improvise, easier to verify

[Milad @m1lad_az]

Tested Codex with tighter boundaries today.
One folder only. Plan first. Wait for approval.
It created time_now.py, ran it, and stopped.
No extra files. No cleanup.
Same model. Better boundaries.
Permission ≠ Precision
Scope → Plan → Execute

[Milad @m1lad_az]

@whereuncle Being able to inspect the context behind an answer makes personalization a lot easier to trust

[Milad @m1lad_az]

@whereuncle That’s a better reliability test. Staying inside scope matters as much as completing the task

[Milad @m1lad_az]

@Kimilooweb3 @IThinkItsArt I’m exploring AI agents and vibe coding through small experiments right now. Always good to connect with people building and creating with AI.

[Milad @m1lad_az]

@Xman_base_eth That’s probably the next test. I want to see whether the boundary is actually enforced or just respected by the model.

[Milad @m1lad_az]

@Alice0nX That’s a better test of accuracy. If the task is vague, the agent should ask before acting.