Positive Delusion, Bounded by Learning
Assume everything is possible until proven wrong — but keep learning what's actually possible.
- Difficulty
- Easy
- Time to result
- ~ongoing to results
- Steps
- 4
- Confidence
- 85%
A working mindset for AI tools: come in 'positively delusional,' assuming anything is buildable until the tool proves otherwise, because not knowing the supposed limits is exactly how non-technical people build things experts say are impossible. The counterweight is disciplined self-awareness — continuously learning through the agent layer what is genuinely possible right now, so your delusion doesn't burn a week brute-forcing something that isn't yet feasible.
Origin
Lazar Jovanovic's framing of why a non-technical background is an advantage in vibe coding — 'people like me don't know that they are not supposed to be building XYZ, and that's how we actually are able to build it.'
Core principles
- 01Positive delusion: assume everything is possible until proven wrong — this unbiased optimism unlocks builds experts dismiss.
- 02Self-awareness is the counterweight: know you need to get better for the delusion to become reality.
- 03The agentic layer is your reality check — agents can browse, read, and reason, and will tell you when something is genuinely undoable right now.
- 04Delusion without learning is a trap: pushing at something truly impossible wastes your scarcest resource — time.
How to run it
- 1
Default to 'this is possible'
Approach every idea assuming it can be built until the tool concretely proves it can't, rather than pre-filtering by what you believe the stack allows.
Pro tip A non-technical vantage point is an asset here — you aren't biased by knowing what 'shouldn't' work.
- 2
Just try it before accepting 'impossible'
When people say something can't be done, attempt it in the tool anyway. Non-technical builders shipped Chrome extensions, desktop apps, and in-app video generation this way — before those were 'supported.'
- 3
Use the agent conversation as a feasibility check
Lean on the agentic layer's reasoning and web-browsing to tell you when something is genuinely blocked (e.g. a missing API) versus merely unfamiliar, and treat each answer as a learning opportunity.
Watch out Don't ignore a factual 'this is undoable at the moment' — delusion must yield to real constraints or you'll waste days.
- 4
Time-box the brute force
If the tool signals a hard blocker, stop pushing and wait for the capability to arrive rather than forcing a workaround that may land natively next week.
Watch out Lazar spent a week brute-forcing a ChatGPT image-gen wrapper before the API existed; once it shipped he rebuilt the app in 30 seconds. Check whether you're fighting a temporary gap.
In the wild
When technical community members explained why Lovable couldn't build Chrome extensions (different stack, React, etc.), non-technical users like Lazar simply prompted Lovable to build one from an app — and it worked. Others built desktop apps and generated video in-app before those were features.
→ Builds that experts said shouldn't be possible shipped anyway, validating the unbiased 'try it' posture.
When OpenAI released native image generation in ChatGPT, Lazar tried for a week to build a Lovable wrapper — not realizing there was no API yet.
→ A week later the API shipped and he built the app in 30 seconds; the failure taught him to check real feasibility via the agent layer before brute-forcing.
Common mistakes
Being delusional without self-awareness
Optimism alone isn't enough. Without the parallel discipline of learning what's genuinely possible, you'll pour time into something the tools can't yet do — Lazar's week-long image-gen dead end is the cautionary case.
Letting technical knowledge pre-filter your ideas
Knowing 'why it shouldn't work' talks you out of builds that actually would. The technical bias that experts carry is precisely what stops them from attempting the things non-technical builders ship.
Is it for you?
Best for
Non-technical or newly-technical builders intimidated by experts telling them what AI tools 'can't' do.
Not ideal for
High-stakes, regulated, or infrastructure work where unfounded optimism about feasibility carries real cost and rigor must lead.
From the transcript
“You have to come with this delusion that absolutely everything is possible until proven wrong.”
“people like me don't know that they are not supposed to be building XYZ and that's how we actually are able to build it”
“I spent at least a week trying to brute my force brute force my way into making this work instead of just waiting for another…”
From the episode
The rise of the professional vibe coder (a new AI-era job)
Lazar Jovanovic (Professional Vibe Coder)