Iโve led teams from 2 to 16 and back down to 6. Here is the hardest lesson I learned about engineering with AI tools.
The engineers who thrived weren't the fastest coders. They were the ones who understood the "Constraint Space".
Leading both startups and overseeing engineering for established quant research has shown me a massive gap in how we talk about AI...
Everyone says: think strategically, let AI handle the tactical work.
But thereโs a massive gap between the two that most people are missing. Itโs what allowed a section thrive while others struggled.
The data backs this up: a study found that 73% of users accepted faulty AI output under time pressure. They could see the code, but they couldnโt see that the "proof shape" was wrong.
The future moat isn't knowing how to code.
Itโs the ability to occupy that constraint space - knowing when a solution is fundamentally flawed, even if it looks perfect on the surface.
AI is moving fast, but the community is moving faster.
A founder of two AI startups and lead at a Quant research firm, my feed needs more builders.
Looking to connect with those into:
๐ค Local AI & Edge Computing
๐ Quant Research & Data Science
โจ Vibe Coding & the future of Dev
๐ ๏ธ Full Stack / DevOps / UIUX
๐ Startups & Freelancing
If youโre shipping, experimenting, or over-optimizing - say hi below! ๐
@X can you help to connect me to real humans?
#BuildInPublic #AI #VibeCoding #Connect #Quant #Startups #TechCommunity
AI is moving fast, but the community is moving faster.
A founder of two AI startups and lead at a Quant research firm, my feed needs more builders.
Looking to connect with those into:
๐ค Local AI & Edge Computing
๐ Quant Research & Data Science
โจ Vibe Coding & the future of Dev
๐ ๏ธ Full Stack / DevOps / UIUX
๐ Startups & Freelancing
If youโre shipping, experimenting, or over-optimizing - say hi below! ๐
@X can you help to connect me to real humans?
#BuildInPublic #AI #VibeCoding #Connect #Quant #Startups #TechCommunity
AI is moving fast, but the community is moving faster.
A founder of two AI startups and lead at a Quant research firm, my feed needs more builders.
Looking to connect with those into:
๐ค Local AI & Edge Computing
๐ Quant Research & Data Science
โจ Vibe Coding & the future of Dev
๐ ๏ธ Full Stack / DevOps / UIUX
๐ Startups & Freelancing
If youโre shipping, experimenting, or over-optimizing - say hi below! ๐
@X can you help to connect me to real humans?
#BuildInPublic #AI #VibeCoding #Connect #Quant #Startups #TechCommunity
Iโve led teams from 2 to 16 and back down to 6. Here is the hardest lesson I learned about engineering with AI tools.
The engineers who thrived weren't the fastest coders. They were the ones who understood the "Constraint Space".
Leading both startups and overseeing engineering for established quant research has shown me a massive gap in how we talk about AI...
Everyone says: think strategically, let AI handle the tactical work.
But thereโs a massive gap between the two that most people are missing. Itโs what allowed a section thrive while others struggled.
The data backs this up: a study found that 73% of users accepted faulty AI output under time pressure. They could see the code, but they couldnโt see that the "proof shape" was wrong.
The future moat isn't knowing how to code.
Itโs the ability to occupy that constraint space - knowing when a solution is fundamentally flawed, even if it looks perfect on the surface.
AI is moving fast, but the community is moving faster.
A founder of two AI startups and lead at a Quant research firm, my feed needs more builders.
Looking to connect with those into:
๐ค Local AI & Edge Computing
๐ Quant Research & Data Science
โจ Vibe Coding & the future of Dev
๐ ๏ธ Full Stack / DevOps / UIUX
๐ Startups & Freelancing
If youโre shipping, experimenting, or over-optimizing - say hi below! ๐
@X can you help to connect me to real humans?
#BuildInPublic #AI #VibeCoding #Connect #Quant #Startups #TechCommunity
AI is moving fast, but the community is moving faster.
A founder of two AI startups and lead at a Quant research firm, my feed needs more builders.
Looking to connect with those into:
๐ค Local AI & Edge Computing
๐ Quant Research & Data Science
โจ Vibe Coding & the future of Dev
๐ ๏ธ Full Stack / DevOps / UIUX
๐ Startups & Freelancing
If youโre shipping, experimenting, or over-optimizing - say hi below! ๐
@X can you help to connect me to real humans?
#BuildInPublic #AI #VibeCoding #Connect #Quant #Startups #TechCommunity
AI is moving fast, but the community is moving faster.
A founder of two AI startups and lead at a Quant research firm, my feed needs more builders.
Looking to connect with those into:
๐ค Local AI & Edge Computing
๐ Quant Research & Data Science
โจ Vibe Coding & the future of Dev
๐ ๏ธ Full Stack / DevOps / UIUX
๐ Startups & Freelancing
If youโre shipping, experimenting, or over-optimizing - say hi below! ๐
@X can you help to connect me to real humans?
#BuildInPublic #AI #VibeCoding #Connect #Quant #Startups #TechCommunity
AI is moving fast, but the community is moving faster.
A founder of two AI startups and lead at a Quant research firm, my feed needs more builders.
Looking to connect with those into:
๐ค Local AI & Edge Computing
๐ Quant Research & Data Science
โจ Vibe Coding & the future of Dev
๐ ๏ธ Full Stack / DevOps / UIUX
๐ Startups & Freelancing
If youโre shipping, experimenting, or over-optimizing - say hi below! ๐
@X can you help to connect me to real humans?
#BuildInPublic #AI #VibeCoding #Connect #Quant #Startups #TechCommunity
AI is moving fast, but the community is moving faster.
A founder of two AI startups and lead at a Quant research firm, my feed needs more builders.
Looking to connect with those into:
๐ค Local AI & Edge Computing
๐ Quant Research & Data Science
โจ Vibe Coding & the future of Dev
๐ ๏ธ Full Stack / DevOps / UIUX
๐ Startups & Freelancing
If youโre shipping, experimenting, or over-optimizing - say hi below! ๐
@X can you help to connect me to real humans?
#BuildInPublic #AI #VibeCoding #Connect #Quant #Startups #TechCommunity
Why do your longest prompts give the worst AI answers?
When you overload the AI's context window, the signal-to-noise ratio drops and hallucinations spike. Stop dumping everything into one prompt. Instead, find the baseline gap and only fill that missing piece.
For the full a2a course, github example link and other excited videos, see the comments