Ishaan Sehgal
@ishaansehgal.bsky.social
Co-Founder @ Omnara (YC S25) | ex-Microsoft & Roblox | Building the AI-agent command center
Building AI agents isn't about replacing people
It's about removing the constraints that kept solo builders stuck at "solo"
The friction that made scale impossible without a team
Now one person can operate like ten
That's not automation. that's amplification
It's about removing the constraints that kept solo builders stuck at "solo"
The friction that made scale impossible without a team
Now one person can operate like ten
That's not automation. that's amplification
January 16, 2026 at 5:14 PM
Building AI agents isn't about replacing people
It's about removing the constraints that kept solo builders stuck at "solo"
The friction that made scale impossible without a team
Now one person can operate like ten
That's not automation. that's amplification
It's about removing the constraints that kept solo builders stuck at "solo"
The friction that made scale impossible without a team
Now one person can operate like ten
That's not automation. that's amplification
The architecture everyone's building toward just became obvious
Skills = reusable recipes
Sub-agents = parallel workers
MCPs = system connectors
Every major AI platform is converging on this exact pattern
Is this the first time we've actually agreed on how to build reliable AI workflows?
Skills = reusable recipes
Sub-agents = parallel workers
MCPs = system connectors
Every major AI platform is converging on this exact pattern
Is this the first time we've actually agreed on how to build reliable AI workflows?
January 8, 2026 at 5:14 PM
The architecture everyone's building toward just became obvious
Skills = reusable recipes
Sub-agents = parallel workers
MCPs = system connectors
Every major AI platform is converging on this exact pattern
Is this the first time we've actually agreed on how to build reliable AI workflows?
Skills = reusable recipes
Sub-agents = parallel workers
MCPs = system connectors
Every major AI platform is converging on this exact pattern
Is this the first time we've actually agreed on how to build reliable AI workflows?
the AI startup playbook inverted
"serious" founders still writing whitepapers
meanwhile some kid cloned an open-source template, slapped Claude on it, and has 10k users
stealth mode is just slow mode now
"serious" founders still writing whitepapers
meanwhile some kid cloned an open-source template, slapped Claude on it, and has 10k users
stealth mode is just slow mode now
January 7, 2026 at 4:59 PM
the AI startup playbook inverted
"serious" founders still writing whitepapers
meanwhile some kid cloned an open-source template, slapped Claude on it, and has 10k users
stealth mode is just slow mode now
"serious" founders still writing whitepapers
meanwhile some kid cloned an open-source template, slapped Claude on it, and has 10k users
stealth mode is just slow mode now
YC-backed AI founders are hitting a brutal paradox
Ship in 24 hours with AI tools
Explode by week 4 from tech debt
Winners use AI to execute faster while keeping architectural control
Losers mistake prompting for product thinking
Ship in 24 hours with AI tools
Explode by week 4 from tech debt
Winners use AI to execute faster while keeping architectural control
Losers mistake prompting for product thinking
January 3, 2026 at 4:52 PM
YC-backed AI founders are hitting a brutal paradox
Ship in 24 hours with AI tools
Explode by week 4 from tech debt
Winners use AI to execute faster while keeping architectural control
Losers mistake prompting for product thinking
Ship in 24 hours with AI tools
Explode by week 4 from tech debt
Winners use AI to execute faster while keeping architectural control
Losers mistake prompting for product thinking
the most successful founders i know still respond to customer support tickets
meanwhile everyone else is too busy "building their personal brand" to actually build anything
that's the whole game
meanwhile everyone else is too busy "building their personal brand" to actually build anything
that's the whole game
January 2, 2026 at 4:16 PM
the most successful founders i know still respond to customer support tickets
meanwhile everyone else is too busy "building their personal brand" to actually build anything
that's the whole game
meanwhile everyone else is too busy "building their personal brand" to actually build anything
that's the whole game
Most people are building AI agents wrong
They're obsessed with complex orchestration frameworks
The real unlock is making AI accessible enough that non-developers can become unstoppable builders
They're obsessed with complex orchestration frameworks
The real unlock is making AI accessible enough that non-developers can become unstoppable builders
December 30, 2025 at 4:10 PM
Most people are building AI agents wrong
They're obsessed with complex orchestration frameworks
The real unlock is making AI accessible enough that non-developers can become unstoppable builders
They're obsessed with complex orchestration frameworks
The real unlock is making AI accessible enough that non-developers can become unstoppable builders
AI made everyone a coder
Now companies are desperate for something else entirely
They don't need more people who can prompt ChatGPT
They need architects who can build AI systems that don't fall apart in production
The gap between "works in demo" and "works at scale" is where the real money is
Now companies are desperate for something else entirely
They don't need more people who can prompt ChatGPT
They need architects who can build AI systems that don't fall apart in production
The gap between "works in demo" and "works at scale" is where the real money is
December 29, 2025 at 4:13 PM
AI made everyone a coder
Now companies are desperate for something else entirely
They don't need more people who can prompt ChatGPT
They need architects who can build AI systems that don't fall apart in production
The gap between "works in demo" and "works at scale" is where the real money is
Now companies are desperate for something else entirely
They don't need more people who can prompt ChatGPT
They need architects who can build AI systems that don't fall apart in production
The gap between "works in demo" and "works at scale" is where the real money is
Everyone's debating agents vs workflows in AI
Here's what actually matters:
Can you ship it? Can you maintain it? Can users actually access it?
The most sophisticated agentic system is worthless if it lives in your localhost
Distribution beats architecture every single time
Here's what actually matters:
Can you ship it? Can you maintain it? Can users actually access it?
The most sophisticated agentic system is worthless if it lives in your localhost
Distribution beats architecture every single time
December 27, 2025 at 5:14 PM
Everyone's debating agents vs workflows in AI
Here's what actually matters:
Can you ship it? Can you maintain it? Can users actually access it?
The most sophisticated agentic system is worthless if it lives in your localhost
Distribution beats architecture every single time
Here's what actually matters:
Can you ship it? Can you maintain it? Can users actually access it?
The most sophisticated agentic system is worthless if it lives in your localhost
Distribution beats architecture every single time
The most exciting AI builders aren't coding in isolation anymore.
They're non-technical visionaries who vibe code a prototype, then collaborate with engineers to ship production systems.
AI just democratized who gets to be a founder.
They're non-technical visionaries who vibe code a prototype, then collaborate with engineers to ship production systems.
AI just democratized who gets to be a founder.
December 26, 2025 at 4:10 PM
The most exciting AI builders aren't coding in isolation anymore.
They're non-technical visionaries who vibe code a prototype, then collaborate with engineers to ship production systems.
AI just democratized who gets to be a founder.
They're non-technical visionaries who vibe code a prototype, then collaborate with engineers to ship production systems.
AI just democratized who gets to be a founder.
Dear builders,
I want to connect with:
AI founders
Coding hackers
YC alums
Tech disruptors
Product makers
Voice interface designers
Mobile devs
Everyday engineers
If that's you, say hi. Let's build smarter together.
I want to connect with:
AI founders
Coding hackers
YC alums
Tech disruptors
Product makers
Voice interface designers
Mobile devs
Everyday engineers
If that's you, say hi. Let's build smarter together.
December 22, 2025 at 5:14 PM
Dear builders,
I want to connect with:
AI founders
Coding hackers
YC alums
Tech disruptors
Product makers
Voice interface designers
Mobile devs
Everyday engineers
If that's you, say hi. Let's build smarter together.
I want to connect with:
AI founders
Coding hackers
YC alums
Tech disruptors
Product makers
Voice interface designers
Mobile devs
Everyday engineers
If that's you, say hi. Let's build smarter together.
what founders are hiring for now:
- can you ship with cursor
- have you built with claude
- do you actually code or just prompt
what they're NOT hiring for:
- read every AI paper
- can explain transformers
- went to that AI conference
talk less, build more
- can you ship with cursor
- have you built with claude
- do you actually code or just prompt
what they're NOT hiring for:
- read every AI paper
- can explain transformers
- went to that AI conference
talk less, build more
December 18, 2025 at 5:14 PM
what founders are hiring for now:
- can you ship with cursor
- have you built with claude
- do you actually code or just prompt
what they're NOT hiring for:
- read every AI paper
- can explain transformers
- went to that AI conference
talk less, build more
- can you ship with cursor
- have you built with claude
- do you actually code or just prompt
what they're NOT hiring for:
- read every AI paper
- can explain transformers
- went to that AI conference
talk less, build more
what VCs think gives big tech the AI edge:
- better models
- smarter engineers
- more compute
what ACTUALLY gives them the edge:
- 3 billion users already logged in
- distribution you can't buy
- they don't need you to download anything
- better models
- smarter engineers
- more compute
what ACTUALLY gives them the edge:
- 3 billion users already logged in
- distribution you can't buy
- they don't need you to download anything
December 16, 2025 at 5:14 PM
what VCs think gives big tech the AI edge:
- better models
- smarter engineers
- more compute
what ACTUALLY gives them the edge:
- 3 billion users already logged in
- distribution you can't buy
- they don't need you to download anything
- better models
- smarter engineers
- more compute
what ACTUALLY gives them the edge:
- 3 billion users already logged in
- distribution you can't buy
- they don't need you to download anything
The future of developer hiring:
"We need someone who [built something impressive]!"
Actually means:
"We need someone who [can prompt Claude to build something impressive]!"
The bar didn't lower.
It just moved to a place we don't understand yet.
"We need someone who [built something impressive]!"
Actually means:
"We need someone who [can prompt Claude to build something impressive]!"
The bar didn't lower.
It just moved to a place we don't understand yet.
December 15, 2025 at 5:14 PM
The future of developer hiring:
"We need someone who [built something impressive]!"
Actually means:
"We need someone who [can prompt Claude to build something impressive]!"
The bar didn't lower.
It just moved to a place we don't understand yet.
"We need someone who [built something impressive]!"
Actually means:
"We need someone who [can prompt Claude to build something impressive]!"
The bar didn't lower.
It just moved to a place we don't understand yet.
The biggest startup killer isn't competition
It's your advisor saying "be realistic"
Your investor suggesting you "pivot to something safer"
Your cofounder questioning this is "too ambitious"
They mean well but they're wrong
Knowing when to nod and ignore is the actual skill
It's your advisor saying "be realistic"
Your investor suggesting you "pivot to something safer"
Your cofounder questioning this is "too ambitious"
They mean well but they're wrong
Knowing when to nod and ignore is the actual skill
December 10, 2025 at 6:14 PM
The biggest startup killer isn't competition
It's your advisor saying "be realistic"
Your investor suggesting you "pivot to something safer"
Your cofounder questioning this is "too ambitious"
They mean well but they're wrong
Knowing when to nod and ignore is the actual skill
It's your advisor saying "be realistic"
Your investor suggesting you "pivot to something safer"
Your cofounder questioning this is "too ambitious"
They mean well but they're wrong
Knowing when to nod and ignore is the actual skill
One of my biggest lessons this year is to stop fighting uphill battles
If your sales cycle is 18 months, you're building the wrong way
Work backwards from your network instead of forwards from theory
If your sales cycle is 18 months, you're building the wrong way
Work backwards from your network instead of forwards from theory
December 9, 2025 at 5:14 PM
One of my biggest lessons this year is to stop fighting uphill battles
If your sales cycle is 18 months, you're building the wrong way
Work backwards from your network instead of forwards from theory
If your sales cycle is 18 months, you're building the wrong way
Work backwards from your network instead of forwards from theory
The best founder communities respect opportunity cost
If your event is just vibes, you're wasting builder time
If your event is just vibes, you're wasting builder time
December 8, 2025 at 5:24 PM
The best founder communities respect opportunity cost
If your event is just vibes, you're wasting builder time
If your event is just vibes, you're wasting builder time
Regulated industry founders think they have:
• A 6-month compliance checkbox before scaling
• A legal problem they can solve after product-market fit
What they actually have:
• 18+ months of existential bureaucracy
• A company-killer that starts the day after they raise
• A 6-month compliance checkbox before scaling
• A legal problem they can solve after product-market fit
What they actually have:
• 18+ months of existential bureaucracy
• A company-killer that starts the day after they raise
December 6, 2025 at 5:14 PM
Regulated industry founders think they have:
• A 6-month compliance checkbox before scaling
• A legal problem they can solve after product-market fit
What they actually have:
• 18+ months of existential bureaucracy
• A company-killer that starts the day after they raise
• A 6-month compliance checkbox before scaling
• A legal problem they can solve after product-market fit
What they actually have:
• 18+ months of existential bureaucracy
• A company-killer that starts the day after they raise
reddit right now:
- AI spam everywhere
- mods can't keep up
- vibe is gone
what's even the point
- AI spam everywhere
- mods can't keep up
- vibe is gone
what's even the point
December 6, 2025 at 1:14 AM
reddit right now:
- AI spam everywhere
- mods can't keep up
- vibe is gone
what's even the point
- AI spam everywhere
- mods can't keep up
- vibe is gone
what's even the point
Watching people 10x their output when you give them real responsibility early.
Not "let's see how they do with small tasks first" - actual ownership, actual stakes.
The upside of betting on raw potential over perfect resumes.
Not "let's see how they do with small tasks first" - actual ownership, actual stakes.
The upside of betting on raw potential over perfect resumes.
December 5, 2025 at 5:04 PM
Watching people 10x their output when you give them real responsibility early.
Not "let's see how they do with small tasks first" - actual ownership, actual stakes.
The upside of betting on raw potential over perfect resumes.
Not "let's see how they do with small tasks first" - actual ownership, actual stakes.
The upside of betting on raw potential over perfect resumes.
Young founders sacrifice everything. social life, stability, sleep. for their mission
Why do so many "founder communities" waste their time with surface-level BS instead of actually helping them build?
Why do so many "founder communities" waste their time with surface-level BS instead of actually helping them build?
December 3, 2025 at 5:04 PM
Young founders sacrifice everything. social life, stability, sleep. for their mission
Why do so many "founder communities" waste their time with surface-level BS instead of actually helping them build?
Why do so many "founder communities" waste their time with surface-level BS instead of actually helping them build?
Most accelerators say they're building "community"
YC built a cult
The difference?
YC founders help each other 10 years later for free
Everyone else ghosts you after demo day
YC built a cult
The difference?
YC founders help each other 10 years later for free
Everyone else ghosts you after demo day
December 1, 2025 at 5:04 PM
Most accelerators say they're building "community"
YC built a cult
The difference?
YC founders help each other 10 years later for free
Everyone else ghosts you after demo day
YC built a cult
The difference?
YC founders help each other 10 years later for free
Everyone else ghosts you after demo day
Leak: OpenAI is building an Ad Network inside ChatGPT. "Search Ads" strings found in latest Android Beta
November 30, 2025 at 11:06 PM
Leak: OpenAI is building an Ad Network inside ChatGPT. "Search Ads" strings found in latest Android Beta
YouTube now: Dislike... but in secret
November 30, 2025 at 5:43 AM
YouTube now: Dislike... but in secret