scaling software services business of https://xitter.cf/t.co/LX9XAmAzi8, deeply into building SaaS & digital marketing
🟩🟩🟩🟩🟩🟩🟩🟩🟩⬜️ $5k/m
Joined January 2009
- Tweets6K
- Following497
- Followers5.5K
- Likes2.7K
Pinned Tweet
My timeline needs more people who build things. 👀
developers,
designers,
AI people,
indie hackers,
founders,
or anyone experimenting with cool ideas.
let’s connect, share what we’re working on, and learn from each other 🤝
Mahesh Chulet retweeted
The next great "language for agents" should be:
- As fast as Rust
- As little design space as Go
- Have the strictest type/verification system ever written
Bend seems like an interesting first effort
The next great "language for agents" should be:
- As fast as Rust
- As little design space as Go
- Have the strictest type/verification system ever written
Bend seems like an interesting first effort
Mahesh Chulet retweeted
Understanding agentic AI system design concepts in simple language:
• Agent: an AI that takes actions not just answers
• LLM: the brain that thinks and decides
• Prompt: the instructions you give the brain
• System prompt: the agent's job description
• Tools: the hands (search, code, APIs, browser)
• Function calling: how the brain uses its hands
• Memory: short-term is the chat, long-term is what it remembers forever
• Context window: how much it can hold in its head at once
• Planning: breaking a big goal into small steps
• Reasoning loop: think, act, check, repeat
• Reflection: the agent reviews its own work and fixes it
• RAG: looking things up before answering
• Embeddings: turning words into numbers so AI can find meaning
• Vector database: the library where those numbers live
• MCP: a universal plug that connects AI to any app
• Orchestrator: the manager that assigns work to other agents
• Multi-agent: a team of AIs, each with one job
• Sub-agents: helpers the main agent spins up for side tasks
• Guardrails: rules that stop it doing something dumb
• Human in the loop: you approve the risky stuff
• Evals: tests that check if the agent actually works
• Observability: logs that show what it did and why
• Latency: how long it takes to respond
• Token cost: what every thought costs you
Chatbots talk.
Agents do.
Bookmark this before your next AI interview.
Understanding agentic AI system design concepts in simple language:
• Agent: an AI that takes actions not just answers
• LLM: the brain that thinks and decides
• Prompt: the instructions you give the brain
• System prompt: the agent's job description
• Tools: the hands (search, code, APIs, browser)
• Function calling: how the brain uses its hands
• Memory: short-term is the chat, long-term is what it remembers forever
• Context window: how much it can hold in its head at once
• Planning: breaking a big goal into small steps
• Reasoning loop: think, act, check, repeat
• Reflection: the agent reviews its own work and fixes it
• RAG: looking things up before answering
• Embeddings: turning words into numbers so AI can find meaning
• Vector database: the library where those numbers live
• MCP: a universal plug that connects AI to any app
• Orchestrator: the manager that assigns work to other agents
• Multi-agent: a team of AIs, each with one job
• Sub-agents: helpers the main agent spins up for side tasks
• Guardrails: rules that stop it doing something dumb
• Human in the loop: you approve the risky stuff
• Evals: tests that check if the agent actually works
• Observability: logs that show what it did and why
• Latency: how long it takes to respond
• Token cost: what every thought costs you
Chatbots talk.
Agents do.
Bookmark this before your next AI interview.
TOP 1% AI PROJECTS that get you hired in 2026.
1.) Terminal Agent
Autonomous shell execution with self-debug loops.
Target: Top 10 on Terminal-Bench.
Shows: Agents that operate real systems safely.
2.) Computer-Use Agent
Browser agent that books, fills, extracts on live sites.
Target: 90% task success, human takeover under 10%.
Shows: You ship agents in the messy real world.
3.) MCP Server Suite
3 production MCP servers with auth, rate limits, docs.
Target: 500+ installs or adopted by one real team.
Shows: You build plumbing others build on.
4.) Agent Red-Team Scanner
Fuzzes your own agents for injection and tool abuse.
Target: Find and patch 10+ vulns, publish the report.
Shows: You think like an attacker, ship like a defender.
5.) Stateful Long-Running Agent
Temporal-checkpointed workflows that survive crashes.
Target: 24h task resumes cleanly after kill -9.
Shows: You build durable systems, not fragile demos.
6.) Cost-Aware Model Router
Routes by complexity with caching and fallbacks.
Target: -50% spend, under 2% quality drop.
Shows: You optimize unit economics, not vibes.
7.) On-Device Private Agent
Quantized local model in browser or edge, offline mode.
Target: $0 inference at 1k users.
Shows: You ship private AI without API bills.
8.) Trajectory Eval Harness
Grades every hop, not just the final answer. Blocks bad PRs.
Target: 100+ golden trajectories, CI gate live.
Shows: You measure what actually breaks.
9.) Self-Improving Data Flywheel
Thumbs-down → synthetic data → LoRA fine-tune loop.
Target: +5% quality per cycle, measured on evals.
Shows: You build moats, not features.
10.) Multi-Agent Research Desk
Researcher + writer + fact-checker with consensus logic.
Target: 1 published report, every claim cited and verified.
Shows: You orchestrate teams of agents, not toys.
Most people stay stuck watching tutorials.
Builders get hired.
(Bookmark & Repost)
Mahesh Chulet retweeted
Using Claude Code in the terminal is mostly an ego thing.
The desktop app looks better and does the job.
Prove me wrong. What does the CLI do that the app can't?
Using Claude Code in the terminal is mostly an ego thing.
The desktop app looks better and does the job.
Prove me wrong. What does the CLI do that the app can't?
Mahesh Chulet retweeted
The 15 stages of every vibe coder's backend:
1. "Just use localStorage, it's fine"
2. "Okay, let's add a real database"
3. Picks Supabase because of a tweet
4. Forgets to turn on row level security
5. API keys pushed to a public GitHub repo
6. "Why is my OpenAI bill $847?"
7. Asks the AI to "make it secure"
8. It rewrites the entire backend
9. Nothing works anymore
10. Reverts to 3 days ago
11. Adds auth. Breaks login for everyone
12. "Who needs migrations?"
13. Production database deleted by an agent
14. Learns what a backup is
15. Hires a backend engineer
Which stage are you on?
The 15 stages of every vibe coder's backend:
1. "Just use localStorage, it's fine"
2. "Okay, let's add a real database"
3. Picks Supabase because of a tweet
4. Forgets to turn on row level security
5. API keys pushed to a public GitHub repo
6. "Why is my OpenAI bill $847?"
7. Asks the AI to "make it secure"
8. It rewrites the entire backend
9. Nothing works anymore
10. Reverts to 3 days ago
11. Adds auth. Breaks login for everyone
12. "Who needs migrations?"
13. Production database deleted by an agent
14. Learns what a backup is
15. Hires a backend engineer
Which stage are you on?
Mahesh Chulet retweeted
Guys, I don’t understand.
Chinese models are clearly very strong, especially at
coding like
for example, DeepSeek is very good at coding
So why does everyone spend $200 on Claude or in on ChatGPT subscriptions? 🤔
Guys, I don’t understand.
Chinese models are clearly very strong, especially at
coding like
for example, DeepSeek is very good at coding
So why does everyone spend $200 on Claude or in on ChatGPT subscriptions? 🤔
Mahesh Chulet retweeted
The MOST obvious (but REALLY good) STARTUP ideas right now:
1. Zestimate for company data (what your data is worth to an AI lab)
2. Google Analytics for agents (see which agents visit your site and what they want)
3. Y Combinator for 1 person companies (back solo founders running a team of agents)
4. Okta for agents (who each agent is and what it's allowed to touch)
5. Marie Kondo for your company's data (clean it up so agents can use it)
6. FICO for agents (a trust score before an agent can spend money)
7. Geico for agents (insurance for when an agent makes an expensive mistake)
8. Datadog for agents (see every step your agents took)
9. Twilio for agents (phone numbers and texting for agents)
10. Upwork for agents (hire an agent for a job instead of a freelancer)
11. ESPN for AI (every model launch covered like a big game).
12. Shutterstock for robot training videos (get paid to film yourself doing chores)
13. Credit Karma for what AI knows about you (see it, fix it, delete it)
14. Norton for AI scams (protect your parents from fake voices and emails)
15. PayPal Buyer Protection for agent purchases (fixes it when your agent buys the wrong thing)
16. Geek Squad for AI (set up personal agents for families and small businesses)
17. Dell for private AI (AI in a box for firms that won't use the cloud)
18. Carfax for used GPUs (a health report before you buy secondhand)
19. Airbnb for idle GPUs (rent out computing power you're not using)
20. Rotten Tomatoes for AI models (real reviews instead of benchmarks)
21. G2 for skills (find the best skills for your agent)
22. Dropbox for your AI memory (take what your agent knows about you to any agent)
23. Canva for physical products (design it with AI and get it made)
24. ADT for your personal agent (alerts when your agent does something it shouldn't)
25. Hertz for robots (rent one for a move, a deep clean or an event)
26. Harvard Business School for internet businesses (a real education in building online)
27. Workday for agents (onboard, manage and offboard your digital workers)
28. Getty Images for your face and voice (license your likeness to AI safely)
29. Meta Muse for older adults (and other high LTV niche)
30. Google Ads for agents (pay to be what the agent recommends)
31. Loom for agents (a walkthrough of everything your agent did)
32. Plaid for agents (approved access to banks, airlines and stores)
33. Jiffy Lube for vibe coded apps (quick fixes when your AI built app breaks)
34. Ancestry for datasets (where your training data came from)
35. The organic label for human-made content (proof a person made it)
36. Strava for builders (track what you ship every week)
37. Honey that haggles for you (your agent negotiates the price)
38. SEM Rush for personal agent platforms (like Muse etc)
39. Frequent flyer miles for agents (loyalty programs built for agents)
The MOST obvious (but REALLY good) STARTUP ideas right now:
1. Zestimate for company data (what your data is worth to an AI lab)
2. Google Analytics for agents (see which agents visit your site and what they want)
3. Y Combinator for 1 person companies (back solo founders running a team of agents)
4. Okta for agents (who each agent is and what it's allowed to touch)
5. Marie Kondo for your company's data (clean it up so agents can use it)
6. FICO for agents (a trust score before an agent can spend money)
7. Geico for agents (insurance for when an agent makes an expensive mistake)
8. Datadog for agents (see every step your agents took)
9. Twilio for agents (phone numbers and texting for agents)
10. Upwork for agents (hire an agent for a job instead of a freelancer)
11. ESPN for AI (every model launch covered like a big game).
12. Shutterstock for robot training videos (get paid to film yourself doing chores)
13. Credit Karma for what AI knows about you (see it, fix it, delete it)
14. Norton for AI scams (protect your parents from fake voices and emails)
15. PayPal Buyer Protection for agent purchases (fixes it when your agent buys the wrong thing)
16. Geek Squad for AI (set up personal agents for families and small businesses)
17. Dell for private AI (AI in a box for firms that won't use the cloud)
18. Carfax for used GPUs (a health report before you buy secondhand)
19. Airbnb for idle GPUs (rent out computing power you're not using)
20. Rotten Tomatoes for AI models (real reviews instead of benchmarks)
21. G2 for skills (find the best skills for your agent)
22. Dropbox for your AI memory (take what your agent knows about you to any agent)
23. Canva for physical products (design it with AI and get it made)
24. ADT for your personal agent (alerts when your agent does something it shouldn't)
25. Hertz for robots (rent one for a move, a deep clean or an event)
26. Harvard Business School for internet businesses (a real education in building online)
27. Workday for agents (onboard, manage and offboard your digital workers)
28. Getty Images for your face and voice (license your likeness to AI safely)
29. Meta Muse for older adults (and other high LTV niche)
30. Google Ads for agents (pay to be what the agent recommends)
31. Loom for agents (a walkthrough of everything your agent did)
32. Plaid for agents (approved access to banks, airlines and stores)
33. Jiffy Lube for vibe coded apps (quick fixes when your AI built app breaks)
34. Ancestry for datasets (where your training data came from)
35. The organic label for human-made content (proof a person made it)
36. Strava for builders (track what you ship every week)
37. Honey that haggles for you (your agent negotiates the price)
38. SEM Rush for personal agent platforms (like Muse etc)
39. Frequent flyer miles for agents (loyalty programs built for agents)
Mahesh Chulet retweeted
My timeline needs more people who build things. 👀
developers,
designers,
AI people,
indie hackers,
founders,
or anyone experimenting with cool ideas.
let’s connect, share what we’re working on, and learn from each other 🤝
Mahesh Chulet retweeted
Hey @X,
Looking to connect with people interested in:
🎮 Mobile game development
🤖 AI agents
⚡ Vibe coding
🛠️ Building in public
🎨 Pixel art and game design
📱 Indie apps and side projects
🎬 Dev YouTube
Building my first mobile game right now. What are you building? Say hi 👋
Hey @X,
Looking to connect with people interested in:
🎮 Mobile game development
🤖 AI agents
⚡ Vibe coding
🛠️ Building in public
🎨 Pixel art and game design
📱 Indie apps and side projects
🎬 Dev YouTube
Building my first mobile game right now. What are you building? Say hi 👋
Mahesh Chulet retweeted
Create these 6 files before you start vibe coding:
• PRD.md: what you're building and for whom
• ARCHITECTURE.md: stack, folders, data flow
• DESIGN.md: colors, fonts, components, vibe
• RULES.md: coding style and what the AI must never do
• TASKS.md: small steps, one at a time
• PROGRESS.md: what's done, what broke, what's next
No files = the AI guesses.
6 files = the AI follows your plan.
10 minutes of writing saves 10 hours of fixing.
Bookmark this before your next build.
Create these 6 files before you start vibe coding:
• PRD.md: what you're building and for whom
• ARCHITECTURE.md: stack, folders, data flow
• DESIGN.md: colors, fonts, components, vibe
• RULES.md: coding style and what the AI must never do
• TASKS.md: small steps, one at a time
• PROGRESS.md: what's done, what broke, what's next
No files = the AI guesses.
6 files = the AI follows your plan.
10 minutes of writing saves 10 hours of fixing.
Bookmark this before your next build.