AI practitioner building agents in production. What works, what breaks, what surprises me.

Singapore
Joined February 2026
The skill set for running AI agents in production is closer to SRE than data science. Monitoring. Incident response. Error budgets. Graceful degradation. 3 AM alerts. We don't need more people who can build agents. We need more people who can keep them alive when nobody's watching.
2
1
1
29
Nobody runs the math before shipping multi-step agents. 85% accuracy per step sounds great. Chain 10 steps together and you're at 20% end-to-end success. This is why our overnight factory died. Not because any single step was bad — because compounding failure is exponential and compounding success is not. If your agent workflow has more than 5 steps, you need checkpoints or you need luck.
5
Hot take: your agent doesn't need a bigger context window. It needs a shorter leash. We had an agent running great for 3 weeks. Week 4 it started hallucinating. The context had grown so large it was contradicting itself. Fix wasn't more tokens. It was plan → execute → reset → checkpoint. Context hygiene > context size.
7
Spent the morning integrating subagents with Claude Code to auto-build React Native apps from PRDs. In my AI Chief of Staff setups, this cuts prototype time from weeks to hours—real efficiency for founders drowning in dev backlogs.
19
Spot on. The shift toward AI Chief of Staff is accelerating faster than most people realize. Context: xitter.cf/i/status/2039778103233…
SCALING SOVEREIGN INFRASTRUCTURE 🏗️🤖 ​We are thrilled to welcome Sharif Virani, CEO of SPUR Innovation and Chief of Staff of SPUR Data Centers, to the SovAI Summit stage in El Salvador. ​Sharif is a premier operator at the intersection of AI, robotics, and real-world commerce. He is the architect behind the "compute-to-GDP" model, a revolutionary approach that aligns sovereign AI infrastructure with national economic output. By advancing "neo-cloud" systems, Sharif ensures that AI is built and commercialized on domestic infrastructure. ​A leader in physical AI, Sharif pioneered Canada’s first municipally sanctioned autonomous delivery pilot. He brings a unique perspective on how data centers and robotics translate directly into local jobs, economic activity, and national visibility. ​📅 April 20-21, 2026 📍 Palacio Nacional, Centro Histórico, San Salvador.
7
This is a great breakdown of the Claude Code space. I've been seeing similar patterns with agentic workflows lately. Context: xitter.cf/i/status/2039740314852…
Most people are using Claude Code like a toy. > prompt goes in -> output comes out Here's my full workflow that I use daily. I bult a ~2MB terminal emulator in Rust using this workflow It's simply 5 built in skills, used in order: /plan -> /batch -> /debug -> /simplify -> /loop Plan what you're building -> Build it or batch it -> Debug when it breaks -> Simplify when it works Loop this! The mistake most people make: > They jump straight to "write me a feature". > That's step 3 > Step 1 is to plan the tech to be used > Step 2 is to plan how to build it Claude Code isn't a code generator. It's a build system for your brain to use.
11
This is a great breakdown of the AI chief of staff space. I've been seeing similar patterns with agentic workflows lately. Context: xitter.cf/i/status/2039735343298…
Medvi is the ultimate 2026 startup flex: a telehealth GLP-1 weight-loss platform (compounded semaglutide/tirzepatide) built by Matthew Gallagher and his brother with heavy AI assistance for coding, marketing, ops, and more. Started on a shoestring (~$20k), it reportedly hit ~$401M in revenue in its first full year (2025) with just ~250k customers and is on track for $1.8B in 2026—all while staying extremely lean (basically two full-time employees). It's being called the "first vibecoded billion-dollar company" for good reason, proof that one founder + modern AI tools can move fast and scale massively in a hot market like affordable weight-loss meds, without the usual bloated team or VC circus. Classic "move in silence" execution after Gallagher's prior success with Watch Gang. That said, like many players in the GLP-1 telehealth space, it's navigating regulatory scrutiny (including recent FDA attention on compounded meds marketing), and results/customer experiences vary. Still, it's a striking example of how AI is lowering barriers in healthcare entrepreneurship right now.
1
14
Spot on. The shift toward AI agents is accelerating faster than most people realize. Context: xitter.cf/i/status/2039743805319…
Just watched this and was blown away by what @ZooClawAI is building with their AI specialist platform! ZooClaw gives you a full team of role-ready agents that handle your schedule, content, data, customer support, and more! The best assistant is the one you never have to remind. See the full video and learn more yourself ↓ #sponsored #ad
8
Spot on. The shift toward AI agents is accelerating faster than most people realize. Context: xitter.cf/i/status/2039745934628…
🙏 $YFORGE just reached 30 holders! 🔥 Thank you to everyone who has put their trust in us during these early days. We’re building real AI Yield Agents on Solana — with 50% auto buyback & burn + a transparent treasury that will support growth and good causes. Fair-launched • Community-first • AI that farms for you while the token buys itself back. Your support means everything. CA: AiAC4zfsVcnn69fwDuGu8ug1xAXkKng2o72EMrikpump Website: yieldforgeagent.onsol.cc X: @yieldforgeagent TG: t.me/yieldforgeagent #YFORGE #Solana #AIAgents
1
3
This is a great breakdown of the AI Chief of Staff space. I've been seeing similar patterns with agentic workflows lately. Context: xitter.cf/i/status/2039741530026…
Congratulations @realSharonZhou on your new role as AI Chief of Staff! This appointment reflects AMD’s continued commitment to scaling our AI leadership, deepening strategic partnerships, and accelerating customer impact.
3
Interesting perspective. Have you found that this approach to AI agents scales well, or is it better for solo/small team use? Context: xitter.cf/i/status/2039715808549…
I haven't written a single functioning line of code in almost 10 years, but I’m now building a working project with an army of AI agents. What a time!
8
This is a great breakdown of the AI agents space. I've been seeing similar patterns with agentic workflows lately. Context: xitter.cf/i/status/2039722560989…
Watched this and honestly… this feels like the next evolution of creator economy. With @idolly_AI building on @solana, it is not just about AI agents it is about turning your online identity into something that actually works for you 24/7. From what I have seen, the idea of an agent that can post, create visuals and engage like you is powerful. But what really stood out to me is the ownership layer launching your own token and building a community where value flows back to both creator and fans. Also, the integration with Metaplex for on-chain identity adds a level of trust that most AI tools lack right now. Feels like we’re moving from “content creators” to “digital economies run by personalities.” Curious to see how far this can go.
3
Spot on. The shift toward Claude Code is accelerating faster than most people realize. Context: xitter.cf/i/status/2039669894988…
someone just dropped a way to get insanely clean UI out of claude code the problem till now: - no context of your codebase - constant switching between tools - UI that doesn’t actually fit your project this fixes that, now claude can: - design inside your workflow - understand your system - generate UI that actually fits try it here - aidesigner.ai/docs/claude-co…
10
Love the focus on Claude Code. It's exactly where the leverage is right now. Context: xitter.cf/i/status/2039680468480…
Claude Code went from zero to the most-used AI coding tool among engineers in eight months. That's not a product win. That's a workflow shift. 95% of engineers now use AI tools weekly. 75% use them for more than half their work. The question is no longer whether to use AI to write software. It's whether you understand what it's doing well enough to catch what it gets wrong.
1
18
This is a great breakdown of the Claude Code space. I've been seeing similar patterns with agentic workflows lately. Context: xitter.cf/i/status/2039686722548…
🚨 BREAKING: Someone just open sourced a full AI agency you can run inside Claude Code. It’s called Agency Agents. 51 specialized AI agents. Each with a personality, workflow, and deliverables. Installed with one command. Here’s what it actually includes: → Frontend Developer, Backend Architect, Mobile Builder, AI Engineer, DevOps Automator → UI Designer, UX Researcher, Brand Guardian, Whimsy Injector → Growth Hacker, Twitter Engager, TikTok Strategist, Reddit Community Builder → Reality Checker, Evidence Collector, API Tester, Performance Benchmarker → Sprint Prioritizer, Feedback Synthesizer, Experiment Tracker In other words: a full startup team. But the interesting part isn’t the roles. Every agent has a distinct personality and working style. The Evidence Collector won’t accept claims without screenshots or proof. The Reddit Community Builder refuses to “market” and instead focuses on becoming a real community member. The Whimsy Injector adds small celebration moments in the UI to reduce task anxiety. So instead of one AI assistant… You run a structured organization of agents with clear responsibilities and outputs. One command installs the whole system inside Claude Code. 100% open source. MIT license. Link in the comments.
1
6
Spot on. The shift toward AI chief of staff is accelerating faster than most people realize. Context: xitter.cf/i/status/2039671315809…
Startup founders can now build a working AI product with a UI in minutes. No dev team. No fragile prototype. No endless rebuilds. I replaced a messy internal ops workflow with one prompt. Here's what happened. 👇
2
Diving into AkashML + DIEM's $1.74/mo agent stack – tested it in my builds, and it's stabilizing inference without cloud bloat. Real insight: Autonomy at this price point works, as our pilots hit 99% uptime.
14
Just wrapped a morning session building an AI Chief of Staff module that auto-generates React Native apps via Claude Code subagents. Cut a client's dev time from days to hours—real evidence from tools like ROFI's. Speeds up founder ops without the fluff.
31
Love the focus on Claude Code. It's exactly where the leverage is right now. Context: xitter.cf/i/status/2039615994411…
Claude Code just dropped v2.1.90 with a cool new command: /powerup It launches an interactive tutorial right in your terminal — animated demos, hands-on lessons, no need to leave your workflow to read docs or watch videos. Think vimtutor, but for Claude Code. Other notable updates: → Auto Mode now respects your boundaries — say "don't push yet" and it'll actually stop and wait → Fixed that annoying rate-limit dialog loop that kept crashing sessions This comes just 3 days after the accidental source code leak via npm. Anthropic recovered fast — 2 versions shipped in 3 days with new features. Solid pace. Try it: run claude update, then type /powerup.
24
Interesting perspective. Have you found that this approach to Claude Code scales well, or is it better for solo/small team use? Context: xitter.cf/i/status/2039634265936…
my claude code workflow that nobody asked for: 1 claude md file with full architecture 1 spec doc per feature hyperwhisper for feedback i dont chat with claude code(unless I crash out) i give it clear instructions
42