I build like an operator, not a hobbyist. The AI tools aren't novelties — they're an operating system for shipping. The edge was never access to the model. It's knowing what to ask, what to throw away, and how fast to move.
AdvisoryCategory strategy for technical infrastructure companies: naming the category, and the developer-to-enterprise story that turns engineer love into enterprise revenue.
operator console — build & ship the Oriya normalizer, in one night · illustrative
Manus
›
GPT
›
Claude Code
›
Replit
›
Evals
›
Vercel
raw readout →one fair 0–100 index
Ourareadiness 82—
Whooprecovery 44%—
Garminbody battery 55—
Appleno native score—
Same night, four devices, four scales that don't compare — 82 vs 44% vs 55 vs nothing at all. Give each one a handicap and they land within a few points of each other: now a squad on mixed hardware competes fairly — and it held across 50 synthetic edge cases before it ever shipped. That normalizer is Oriya.
01AI · building
Claude Codeagentic coding
My primary pair on real codebases. I scope tightly, let it hold the architecture, and review every diff like a staff engineer. The model isn't the moat — knowing what to ask is.
Replitprototype → deploy
Zero to running app in one tab. Where an idea gets tested before it earns its own repo.
Vercelship the frontend
Preview deploy per branch. “Show me” becomes a live URL in minutes, not a meeting.
Manusautonomous runs
I queue the grunt work — competitor teardowns, data pulls, first-draft ops — then audit the output. Delegation, not magic.
GPTreasoning · adversary
My scratchpad and sparring partner. I make it argue against my own thesis before I commit code or capital.
Geminilong context · multimodal
Drop in a whole deck, transcript, or spec and interrogate it as one object. Breadth where the others go deep.
02Growth · the loop
The other half of the stack — not how I build the product, but how I see what's working and move the number. Roughly in loop order: measure, experiment, reach, keep, expand.
PostHoganalytics · replay · flags
Where I start on day one. Product analytics, session replay, and feature flags in one place — I can watch what a user actually does instead of what a dashboard rounds them into. Open-source, generous free tier, ships fast.
Amplitudebehavioral depth
When the questions outgrow PostHog — cohorts, funnels, and retention curves the whole org will trust. What I reach for once activation is the number everyone's arguing about.
Statsigexperimentation
Ship, measure, decide — the loop that separates operators from opinion-havers. Feature gates and experiments in one system; it's what the serious AI teams run, and for the right reason.
ClayGTM data · outbound
The unfair advantage in outbound right now. Enrich, score, and personalize at the row level, then let it run — it turns a flat list into a growth motion without a ten-person SDR floor.
Customer.iolifecycle
Behavioral messaging that fires on what a user did, not what day it is. Where onboarding and retention stop being a newsletter and start being a system.
AttioCRM · pipeline
The CRM that finally reads like software built this decade. Data-model-first and fast enough to shape around your actual motion, instead of forcing you into Salesforce's.
Hightouchreverse-ETL
The quiet unlock: the warehouse becomes the source of truth and every downstream tool gets the same clean data. Activation without the spaghetti of point-to-point integrations.
03Productivity · SaaS hacks
Build the throwaway first. A 30-minute prototype kills more bad ideas than a three-hour debate.
Prompts are infrastructure. Version them, reuse them, treat the good ones like code — not clever one-offs.
Automate the second time, never the first. The first pass teaches you where the edges are.
One triage queue. Route every tool's pings into a single surface — context-switching is the tax nobody prices.
Default to the boring stack. Novelty is a cost; pay it only where it's the actual product.
04Migrations · the switch
Namecheap → Cloudflaredns · edge
Left the domains at Namecheap out of inertia until a record change took hours to propagate instead of seconds. Cloudflare runs the fastest anycast resolver on the internet, registers domains at wholesale cost with zero upsells, and hands you a WAF, unmetered DDoS mitigation, and edge caching on the free tier. The pitch that finally moved me wasn't a feature — it was that I stopped thinking about DNS entirely.
Netlify / raw AWS → Verceldeploy
Netlify was the gateway drug; raw AWS was the tax I paid to feel serious. Vercel won on one mechanic: a preview URL per branch turns “show me” from a meeting into a link. Framework-defined infrastructure means I stopped hand-writing the deploy, and because every build is immutable and promotion is just an alias swap, rollback is instant — no rebuild, no downtime.
Postgres read-replica → Snowflakeanalytics
Analytics rode a groaning Postgres read-replica until the dashboards started timing out and every heavy query threatened prod. Snowflake's unlock is boring and total: storage and compute are decoupled, so a brutal backfill spins up its own warehouse, never touches the app, and you pay for the seconds it runs and switch it off. Separation of concerns, billed by the second.
EC2 GPU babysitting → Modalai infra
A newer one I genuinely enjoy. For anything that needs a GPU for ninety seconds, Modal killed the old loop — spin up an instance, forget to kill it, find a $400 bill next week. You decorate a Python function, it runs on their fleet, you pay for the execution and nothing idles. Serverless the way the word was supposed to mean.
05Tools I back · referral
The stack behind the teams I back — the tools I actually reach for in the trenches with a company, and the AI ones I'd stake my name on. As partner links land, readers get a discount or an extended trial — I'll mark each one here. I keep the list honest by only naming what I use or would recommend without one.
The deploy layer I moved my own stack to. Preview URL per branch, instant rollback, zero infra babysitting — the reason “show me” becomes a link, not a meeting.
Serverless GPUs the way the word was meant. Decorate a Python function, pay for the ninety seconds it runs, nothing idles — the end of the surprise $400 bill.
My primary pair on real codebases. The model isn't the moat — knowing what to ask is — and this is the one I trust to hold the architecture while I review every diff.
AI that ships a real front end from a sentence. What I hand a non-technical founder who needs a working product in front of users this week, not next quarter.
An AI exec assistant that clears the inbox and drafts in your voice. The first one I've kept past the trial — it buys back the hour founders pretend they don't lose to email.
Fast, cheap inference for teams shipping on open models — where the inference bill stops outrunning the product, the exact math that decides whether an AI feature survives.
The boring, reliable place to run the backend so you don't hand-roll infra. Heroku's ease without the ceiling — I reach for it when the app just needs to run.
The voice layer for anything that talks — agents, media, localization. The one I'd bet on for going multilingual into markets your current product can't yet speak to.
Generative video a real creative team will actually use — where the cost of a moving image drops far enough to change what a small team can ship.
Where a link above becomes an affiliate or referral link, it costs you nothing extra — often less — and I'll say so. I only list tools I use, or would recommend without one.
06Takes · what I actually believe
The bottleneck moved from writing code to specifying intent. Taste and judgment are the scarce inputs now.
Most “AI products” are a system prompt in a trench coat. Defensibility lives in the data loop and the distribution — not the model call.
Speed is strategy, not a personality trait. Whoever learns fastest in-market wins; tooling just compresses the loop.
Owning the workflow beats owning the model. The model is rented; the workflow, the data, and the trust are yours.
Developer mindshare isn't revenue. The devtools that grow up run two tracks at once — an open-source wedge engineers love, quietly generating the compliance story a CISO signs.