Recent content on Cameron Westland
Each user in our FIRE alpha talks to an investing agent we run for them. Every day at noon, a second agent I call the watcher checks on one of those users. It reads four systems: Supabase for what his agent sessions did, PostHog for what he did in the app, Knock for whether our morning briefing email reached him, Braintrust for how the agent behaved. It posts four lines into an ops thread I read:…
On July 16th I recorded a ten-second screen capture of a scroll bug that had escaped to production. Our chat transcript renders interactive cards: a portfolio card you can drill into, tap a ticker, get the detail view. Tapping a ticker made the whole transcript lurch. I handed the video to Codex running GPT-5.6 Sol at extra-high reasoning, a model that had been the state of the art for exactly one…
On June 7, Peter Steinberger tweeted : “Here’s your monthly reminder that you shouldn’t be prompting coding agents anymore. You should be designing loops that prompt your agents.” Eight million views. Around the same time, Boris Cherny, the creator of Claude Code, said on Acquired Unplugged : “I don’t prompt Claude anymore. I have loops running. They’re…
My 30-day token bill is $11,232.54. Fourteen billion tokens, almost all of it GPT-5.5 on a ChatGPT Pro 20x plan. Yesterday alone I put $1,157.78 and 1.4 billion tokens through the harness. Today, mid-afternoon, I’m already at $226 and 270M. 
 That’s nothing next to Peter Steinberger : $1.3M and 603 billion tokens against the OpenAI API in 30 days. Peter noted in replies that…
I had coffee today with a founder I see every 90 days. He’s pre-PMF, mostly solo, a couple friends helping. He always closes the same way, grateful for the time and the perspective. On the way out I gave him the next move: “If you want to take this to the next level, start emailing me weekly. Treat me like an investor.” 
 The email is a forcing function disguised as a…
Ask Apple Intelligence something it can’t handle and watch what happens. It delegates to ChatGPT. You get a brief “Working with ChatGPT…” indicator, then results come back. Claude Code delegates to subagents . Linear lets you @mention agents that run entire sessions with their own lifecycle. Stagehand wraps an internal LLM loop behind MCP tools for browser automation.…
Mario Zechner wrote a post today called “Thoughts on Slowing the Fuck Down” . His argument: AI agents compound errors in codebases because humans are no longer the bottleneck. “A human cannot shit out 20,000 lines of code in a few hours.” That friction was a feature. Without it, tiny mistakes accumulate until the codebase is untrustworthy. 
 He’s right about code.…
I pointed an AI agent at a performance-sensitive Python code path, gave it a 40-line spec and a replay harness, and walked away. An hour later it had tried 49 optimizations, kept 20, and taken the p95 latency from 339ms to 34ms. The whole thing cost $24. 
 Here’s the artifact it produced. A JSONL file where each line is a structured experiment: 
 { 'run' : 1 , 'metric' : 338.57 ,…
There’s a talent war happening in AI right now, and it’s changed the tone of a lot of career conversations. 
 I’ve been thinking more explicitly about the framework I use for deciding where to work and when to stay put. I realized later on a walk home from the gym that I’d never actually written it down. This is that post. 
 Three dimensions: 
 
 Meaningful…
I’m recording a Loom walkthrough of a new feature when I notice something broken—unrelated to what I’m demoing. A year ago, I would have logged a Linear issue. It would sit in triage for a week, get prioritized in a meeting, eventually get picked up. Now I open Claude Code, screenshot the bug, and type: “Why is this broken? Make me a plan to fix.” 
 I have a backlog of…
I watched our CEO and a senior technical leader debate “everyone ships code” in a Slack thread this week. The technical leader is exactly the kind of person you’d expect to embrace this—ex-founder, high agency, ships fast. He pushed back. 
 Our CEO cited Tobi Lutke’s tweet about shipping more code in three weeks than the previous decade. The response: he’s seen…
A tweet from Amir Salihefendic hit 288K views last week: 
 
 “Cursor feels like a company without product management. Just devs and designers running seemingly random experiments with little higher-level strategy. I use Cursor a lot, and it is becoming a very frustrating experience.”
— @amix3k , Dec 18, 2025 
 
 He wasn’t alone. Two days earlier, kyzo’s…
My post about DHH moving Rails off GitHub got flagged on Hacker News within an hour: 
 
 They’re not wrong. I did use AI—for transcription, for structural editing, for finding holes in my argument. But every claim was mine. Every link was verified by me. I spent hours on it. 
 The problem wasn’t that I used AI. The problem was that engagement-optimized writing now looks…
GitHub proposed charging for the Actions control plane , then backed off after community backlash . At the same time, services like Blacksmith can run CI machines faster and cheaper while GitHub still brokers the workflow. GitHub charges for the parts it still controls. 
 DHH should move Rails off GitHub. The tooling isn’t ready—but moving Rails would make it ready. 
 How I got here:…
Why some AI feels like translation—and some feels like replacement 
 TL;DR — AI gets adopted fastest when it translates human judgment into a different format (notes → worksheet, jargon → plain English, transcript → summary). It gets scary when it replaces the thing you were paid to produce. 
 I’m at Nathan Phillips Square watching my kid’s class skate. A handful of parents, a couple of…
After thousands of AI-assisted commits, I discovered something surprising: the best AI framework is the one that doesn’t feel like a framework at all. 
 I write markdown files that tell AI what to do. I use bash commands to make things happen. That’s it. 
 gh issue create --title 'Add user auth' --body '...' 
 gh pr create --title 'feat: implement OAuth' 
 psql…
I killed our weekly triage meeting last month. Three hours compressed to five minutes. But here’s the thing—it took me six failed attempts to get there. 
 The breakthrough wasn’t making the AI smarter. It was making the task more structured. This is what context engineering actually looks like—messy, iterative, and focused on constraints rather than capabilities. 
 Let me show…
I watched our AI completely melt down during a CTO demo last month. 
 The model started confidently mixing up “deep research” (an existing feature) with “thematic analysis scans” (the new feature I’d just shipped). Every response made the confusion worse. The CTO’s face went from interested to puzzled to concerned. I sat there watching my careful prompt…
I learned this the hard way: when someone else pays the cost of understanding your AI-generated content, it isn’t real productivity. 
 I generated a Dagster pipeline plan in 6 minutes. My colleague Sam spent nearly an hour trying to review it. Here’s what the AI gave me: 
 
 Pipeline Overview : Implement staged rollout with A/B testing framework, targeting 5% initial…
Here’s the thing: I can’t sit still for three hours. 
 A friend sent me this Andrew Huberman podcast last weekend. “You’ll love this,” they said. And they were probably right. But three hours? I tried. Really tried. Made it about 20 minutes before I started cleaning my desk. Then checking email. Then I lost track completely. 
 Sound familiar? 
 The Real…
TL;DR 
 
 Make PR submitters show proof of work. 
 Build on old rituals like code walk-throughs. 
 Experiment with small PRs, code screencasts, and AI video checks to realign incentives. 
 These could prevent teams from getting buried in review debt. 
 
 A Thought Experiment 
 In “ The Externality of AI Velocity ,” I wondered if the real issue with…
Cursor’s Pricing Puzzle 
 Cursor’s 4-July “ Clarifying Our Pricing ” post explains that the new Pro plan’s US $20 credit buys roughly 550 Gemini 2.5, 650 GPT-4.1, or 225 Claude Sonnet calls before metering kicks in. My reaction: that covers half a power-user day, not a month. If Cursor needed a second post to spell this out, plenty of devs are lost in the math. 
 My $200 Experiment Card…
TL;DR *: Claude Code hooks let you inject shell commands at key points in your development workflow. I built two simple hooks that enforce branch protection and run code quality checks automatically. It’s like having a pair programming partner who never gets tired of reminding you about the basics.* 
 I’ve pretty much switched to Claude Code exclusively for development work. The…
Or: How the AI wars just created the biggest startup opportunity in history 
 TL;DR: AI power is compounding faster than usability. The resulting complexity gap is the best greenfield for vertical software since the spreadsheet. 
 
 Excel’s trick, in fast-forward 
 
 1985 Excel 1.0. Rows, columns, arithmetic. 
 1990 Charts and a macro recorder arrive. 
 1993 PivotTables…
TL;DR - Early-stage startups run on conviction. A CEO who radiates certainty turns rough ideas into believers, believers into runway, and runway into time for the numbers to catch up. Guard that spark. 
 CEO Conviction: the real unit of survival 
 Smart calls and lucky breaks help. Belief pays the bills. Your job is to grow the pool of believers (employees, investors, lighthouse customers)…
TL;DR — LLMs can turn a spoken outline into a 500-line plan in minutes, but the time you save up front can re-appear downstream as reviewer pain. My six-minute, AI-generated Dagster plan cost a teammate an hour of forensic code review—proof that velocity has externalities. Collapse the slop, flag uncertainties, and optimise for their minutes before you open the PR. 
 When vibe-coding dumps…
How MCP‑style LLM Wrappers Unlock the Integrations We Always Wanted 
 TL;DR — When LLM-powered agents can plan and call REST APIs for you, the activation energy for integrations collapses. MCP sticks to plain JSON + standard auth while the model handles schemas, retries, and pagination—ideas that once took weeks of Zapier flows now fit in a single prompt. 
 My brother—a twenty-year IT…
The tech world is buzzing with dramatic predictions about AI and coding: 
 
 “In 12 months, we may be in a world where AI is writing essentially all of the code.” - Dario Amodei, Anthropic CEO 
 “I think software engineering by the end of 2025 looks very different than software engineering at the beginning of 2025.” - Sam Altman, OpenAI CEO 
 “We at…
Someone recently asked me about a program called “ BetaKit’s Most Ambitious ” for Toronto Tech Week. They wanted my opinion on who’s the Canadian doing the most ambitious tech-related work outside of Canada, and who we should “repatriate.” 
 While I chose not to participate directly in BetaKit’s program due to previous experiences with their coverage,…
In my last post , I shared how my local reinforcement learning experiments slammed into a hardware wall. Despite my M4 Max’s impressive specs, Apple’s Metal framework proved inadequate for the tensor operations needed in RL training. After weighing my options, I decided to migrate to a cloud GPU solution—a decision that opened up a whole new adventure in debugging, optimization, and…
Last time I shared how I was dipping my toes into reinforcement learning with my “side side quest” project. I had just set up the environment and was waiting on benchmark results to decide whether RL was worth pursuing for my hypothesis validation component. Well, a lot has happened since then - some promising developments and some humbling reality checks. Let me take you through the…
Using LLMs to Break Down Barriers in Technical Learning 
 I’ve always been fascinated by reinforcement learning since the AlphaGo days. As an avid gamer, the concept of learning through repetitive play strongly resonates with me. But despite my programming experience, I always viewed reinforcement learning as “high science” – the hardest aspect of machine learning, beyond my…
I had an early test user trying out my app , and he sent me three Loom videos. One of them showed a genuine bug—he was working through a list of tasks, hitting reject on one, accept on another, and somehow the whole thing kept resetting. Very frustrating. The other two videos? Just him expecting a capability we didn’t have. 
 The first bug was obvious. It wasn’t the expected behavior. But…
I know it’s a lot less cool to talk about these days, but I was one of the first people to get a Tesla Model X. My VIN showed it was the 2550th off the line, and I even toured the factory. I was excited. But the thing that really stood out to me wasn’t the battery or the acceleration—it was self-driving mode. 
 Cars had lane-keeping for a while, but I always struggled with driving. It wasn’t…
Benedict Evans recently wrote two sharp critiques of AI research tools: “ The Deep Research Problem ” and “ Are Better Models Better? ” His argument is compelling and straightforward: AI tools like OpenAI’s Deep Research confidently present facts that are sometimes wrong, creating an “unknown unknowns” problem. If you’re not already familiar with a…
Sam Altman posted about the future of AI recently—his “ Three Observations "—and I found myself thinking about it this morning while lying in bed. His core message was that by 2035, anyone should be able to harness the skills of anyone from 2025. Think about that: someone who can barely read having access to the capabilities of a Harvard lawyer. That’s a massive shift. 
 I’ve…
As a product engineer , I’ve discovered how generative AI is transforming my code refactoring process. I’m not claiming AI is a cure-all, but I want to share a practical example that many developers facing messy component states will understand. 
 The Setup: A Tale of State Management Gone Wild 
 I was working on a multi-step form component “PinContent” when I…
I’ve been thinking about how we use Large Language Models (LLMs). While many focus on AI-generated content, I believe we’re missing a more powerful use: AI as an editor. 
 The Writing Revolution We’re Living Through 
 We write more today than any generation before us. Our parents wrote occasional letters and grocery lists. We write constantly - through Slack, email,…
Let’s talk about a transformation in my professional life. It’s a story of rediscovery and finding my true calling in a rapidly evolving tech landscape. Join me as we dive into the world of product engineering, AI, and why I’m making a dramatic shift in my career. 
 The Management Merry-Go-Round 
 The call to management is tempting. “We wish we could clone…
Let’s talk about Figma and AI startups. Don’t get me wrong, Figma is a powerful tool. It’s great for inspiring teams and aligning visions. In many ways, it’s like concept art in video game development - beautiful, aspirational, and great for setting the tone. 
 But here’s the thing: concept art never tells you if a game is fun to play. And Figma prototypes…
When I first started fforward, my goal was to explore the intersection of decision-making in startups and the scalability potential of new technologies like LLMs. I experienced significant pain in my previous startup around building alignment and adapting rapidly within a large team. My initial concept was to create a decision memo or decision log system. The idea was to use natural language…
Traveling to conferences is a rollercoaster ride of emotions for me. 
 One moment, I’m savoring the solitude of an airport lounge, and the next, I’m engulfed by the ache of missing my family. As the co-founder of an early-stage startup, fforward , attending events like TechCrunch Disrupt offers a whirlwind of opportunities and challenges. 
 In this blog post, I’ll take…
The 50-50 Ideal: Fair, But Is It Enough? 
 For much of my life, I’ve held onto the belief that a fair and equitable partnership—whether in marriage or parenting—should ideally be a 50-50 split in duties and responsibilities. You take care of half, I take care of the other half, and we make a whole together. On paper, this sounds perfect. 
 Yet, I’ve begun to question its…
The Backdrop 
 
 Four years after saying goodbye to Northern California, my family and I returned to Marin County for a two-week vacation. There’s an age-old saying, “You can never go home again,” but what does it mean when that place once was home? The experience was emotional. It made me ponder the complexity of belonging, the notion of home, and the inevitable…
This blog post is based on a thought experiment and explores the future potential of generative AI in redefining Spend Control within companies. 
 Introduction 
 People have previously misunderstood Spend Control as a strict enforcer of budgets. In reality, it serves as an organizational keeper of principles, challenging spending that exceeds certain thresholds. But the landscape could be…
Over the past five years, I’ve had regular talk therapy. It’s been crucial to my self-growth and mental stability. But I’ve found a surprising aid: good, old conversations with friends. 
 I’m cautious about saying this because it’s my experience. Not everyone may feel the same. In my view, men often shy away from deep talks with friends. For me, building this into my life was a conscious act.…
Each dawn at 7 a.m., I have a routine. 
 I walk my dogs. It’s a short walk. We tread to a roundabout, where a small park lies. We call it the Circle Park. The dogs attend to their needs, and then we head home. In the school year, my kids join. In summer, I’m usually alone. 
 This walk, simple as it is, clears my head. 
 Have you ever had pickled ginger? It cleanses your…
As I write this, my thoughts are tumultuous, and my emotions are raw. 
 The start of this year marked a significant turn in my life. I said farewell to my long-standing company and the security it provided. I found myself in the vast sea of uncertainty for three months, asking myself - ‘What’s next?’ 
 A new opportunity soon emerged from the chaos—a light at the end of…
Sometimes life is hard, a constant battle that keeps you on edge. It throws punches, and sometimes you’re unprepared to dodge. But luckily, we all have our unique coping mechanisms. 
 Mine? Video games. 
 Imagine feeling down and unable to think straight, only to find solace in digital fantasy. That’s been my experience. 
 With a recent bout of strep throat causing havoc in my life, I…
The sun streams through the windows this Sunday morning. I find myself wrapped in the warmth of my bed, succumbing to the siren song of the “snooze” button a few times more than usual. 
 I had a late-night filled with bouts in The Witcher 3 and unraveling the mysteries of Silo’s first season. I was waiting for my wife’s return from a Saturday night Beyoncé concert,…