Sabitlenmiş Tweet
Jean-Michel Lemieux
7K posts

Jean-Michel Lemieux
@jmwind
Full time apprentice human.
Ottawa ⮕ 🇨🇦 Katılım Mart 2009
188 Takip Edilen47.6K Takipçiler

V… you still in there? Blink twice if the AI has fully optimized you 😂
I get the take. I just don’t buy it.
Efficiency is a tool, to buy time for impact, meaning, beauty, and purpose. Nothing to do with relaxing. Efficiency clears the mundane to make way for time on the impactful.
If efficiency is your source of dopamine hits, you’ll end up a hamster on that wheel and not go anywhere.
English

The more I use ai the more I’m convinced we’re hardwired as humans to be addicted to efficiency. It just feels good. The fallacy is that we want to make tools so we can relax. The truth is the more efficient you become, the more you want to be EVEN more efficient. It feels powerful. Ai will probably turn more people into workaholics than it will generate lazy people.
English

@graceongrid The good news is you’ll get tons more practice over the next 2 decades 😆!
everand.com/book/251375722…
English

Technology has a funny habit of doing the opposite of what people predict. Spreadsheets didn’t eliminate accountants, they created more accounting. Email didn’t reduce communication, it multiplied it. AI will do the same thing to contracts.
For the past year I’ve been helping @scottastevenson. The reason is simple: contracts are one of the core technologies of civilization. They let strangers cooperate safely. When contracts are slow and expensive to produce, deals don’t happen or they happen without the right protections.
Small businesses feel this the most. Good lawyers are essential, but they are often too expensive for the number of agreements a growing company actually needs. Founders either delay formalizing things or pull questionable templates from the internet. That works right up until it doesn’t. I run businesses and have found myself in a few litigations; good contracts are a lot cheaper than bad disputes.
Scott Stevenson@scottastevenson
We’ve raised $40m in addition to the $50m we raised last October. We’re seeing record-breaking growth in 2026, with lawyers booking 410 demos of Spellbook last week. We now service 4,000+ in-house legal teams and law firms in 80 countries. theglobeandmail.com/business/artic…
English

@MParakhin Generally agree. We used to obfuscate on purpose for many shops because of those pesky bots. We'd embed honeypots into the html to trick bots and have them accidentally checkout $10,000 sneakers. Stores should still be allowed to go into "human" mode.
English

A product page can have every piece of information a buyer needs, but if it lives in JavaScript logic, an agent can't read it. Schrödinger's product data: it is there and not there, depending on who’s asking :-)
That's one of the distinctions between data that's human-readable and data that's machine-parsable. Here’s an interesting look at how to solve this.
shopify.com/enterprise/blo…
English

@fahdananta Hahah I’ve had too many life and death close calls at sea that you triggered me! But as you were my friend 🙏
English

@jmwind Fair, i'll stick to tweeting and leave the sailing analogies to you 😂
English

@jmwind There's a video of this insanely buff speed cyclist with 30-inch thighs barely being able to toast a single piece of bread
youtube.com/watch?v=S4O5vo…

YouTube
English

@dps Congrats on the launch! Can’t wait to try.
I’m open to building the 3d cad room so that we can bring agents IRL. Take a pic of something broken in the house and get it designed and printed at home.
English

Jobs called computers "bicycles for the mind" -- tools we could shape to our will.
But they never were. Until now.
Every morning an agent preps me for my day -- calendar, news, last 24hrs of Slack -- in a personal podcast. I made it by asking. Same for hundreds of other things.
Launching @dreamer in beta today.
That 🧠 bicycle, finally.
dreamer.com
Dreamer@dreamer
Introducing Dreamer. A place to discover, build, and enjoy agentic apps. It’s your home for personal intelligence. Now in beta. Sign up👇
English

@pammcgugan @RaymondGagne10 If you think my boss would fire me, you haven’t met my boss. He’s unbearable.
English

@jmwind @RaymondGagne10 This is so unprofessional! If I was your company I would fire you asap. Your lack of professional respect for confidentiality is appalling.
English

I interviewed Mark Carney for a job.
It was 2021. We were looking to bring in a board member or executive who could help us think about entrepreneurship in the context of the global economy.
I was asked to interview a guy named Mark. I Googled him. He had a decent CV.
In our call after brief introductions, he jumped straight in and said, “I see on Strava that you’re a runner and you live in...” It threw me off a bit. He had clearly done his homework. Borderline stalking. We talked about running for a while. He was good at small talk. He casually mentioned that he runs marathons now and then.
Mental note: he does his homework. Also, is he distracting me so I do not get to my questions? I started worrying he would try to schmooze me instead of getting into anything substantive.
In my defence, I had done some homework too. I had not stalked him on social media, but I had read several articles he had written. I learned that he leaned strongly globalist and believed that most of our challenges do not respect borders and will only be solved collaboratively.
I wanted to see if he could argue against his own position. More specifically, I asked when countries should invest in self-reliance. For context, I explained how, in software, we have learned that purely centralized systems fail in obvious ways. But overly distributed systems fail too, when small issues propagate across too many dependencies.
I asked whether societies behave the same way. Are we sometimes too decentralized? When should countries accept less efficiency and invest in more centralization or self-reliance?
He smiled. I could not tell whether he thought the question was childish or whether I had annoyed him. Then he broke the silence and said, “This is a great parallel. Give me a second to think about it.”
We ended up having a great conversation. That said, it took him a lot of words to make his point. Professional talker.
He liked the exercise. I could tell he had spent so long defending global collaboration that he had not fully prepared for this angle. I did not know it at the time, but he was in the middle of writing Value(s), which is essentially an ode to global cooperation.
We went over time. It did not faze him. He cared about finishing the discussion. At that point he was improvising, and it felt natural and fun. It was a genuinely thoughtful discussion. I learned a lot.
In the end, he did not join us. But we all wanted him to.
When I see him in his current gig, a small part of me laughs that I might have helped warm him up.
The more I think about that conversation, the clearer it becomes that he probably did not want his current Prime Minister role. Not in the way people want promotions or titles. Some people spend a lifetime preparing for problems they hope never arrive. When the moment shows up anyway, they step in. Not because it is appealing, but because it is necessary.
This just happened to be his moment.
English

@tomnewton I'll have to try this and let you know. My challenge is that I'm obsessive over doing one thing at a time. Even with AI, I'll wait and think about the problem while it works. I tried to run parallel agents and my brain exploded swapping between contexts.
English

9/9 The mental model shift: instead of isolating everything from your machine, you make your machine itself reproducible. One declarative config file. Simpler, faster, and you stop fighting your own hardware. Sometimes the answer isn't more abstraction—it's less. ps. I really had to channel my inner @jmwind for this post. :)
English

1/9 We just mass migrated our dev setup from devcontainers + git worktrees to devenv.sh (Nix). Here's what we learned and why we did it 🧵
English

My main focus is always on machines (especially now—making sure Shopify’s infrastructure is in tip-top shape for BFCM), but humans, it turns out, are still pretty useful :-) toloka.ai/tendem-benchma…

English

Today my son started his first day at his full-time job as a software developer. Exactly 30 years after my own first day of full-time work. Proud papa.
Thirty years ago, my parents didn’t personally know anyone whose job was “software developer.” If you believe the news today, this has been the shortest career in human history before vanishing. Barely outlasting medieval leech collectors. Or maybe it’s just getting started?
Thirty years ago, I spent my first week chasing segfaults in a crusty C codebase, stepping through gdb, and fighting a cursed Makefile. My son installed a few new skills, set up his MCP servers, and asked @stillaai to summarize 300 Slack threads.
On his first day, I spent today coding too. Same as I did 30 years ago. Different tools, same thrill.
English

@jmwind @BrandonMChu Sprinkle in some devenv.sh and with a couple of smart tweaks you can be running as many parallel sessions as you want on one box.
English

You don't open files to read them anymore. It's just diffs and text instructions. I also like looking at the blast area of changes as sniff tests for the agents going out of line and need a timeout. That's what Warp has optimized around. The next big gap is validation. The Vercel brower is good but need for mobile. Then agent can run the sim and test and ensure there are no regressions. That is where I spend most of my time now.
English







