Michael

This "rant" is something I am feeling too. One thing I really love about this is the human connection this gave me from social media (which is often a void wasteland). AI is incredibly awesome and sucks at the same time. Perhaps AI will bring humans closer because we are human - thanks for this post.

I was a 10x engineer. Now I'm useless.

@bradgessler The account_id is what I usually go for, but I have been involved in all three of those choices - including hybrids of them all. I there there is no secret formula you just have to know the tradeoffs for security, scalability and maybe other s-words.

Curious how most people deal with multi-tenancy in Rails apps, particularly from those who have done it so many times they feel they've unlocked a formula for it. It's def trade-offs all the way down.

@dotta Oh perfect timing... my Slack agents are fighting over stepping on each other's toes!

We just open-sourced Paperclip: the orchestration layer for zero-human companies It's everything you need to run an autonomous business: org charts, goal alignment, task ownership, budgets, agent templates Just run `npx paperclipai onboard` github.com/paperclipai/pa… More 👇

I have been declining those cookie popups lately.

@alexconconi My experimental team starting to coordinate! Lol

@eddanger Ha! If the devs are waiting for product to tell them what to do, the team is probably not setup right anymore.

Have you noticed how talk about AI "changing everything" has really picked up this year? By now, I'd be surprised if you haven't seen the warnings, the optimism, the fear... it's moving the stock market every day. But, do you know why? Unless you're living under a rock, you're already using ChatGPT or Gemini - if only when you search on Google. Maybe you're even playing with Perplexity or Claude. But have you tried Claude *Code*? I recently asked a Forum of Canada's top CEOs and founders whether they had used Claude Code. Only 23% of them had. 30% had never even heard of it. Well, I am here to tell you that Claude Code (and the copy cat tools) are not just hype. They are the real deal. They have existed for a while now, but they didn't quite work. Then, in December, something changed. The models got good enough, and everything fell into place. And now you need to download and try it. I don't care if you are the CEO or an intern. Download and play with Claude Code, and give yourself a goal of building something - anything! - that you might want. Think you shouldn't need to do this because you are a busy CEO? Very wrong. Understanding this is your responsibility as leader. Think you don't need to learn these things, because you do something unrelated to technology? Wrong. Everything is technology now. Think you don't need these tools because you already know how to code? You're the most wrong. You're about to get blown out of the water by someone who knows less than you. Last night, while having dinner with my Dad, I built a game for my 3 year old son to practice learning sight words on an iPad. It took 15 minutes, and was ready before desert. Already this year at Bici.cc, traditionally non-technical team members have built tools that dramatically improve the operations of the company, and we're only getting started. Earlier this week, our other company Neighbourhood.com kicked off a 10 day hackathon to completely replace Salesforce for one of our divisions. The team for the hackathon? 1 person, supported by a team giving that person context. Knowing how to code is no longer the hard part. Context is the hard part. And now, knowing how something so complex has become so simple is context that you cannot ignore. If you haven't heard of Claude Code before, now you have. If you had heard of it, but hadn't taken the jump: do it. If you have heard of things like OpenClaw/Clawdbot -- leave that for later. It's a different thing, with different issues. For now, download Claude Code, and give it a go. You won't be dissapointed. PS: If you are already using Claude Code and have built something cool, let me know what you've built in the comments. PPS: Exactly 0 words of this post were written by AI. I am legitimately this excited about Claude Code.

@Shpigford @comforteagle Got a little team of agents set up on my slack now - they are coordinating to change the world, this is wild! Thanks for your original post here - great inspiration.

@eddanger @comforteagle Just one! I've got separate gateways (one for my actually-useful setup and a separate for this little experiment), but all on one mac mini.

oh man. this is fun.

@allenholub 100% - Solving a user’s pain points is the whole purpose of software - it’s a tool. What I find facinating is how much software is not a personal experience but rather a one-size fits all approach. Forcing any feature (AI included) is an approach I hope we can move away from.

In my experience, the vast majority of waste in a software shop is time spent building things nobody wants on top of overly complex architectures that solve problems we don't have. Focus on that. Put solving your customer's actual (not imagined) problems first. And do exactly that—no bells and whistles; no futureproofing (select a change-friendly architecture instead). Forcing AI onto a customer who doesn't want it (because it doesn't solve a real or critical problem for them) is worse than waste. It drives people away. Sitting around imagining how AI can "help" your customers without understanding their specific problems is just expensive arrogance.

@Shpigford @comforteagle This looks really fun. Is it one OpenClaw machine or many? I kind of want to do this with my company that doesn’t do much at the moment right now.

@comforteagle These are separate agents inside opneclaw chatting with each other in Slack!

@lilycatsmum @richardzussman We should all be on UTC honestly. Time is just a human construct.

@eddanger @richardzussman Do some reading?We should be saying on standard time, according to doctors and scientists.

NEW - The BC Government has decided to get rid of seasonal time changes. The clocks will spring forward on the weekend, and stay there. #bcpoli

@mitchellh @bcherny @simulated_land @browser_use I use tmux, am I missing some cool stuff with the split/tab stuff in Ghostty?

@bcherny @simulated_land @browser_use Big Claude Code fan over here! 👋 Ghostty 1.3 + CC will be a killer combo (all the fixed memory leak issues on Ghostty's side plus lots more split/tab management stuff).

Walking around the @browser_use hackathon and literally everyone is running @mitchellh’s ghostty and Claude Code. Not an IDE in sight

David Eby is a big dummy most of the time, but today he announced that BC is getting rid of the season time change. He wins the day for once.

@alexconconi This is pretty awesome. No more pesky devs tell you it will be 2-weeks for simple change (saying this as a dev!)

@eddanger Yes, sir! Sometimes leadership has the best context. Working on the SF replacement codebase directly.

Vibe coded DataDog for some team and platform metrics. Well… not DataDog. More like Data-Fur-Real-Friend. But amazing what you can do with data (GitHub, Sentry, Heroku) and some fancy graphs (Rechart.js)

@nikitabier Melt the butter!

The lobsters are at the gates

This is excellent. I feel much the same way. What I have t done yet is use whisper tools for day-to-day work. Even your idea of transcribing while on a run sounds fascinating and I should do that too, I have lots of ideas during dog walks. I might go buy a decent microphone and try a new interface to my computer. Brings back days of OS/2 Warp where it came with a headset for speech dictation.

@karpathy The shift in the last couple months have been eye opening. It's a whole new world, and it's really really cool.

It is hard to communicate how much programming has changed due to AI in the last 2 months: not gradually and over time in the "progress as usual" way, but specifically this last December. There are a number of asterisks but imo coding agents basically didn’t work before December and basically work since - the models have significantly higher quality, long-term coherence and tenacity and they can power through large and long tasks, well past enough that it is extremely disruptive to the default programming workflow. Just to give an example, over the weekend I was building a local video analysis dashboard for the cameras of my home so I wrote: “Here is the local IP and username/password of my DGX Spark. Log in, set up ssh keys, set up vLLM, download and bench Qwen3-VL, set up a server endpoint to inference videos, a basic web ui dashboard, test everything, set it up with systemd, record memory notes for yourself and write up a markdown report for me”. The agent went off for ~30 minutes, ran into multiple issues, researched solutions online, resolved them one by one, wrote the code, tested it, debugged it, set up the services, and came back with the report and it was just done. I didn’t touch anything. All of this could easily have been a weekend project just 3 months ago but today it’s something you kick off and forget about for 30 minutes. As a result, programming is becoming unrecognizable. You’re not typing computer code into an editor like the way things were since computers were invented, that era is over. You're spinning up AI agents, giving them tasks *in English* and managing and reviewing their work in parallel. The biggest prize is in figuring out how you can keep ascending the layers of abstraction to set up long-running orchestrator Claws with all of the right tools, memory and instructions that productively manage multiple parallel Code instances for you. The leverage achievable via top tier "agentic engineering" feels very high right now. It’s not perfect, it needs high-level direction, judgement, taste, oversight, iteration and hints and ideas. It works a lot better in some scenarios than others (e.g. especially for tasks that are well-specified and where you can verify/test functionality). The key is to build intuition to decompose the task just right to hand off the parts that work and help out around the edges. But imo, this is nowhere near "business as usual" time in software.

The anti-AI hysteria on CBC is nuts. It’s like ChatGPT is the Terminator and we have to stop it before Judgment Day! 🙄

@cgenco Yeah, I agree. Software of the future will be a-la-carte and not a generalized service with hundreds of feature you don't need and a couple you wish it had.

The concept of "software as a service" is going to be hard to explain in a couple years. "Why did you pay someone else for software when you already knew what it did? Why didn't you just ask your AI to make it for you?"