1/ Most companies reach for AI before their processes are readable by a machine. I just walked away from a client because of it. What happened β and what I'm building instead. π§΅
In 2016 a startup idea needed a development team I didn't have. Today cloud and AI have merged so far that one person with a team of AI agents can build what used to take several developers. The cost of building software has collapsed, so courage is now the scarce resource.
@itsdannyyt I'd like to take a similar path once Apple finally reviews my app. Good luck with everything that comes next, I'll stay in touch and keep following your journey.
@jeffbullas AI, honestly. I spent a weekend playing around with Claude Code and saw how much potential is in it. That was my "holy shit, this changes everything" moment, with literal goosebumps on my forearms. After that, staying put felt like the bigger risk.
@georgeguimaraes@georgeguimaraes honest question, what keeps you on Notion in a world where AI can do almost everything? I dropped it from my workflows, because other tools covered the same things more simply. Curious what it still does for you that nothing else does.
@sherifgjini I picked A in 2016, right after working out a startup idea with a friend. It took me ten years to switch to B. So my honest answer is A first, and then B a lot later than I should have.
@vassallo Optimism alone didn't get me moving. In 2016 I believed in my idea and still took the safe job. What I was missing was exactly that low tolerance, and it took me ten more years to find it.
@nicomoel I'd add one thing from my own case. The knowledge was never my bottleneck. In 2016 I had the idea and the years of experience, but no development team. Today one person with a team of AI agents can build it, so I expect a lot more of these founders.
@georgeguimaraes This took me straight back to my childhood. My agents have names too, Tessa tests my apps on iOS and Android, but none of them is as cool as Splinter running the show. I hope Michelangelo gets pizza when the build goes green.
@karthiknish For me it comes down to who checks. The agent that builds something is never the one that checks it, and what goes live stays my call. And wherever security is touched, a human should always be involved. I wouldn't hand that over.
@harrisonju Good that you measured it. Mine keep going while I'm away too, but a good part of that time is them waiting for me. Anything that needs my decision gets parked and the rest moves on. I'd be curious how much of your 11% was real progress.
Eight months ago I built a trading app framework in one weekend without writing a line of code. My first thought afterwards was that the two-week sprint might be dead.
Today agents take my ideas through concept, plan and user stories with test scenarios. I still group stories into sprints out of habit, but I really plan per feature, because most features are done in a few days.
What I didn't expect is where the time went instead. Most of it now goes into reviews and decisions, and that's the part I'd never hand to an agent.
@mathiasvrb Fair, and that job paid mine too. Looking back, the bills weren't what stopped me though. The safe path just felt easier than the idea, and I didn't dare.
In 2016 I had just quit and was lying in a hotel room with an ice pack on my leg when the news came that I was released effective immediately. I worked out a startup idea with a friend and then took the next consulting job, because the safe path felt easier than the idea.
@harikesh_devops And that is actually the hard part, which everyone forgets throughout all the AI coding hype. If you are capable to create products conpletely by your own, who is challenging your own BS?