I'm running Claude Code, Codex, and Gemini CLI in parallel.
The worst part isn't the token cost. It's the tab switching.
AI DevKit's Agent Console fixes this. One view, all your agents, send messages without leaving, quick hotkey to focus when you need detail.
Works in iTerm2, Ghostty, your default macOS terminal. No new app.
npx ai-devkit@latest agent console
Classic โbig upfront planningโ still fits poorly inside daily/weekly/bi-weekly iteration for most product work. When that uncertainty is high, locking in a large plan and then executing for weeks creates lag between assumption and evidence. AI makes the execution cheap, but the cost of a wrong assumption still grows with how far you travel before checking.
Where AI does change the calculus is that you can now afford more precise thinking inside short cycles rather than instead of them. Because an agent can turn a reasonably detailed spec into working code in minutes or hours, the lag between writing a plan and seeing its consequences collapses. What used to take months of waterfall lag can now be an afternoon of revise-and-rebuild. That is not traditional waterfall; the plan is never frozen and the feedback is continuous.
Fast feedback, responding to change, customer collaboration, and working software over comprehensive documentation become more important. Fixed two-week sprints and story-point estimation may matter less; the ability to change direction based on evidence and to keep shipping increments that produce learning still matters a great deal.
So big upfront planning can be useful for certain classes of decisions and can be done more rigorously than before. It fits continuous iteration when it is scoped, revisable, and subordinated to the learning cycle rather than allowed to freeze the path from idea to validated production. The practical question is how large a plan can we write, execute with AI, and validate before the assumptions are likely to be wrong? That size is usually smaller than classic waterfall and larger than pure just start coding.
Hear me out:
We hated "waterfall" because after Big Upfront Planning, implementation took FOREVER, and by the time we were done, the plan was out of date
But with AI, implementation is rapid (100x faster, or more, than before)
So Big Upfront Planning can now make sense?
Classic โbig upfront planningโ still fits poorly inside daily/weekly/bi-weekly iteration for most product work. When that uncertainty is high, locking in a large plan and then executing for weeks creates lag between assumption and evidence. AI makes the execution cheap, but the cost of a wrong assumption still grows with how far you travel before checking.
Where AI does change the calculus is that you can now afford more precise thinking inside short cycles rather than instead of them. Because an agent can turn a reasonably detailed spec into working code in minutes or hours, the lag between writing a plan and seeing its consequences collapses. What used to take months of waterfall lag can now be an afternoon of revise-and-rebuild. That is not traditional waterfall; the plan is never frozen and the feedback is continuous.
Fast feedback, responding to change, customer collaboration, and working software over comprehensive documentation become more important. Fixed two-week sprints and story-point estimation may matter less; the ability to change direction based on evidence and to keep shipping increments that produce learning still matters a great deal.
So big upfront planning can be useful for certain classes of decisions and can be done more rigorously than before. It fits continuous iteration when it is scoped, revisable, and subordinated to the learning cycle rather than allowed to freeze the path from idea to validated production. The practical question is how large a plan can we write, execute with AI, and validate before the assumptions are likely to be wrong? That size is usually smaller than classic waterfall and larger than pure just start coding.
it's shocking that software engineering has transitioned to 100% AI written in such a short time.
it's more shocking that somehow we, the SWEs, are all just continuing on as if nothing happened. It writes C++ better than I ever could, but the problem of making sure it works didn't magically disappear. we are just shipping enormously more volume at higher quality.
Glad that this idea of having multiple layers of agents orchestrating other getting more and more adoption https://t.co/0KuhM72sGG
If you are finding a seamless solution for building your setup, give ai-devkit a try
I've changed how I work over the last few weeks.
I now have one lead agent for each area of focus, running all the time, and a chief of staff agent that oversees them. Each lead manages its own sub-agents that write code, test it, chase reviews, and coordinate with team members.
Everything I need to decide or review comes into one prioritized queue, a simple local web app. Each product review includes screenshots and a video demo of the feature. I clear the queue in batches, mostly by voice and keyboard shortcuts. The leads also bring back ideas I hadn't considered.
Working from one queue, instead of switching between dozens of chats, has cut a lot of mental overhead. The system keeps learning my preferences and distilling my judgment into a "Brian brain," so its recommendations and first drafts get better before they reach me. The goal is fewer corrections per change.
Every change still goes through the same human review and security checks. It is eliminating tasks, not human judgment.
What an exciting time to push the frontier of productivity, especially as we build the economy for agents (agentic finance). We should all keep pushing in this direction.
@thekitze my eyes are bleeding though.
I think this would be a good designer test. "find all the ui problems in this one image".
- vertical and horizontal alignment is missing
- buttons are rounded and square. also various sizes
- spacing
- ...
send this to your agent to fix :)
I get this. I spent a few months struggling with the same feeling. Seeing machines handle things Iโd spent years learning and doing was honestly depressing. I kept asking myself what would make me relevant anymore. I still donโt have a settled answer.
But Iโve also started to see the other side. I can learn faster, explore several ideas at once, and build things I previously couldnโt even attempt. Things that used to take a whole team are now within reach of a few people. That has made me curious about building again.
I keep reminding myself that execution is part of our job, but it isnโt the whole job. We build things to solve problems for people. Understanding what someone is struggling with, deciding whatโs worth building, and caring about how it fits into their life still matter. To me, thatโs the human soul of a product: someone cared enough to make it useful.
I canโt promise AI will never do that too. If it eventually understands us and brings that same care to what it builds, then weโll have much bigger questions to work through as humanity. That isnโt something software engineers alone should have to figure out.
None of this makes the uncertainty disappear. But being able to build things I care about, which used to feel out of reach, gives me something to look forward to.
there will remain some niche holdouts, where some bespoke, artisan, well crafted piece of software is still appreciated.
but i'm too tired to find such a niche.