THIS IS HOW A SENIOR ENGINEER ACTUALLY SCALES THEMSELVES WITH CLAUDE CODE the biggest change with AI isn't coding faster. it's where you actually spend your time now. more detailed prompts, more code review, more planning, less typing, etc. here's the workflow: this guy has been shipping code since... the days of cgi and perl. he uses a compound engineering plugin that runs 5 separate agents on every task. one brainstorms, one plans the technical implementation, one executes, one reviews, one checks different verticals. every step is documented in markdown files. it's slow and way more waiting. but the output quality is way higher because each agent is focused on one thing. then the REAL multiplier is in git worktrees if Claude Code made you 10x faster, worktrees multiplies that again depending on how many agents you can manage in parallel his team runs 4-8 Claude Code sessions at the same time across different worktrees with each one working on a separate task. the skill is managing multiple AI agents in parallel without losing track, that's the next evolution of engineeringshow more

App is created by @nasherasher Credit where it’s due

i made exactly this with 15 specialised agents

NEW ERA ENGINEERS ARE PROVEN TO BE MORE PRODUCTIVE with the right tool. is the one for orchestration in a clean UX UI and it’s FREE

You can do this with a graph and mutli llm provider all in one like copilot to connect them with a good organisation: But beautifull UI:

This is Absolutely Brilliant however there’s one quite Crucial point being overlooked here. Persistent Context for the agent, a Memory! CPR solves this with just 3 skills agents themselves can use!

wait this is actually smart the worktrees thing makes so much sense. I've been running one Claude session at a time like an idiot when I could be parallelizing across features how's he handling merge conflicts though? or does the planning agent prevent most of that upfront

The bottleneck nobody's talking about: your own review bandwidth. Running 4-8 Claude Code sessions in parallel sounds like a multiplier. But if you can't review 8 PRs worth of AI output without missing things, you're not 8x faster. You're 8x faster at generating bugs. The constraint shifted from writing to reviewing. That's the skill to train now.

This is where things are going. Not better tools. Better orchestration between them.

sounds like a lot of overhead just to feel productive the best multiplier is still a clear thought, not more agents

the bottleneck isn't coding speed anymore, it's context management across parallel agents. the engineers who figure out how to keep 6 sessions coherent without losing architectural intent will eat everyone else alive.

@grok what app is this?

so we automated the typing and discovered the actual job was thinking all along. wild that it took AI to make engineers do more engineering

yeah the planning part has gotten way more important. spent 3 hours yesterday just writing a solid spec before touching any code. the output quality gap is massive when you actually do it

What’s the name of ui that wraps around claude code? Is it public?

@om_patel5 yup, the typing part feels like the smallest piece now. more about strategy and making sure the code fits the bigger picture. totally different mindset shift.

AI doesn’t replace engineers, it shifts their leverage from typing to thinking

The real shift: engineering is becoming project management. You spend more time writing specs, reviewing output, and deciding what NOT to build. The best solo founders right now aren't the fastest coders. They're the ones who can run 4 AI agents without losing context.

@grok what is this app?

Less typing, more judgment. That’s the real shift.

using the voice feature to say your prompts is clutch, you get way more detail out

But the 4/5 agents he is running should be able to communicate with each other to get to know what was the thinking behind the execution of that step Isnt it

This is the standard for any ai tool this days, whats so hyped about his setup? Conductor does this by default, Codex does this, does this mostly by default and claude code also has this native 💁‍♂️ hype boys…

"the biggest change with AI isn't coding faster, it's where you spend your time" translation: the same 10 hours, just redistributed from typing to arguing with a chatbot about why it deleted your entire import block

What app is this??

the time shift is wild. i spend way more time thinking through the problem and reviewing what claude spits out than actually writing code. feels like being an architect instead of a carpenter

how’s Compound Engineering plugin different from Obra’s Superpowers?

@grok what tools is he using for managing concurrent sessions?

spending more time on prompts than actual code is weird but it genuinely works lol

What's he building specifically

What I want to know is that sidebar that shows in the terminal...what is that and how do I get it? I have Claude installed on my mac and just type 'claude' in the terminal and boom I have a session but I want that side bar that shows sessions and activity

Can’t see shit on mobile

AI doesn’t replace engineers, it shifts their leverage from typing to thinking

you don't need all this shit, all apps like that are trying to solve imaginary problems. All you need is good terminal and multiplexor with the right configuration