Building fast with AI ?
Your codebase is accumulating entropy. Findout with vibecop.
Check your Codebase against 70+ industry standard vulnerabilities for free ! Or try with sandbox repository.
Engineering Intelligence Platform
https://t.co/28INkUbqWm
Claude design and Stitch by Google are the opposite end the spectrum.
One has good taste and the other is sloppy.
One has decent tooling to tweak around the generated design where as the other is 'turn based' and feels like 2024.
Where's the best of both worlds ? Kimi ? Noo
Tokenmaxxing to the point, named all frontier as family members and now youngest daughter sonnet has to get family's approval before actually something. Daddy fable is proud.
Releasing Feature-Factory as part of vibecop's MCP skills.
Feature factory - Runs a full spec review, builds feature into a graph, breaks your new feature's spec into nodes and runs agents to implement and review, raise and review PR.
Claude/cursor/codex orchestration. Try ;)
New in Claude Code: your sessions can now message each other.
Instead of having to re-explain yourself in another session, you can now tell Claude to do it. It sends a summary (not your history or files), and the other session picks it up mid-task.
60% all tokens you spend on writing code is a downstream waste.
The larger the codebase the waste increases.
Even with the best skills, efficient thinking and experienced engineers, more than 30% of your tokens are downstream waste and it's not garbage collectible.
@vaibhav__upreti@vaibhav__upreti why don't you give a shot at https://t.co/28INkUbqWm ? Keen tto give you a demo and get you set up if vibecop can help you build cleaner.
@solopribuilds Write clean spec, get approvals before your AI writes code.
https://t.co/srlqwhDlSd
25 detector agents, 7 tech stack support, codebase semantic brain, Findings listed with fix it prompt and get Architecture integrity score. Dm for a demo :)
Tokens are 90% cheaper than two years ago.
Spend tokens writing better plan and spec, grounded on existing source and architecture must be mandate than doing line by line PR.
Solution Architects and Staff engineers used brainstorm on features and how it impacts existing architecture, feasibility and blast radius etc., With AI first coding, these responsibilities are often offloaded to the agents.
Less code to build a feature must be a benchmark