Bro 🤦‍♂️
Yes, you can code 8x faster with AI, but can you review 8x faster? No. Not without being sloppy.
Oh, let me guess—you'll use AI to review too?
I use AI to work, and I've worked for two US tech companies (ex-Uber, current Pinterest) where we had access to all AI tools plus direct training from frontier AI labs. AI is unavoidable for our job. It's deeply integrated into our workflow, and without it we'd take longer to ship or even find internal docs.
Yes, we can code 8x, 10x, even 100x faster. But the bottleneck is reviewing. The more code AI writes, the harder it becomes to understand what it did, why it did it, and how it integrates with the rest of the codebase.
In real terms, if you write code 10x faster, you're maybe 2x more productive because of the time you'll spend reviewing and wrapping your head around it.
8x code delivery is dangerously "vibe-coded startup SaaS" level of "productivity."
Your comment makes me skeptical of Magi's safety. You responded well to the proposal, but now you're doubling down on AI to ship even more. That worries me.
If you think 8x code delivery comes for free, that's concerning. Combined with your earlier comment about those three vulnerabilities Claude found being known but unaddressed—by the time they're fixed, there will be many more. And I'm afraid you'll address them with even more AI. Why would you ship code with known vulnerabilities? Why not fix it first? I know why: because that requires reviewing, and it sounds like you really want to ship 8x more code, regardless of quality. That scares me.
I do agree AI can produce better, faster code. But if AI makes you 8x more productive, you're not actually 8x more productive—you just look 8x more productive. That's dangerous, and it makes me skeptical. You almost had me with the proposal, but you just pushed me away.
I really hope Magi works, everything gets fixed, and your approach is right. I hope I'm wrong. But as it stands? What you're saying is scary. It sounds a lot like my friends who just discovered vibe coding.
Edit: ran ecency AI to fix typos and grammar. Sorry for the excessive dashes it adds, I will take them over grammar mistakes.
RE: Ecency Proposal Follow-Up: Q&A and Roadmap Clarifications