Andrew Ng: With AI making software engineers much faster, product management is increasingly the bottleneck.

"One of my teams came to me, and proposed not to have 1:4 PM/engineers, but to have 1:0.5 PM/engineers"

Many of you have asked me about this:

𝗜𝘀 𝘁𝗵𝗲 𝗳𝘂𝘁𝘂𝗿𝗲 𝗼𝗳 𝗣𝗠 𝗮 𝗹𝗼𝘄𝗲𝗿 𝗣𝗠:𝗲𝗻𝗴 𝗿𝗮𝘁𝗶𝗼?

I doubt it.

After talking to so AI PMs every week, what I've heard is that the nature of the PM role is really changing:

1. Software engineers are becoming product engineers

They're happy to do work without a detailed PRD and design, they just let AI build a simple design for them.

2. This will lessen the burden for "old style PM"

In the old era, the time to build a feature was weeks. So you wanted to make sure it's right. That responsibility fell on PMs. Now that the time it takes is days, eng are more open to taking those decisions themselves.

3. The PMs of the future are AI-powered

The PMs of tomorrow are going to be 10x more productive than their counterparts with AI agents, Ai copilots, and AI doing all the scaffolding of their work.

For these 3 reasons, I reckon that the ratio isn't going to change that much.

Yes, software engineers are 10x more productive with AI.

But PMs are just yet to catch up.

So how do PMs master AI to be more efficient? This is where I think the longer context of the Andrew Ng clip really helps. Because he lays out the roadmap himself!

1. AI Understanding: news.aakashg.com/p/ai-f…

2. AI Skills:

• AI PRDs → news.aakashg.com/p/ai-p…

• AI Strategy → news.aakashg.com/p/ai-p…

• AI Discovery → youtube.com/watch?v=wGu…

3. AI Tool Use:

• AI Prototyping → news.aakashg.com/p/ai-p…

• Vibe Coding → news.aakashg.com/p/vibe…

• AI Copilot → open.spotify.com/episod…

• AI Agents →

The future AI PMs are going to be technical folks who are experts at using AI to improve their efficiency.

Follow Aakash Gupta for daily advice to get there.

📌 Want to become an AI PM without experience? Comment 'transition' and DM me. Repost to cut line.