I learned a lot in my PhD that I wouldn't have learned in my subsequent engineering job: Public speaking with confidence, how to be proactive in your job, how to apply for public funding, teaching experience. How to present results, and how to write. I had exposure to what (at that time) were rare computing resources, which is useful now with in the cloud era. And I learned a ton about my specific field of course - which is probably not monetarily useful, but was a lot of fun and you can't learn it in such depth anywhere else.

As someone with a CS PhD and 7 years in industry, I can tell you that they will teach you different stuff. Unsurprisingly. Hard to compare the "amount of knowledge" though. Even if your only goal is becoming a good engineer it is not obvious to me what the best path is. Yes, you learn a lot in industry the first couple of years, but those things you will also learn when you enter industry after the PhD. So 4 years after a master you are definitely a better engineer if you went straight into industry. But 10 years after a master, comparing 10 years in industry vs 4 year PhD vs 6 year industry it's not as clear cut anymore.

Okay, do it. Indeed I've seen it happen, but... it requires a "certain kind of person" to pull it off. And I certainly wouldn't bet on it as a probability.

It's very easy for newly minted engineers to get sucked into the grind. Most entry level engineering work consists of organizing and arranging things, fitting things together, troubleshooting, bureaucracy, and attending meetings. It's easy to become so busy with those things that you forget your math & theory (or whatever fundamentals you learned in your field). Many engineers are happy to operate the machines (CAD, IDE, etc.) all day long, and get paid a lot.

To learn more in those 3 years requires some combination of potential, motivation, and luck. You might be someone who refused to join the grind, and were strong enough to get away with, or found a great mentor. You might be capable of grinding for 8 hours, then going home and immersing yourself in a fructifying side project.

Granted, a PhD doesn't guarantee any of those things either. It also requires potential, motivation, and luck. There's massive attrition in PhD programs, and they will let you render yourself unemployable. There's no filter at the start of the PhD program that weeds out people who will struggle with the soft skills required for success in higher level jobs.

It will teach you more about engineering. Most jobs aren’t pushing the boundary of CS knowledge, or doing important discoveries

Depends pretty much if the company is doing real engineering, the one accepted by the Engineering Order, and not putting an "engineer" label on coders out of boot camps.

They teach entirely different things. My first 3 years working taught me a lot about the real world practicalities of writing useful software, as a team, that you can charge money for. It taught me very little about doing research or the cutting edge of my field. I suspect a PhD would be the exact opposite.

The biggest difference between working and research is that you're free to just try things out without so much pressure.