This is an extremely important piece of advice. To drive change in a large org, you will need support, which will either come from management, or people on the ground - this is purely people skills. Just writing amazing code and shipping it isn’t enough.

hmm surprising because I had mentioned how to win friends and influence people book by dale carnegie in some other comment on this post and one of the most sticky / remembered lines of this book is that even in a highly technical post , much of the success (approx 70% IIRC) can be because of having people skills.

People skills matter , But I suppose in large organisations. I don't think people skills matter "that much" in open source (other than documentation and hey not messing up like attomatic and being a little bit nicer in general)

Hell we should all try to be a little nicer in general , not just in open source.

Well, I’m not sure that the Linux kernel would have been that successful without Linus’ people skills - granted some it is slightly subpar, but his ability to organize a large number of people and have them follow his lead ( that’s the tricky bit ) is actually quite striking.

Now, you might say that the Linux kernel is a large project, and you’d be right

While this is true, I find myself depressed by the idea of exchanging coding for herding everyone in the right direction.

I’ve been trying to make that switch as it’s the only career prospect I have, but…

I can relate

Would love to hear how others have solved this

In some sense, you don’t solve it. You are correct to perceive you are at a crossroads. And the hard part is knowing that either choice may foreclose some possibilities on the other path.

There’s still plenty of time to change gears, or move back and forth between the two choices. There are also other paths; see my comment above.

Important part is to not ignore the discomfort. I suggest reading Charity Majors’ blog posts around leadership vs engineering roles.

I’m in early 40s and going thru a minor midlife crisis around this. Feeling that I’d wasted time in my 20s building technical prowess only to see myself still be viewed as mostly fungible.

I think my future is moving towards consulting and building my own products. If people want to fight each other to get a seat at the table, be my guest. I see it as a waste of life. I can already build significant projects on my own, and it’s long overdue for me to reap all of the profit that those efforts generate.

Don't feel bad about the technical prowess part. Just consider you are "done" with growing the tech skills and start stacking other skills on top and you'll do great.

Many companies have parallel IC tracks. And even though higher level ICs do need to “herd” more you’re still doing a fair amount of technical work (maybe at least until director equivalent).

> maybe at least until director equivalent

Hence the issue xD I’m at the top of what my track offers, but it feels silly to stay in the same place for the next 30 years.

The director equivalents I know ("distinguished engineeer") still do plenty of individual work if they want to; it's simply expected that they know how to do things in a way that has the same leverage as a director. And yes, that can include temporarily recruiting people to do work outside their team.

I agree and this is good advice that works well. My counterpoint is that if you are very good at:

finding things that could be better or tools that should exist but don’t, and that if materialized would solve multiple problems at once instead of just one; and

explaining why that’s important in a way that everyone can buy into and dealing with naysayers who’ll want to cut you down and doomers who will say it’ll never work; and

have the focus, discipline and grit to make it through a 10k LoC project and land it on the other side while ekeing out code review along the way and not tripping over loose cables and uneven steps of your initial design / architecture / model that was either flawed from the outset or not malleable enough to fix after a month of entrenchment into your code; and finally

once you’re ready, to then go over the entire codebase and lift the old code up into your new way of doing things so that there is no split brain between the old way and your new tool:

then you can make it as a 10x engineer. It is exhausting but so is running up and down a football field, cycling laps around France, or serving tennis balls at 90mph for three hours at a time. Taking the elite-coder staff-engineering path isn’t for the faint hearted.

first i think what you’ve presented is a fiction, in most environments such large projects are surely going to involve multiple people if not multiple team. and then why? unlikely you’re be financially compensated for all the impact you’re really delivering. anything less is just exploitation

My experience is that this is quite common. Especially at startups but also at fortune 500s there exist staff engineers who have no team and just exist working on whatever catches their fancy that day/month. The grandparent talked about a month long project which is where these engineers tend to live in my experience. Its not that they get no help more that they jump around and maybe grab someone from the team when theyre almost done to integrate the new feature with the existing codebase.

I wouldn’t say people skills “always” trump technical skills. Some projects require strong technical chops such that no amount of people skills will be sufficient. If everyone thought this way we wouldn’t have made so many of the recent major tech advances.

Now given a fixed level of technical skills it’s obviously always better to have more people skills, perhaps that’s what you meant.