It's not an especially crisp piece, but it comes down to this.

- Managers need to pay close attention to where they spend their energy (=time). For example, the CEO should not be spending time picking items for the company snack cupboard.

- At the lower tiers of management, the "what" necessarily becomes much more granular and task-oriented.

- The risk is that managers at this tier end up telling people precisely what to do and how to do it. This is not scalable, leads to frustration on the part of the manager and managee, and burnout.

Better is that the manager describe the desired outcome and set of standards, kpis, norms, etc, that must be met, and let ICs do it any way that fits. (Standards and norms enforce a culture that lets people collaborate more productively. No point letting a dev loose to code up a feature if they turn around a week later and tell you they did it...but in C#, when you're a Rust shop.)

An effective IC should aim to internalize the standards and norms that a manager (and their boss) use, which can be explicit or implicit, take the time to understand the manager's own targets, and offer up ways to make their work product make their life easier. For example, if you know that your manager has to write a weekly report with kpis like code coverage and test cases, then take care to produce easy-to-consume material that they can use. Eg text they can copy paste into an email, a dashboard they can screenshot or link to, etc. You want to be thought of as someone who they "dont have to worry about" and who comes to them with solutions and not just problems.

Sounds a bit Machiavellian to some, but a good manager relationship cam make or break your time at an org.

It is a bit too verbose, but the gist of it seems to be, that delegation is about giving goals and then giving freedom to the delegates to achieve those goals. The article discourages micromanagement(with some bad examples), and encourages that giving general tasks and giving some goals with little bit of hints should be the appropriate approach.

For example, you tell your delegate to post a link on HN, giving them goals that it is properly formatted and correctly posted, a hint that there should be a prominent “submit a post” link somewhere on the navigation bar, so that they can try to figure out about how HN works, how to post, figure out the formatting guide, check when their post was successful.

It is discouraged to set out the entire tutorial steps(micromanagement with too much how), like type this url in address bar, signup/login by clicking the anchor, click submit post/link, type in the content …

The aim is that, the delegate learns how to achieve the goals with as little hint as possible, so eventually if you ask them to post a link(say doing a SHOW HN) or a job, they can achieve the goals on their own.

Some sound advice but too much text to explain.

(Edit: reduce word count and formatting update)

I'm a new manager and I'm really struggling.

Reading this article put into words my own struggles and the candid feedback I've received.

I'm certain that had I read this a few months or years ago, I would be completely lost.

But your instinct to use these pieces to understand your managers’ thinking is really cool

I'm not the author, but thank you for sharing that this article is not landing clearly. I write pieces like this occasionally and your comment made me realize how easy it is to miss the point.

Here's how I would put it, looking at it from your perspective:

You are hired to do some work. Often, you know what needs to be done. Sometimes, your boss needs you to do something else. He needs to delegate that something to you.

When he delegates, it means he tells you to do it and explains what to do. You need to be clear what he wants done, when he wants it done, and how he wants it done.

To explain "how", he needs to tell you what he expects to use to get the job done. If there are limits, he needs to tell you what they are. He should not provide every details - that is micromanagement. He should think about what might not be obvious to you, and provide just those details. This is hard and bosses get this wrong a lot of time. It's also hard from your side, as it takes effort to really deeply understand what they want from you.

From your side, the most important thing is to listen, try to understand what he wants, and ask questions if something is unclear.

Example: build an integration with XYZ API. Don't use the REST library we've used so far, as we will be deprecating it in the near future. Find another library, but don't spend more than a day researching it, as the decision is not THAT important. You get to decide which library to use, but let me know, so that I have a chance to review it and comment in case it does not meet some of my requirements. If you don't hear from me, assume everything is OK. We need the integration to be ready for beta testing in 2 weeks. The functional specifications are in Jira, and XYZ Corp. has great documentation. Also, Bob Franklin from XYZ is really helpful in case the docs are missing something.

You responsibility is to see if this makes sense and gives you everything you need to get started. If any of this is unclear, contradictory, or weird, ask without delay.

This is hard. If you wait until you start to carefully read through what he delegated, and you have questions or concerns, it will delay the project. You can't start until those are cleared up. That is why it's super important to try to understand the instructions and think whether everything is clear as soon as possible.

Perhaps a better way to phrase what the author was saying is that you need to tell your reports the things that you know more than them about (which incidentally usually is direction), but should avoid trying to prescribe things that it is their job to know more than you about (which incidentally covers most things about the codebase they see every day).

"I'll run so far ahead of my team that I won't need their advice" - last spoken words of a new lead who died trying to do eight jobs faster than the people hired to do them. ;)

Although how and what are connected, they are not the same. For example, 'I would like to become fit and healthy.' 'What' is getting fit and healthy, but 'how' has almost endless possibilities, each with their own success level and compatibility with you and your lifestyle. Just like nouns and verbs are connected, they are not the same.

What and how are most definitely not the same thing. They aren't necessarily even related:

If you tell people where to go, but not how to get there, you'll be amazed at the results.

George S. Patton

Delegation doesn’t demand strict separation between "what" and "how"

IDK, but I appreciate the domain name being honest about the point of it all.

Well HN articles should be a good mix between serious, concise & boring on one side and on the other hand easy to read articles like these.

While reading I still have some cpu cycles left to think about the broader context and how this applies to my own situation.

"Savant" is an iffy word choice I'd agree, but lots of programmers tend to take instructions very literally. Strict logical thinking is part of the job, and most low and mid level managers are (or recently were) also programmers.

When people are giving me instructions that seem overly detailed to me, I usually assume they have a reason for giving those details because otherwise they would have left it up to me. If I disagree with some design choice or don't understand I'll usually challenge it or ask for that reasoning, but I don't always have time or energy to challenge every last thing. Nothing is more frustrating to me than when I ask for the reasoning and they can't (or won't) explain any motivation behind it. That goes both for my superiors and for my peers during architectural discussions and things like that.

And of course, there's a very high rate of autism in software engineering. Autistic people tend to take things more literally than most.

I believe the savant in this case is the computer that takes instructions literally?

I think there’s a long-standing stereotype that programmers are "literal" thinkers who operate only with exact instructions