I cant work for someone who doesn't understand what I do.

An unused sword rusts in its sheathe.

I remember working for a gent years ago, who was stressed out that my output was so low. He declared "I started this business in my living room let me show you I can do any job in this building"

He came to my workspace, where I had 20 servers stacked on my workbench. He looked at them. Attached a single power cord. And then wandered off telling me he could definitely do the work if he wanted to.

I dont think a manager of software engineers needs to code the application he manages, but he should be continually coding something to remain sharp.

TBH one of my current clients produces hardware and software, medium to large enterprise with close to 200 staff. Their CEO can operate all their products, operate the machines that place chips on the circuit boards, operate the injection moulding machines, write SQL queries to pull data out of their CRM and write code. He tries his best not to do it, but he maintains the skills. That's the goal I reckon. Someone who understands the job all the way to the top.

Strong yes for 3 reasons

1. Reducing dev friction.

When I had managers who coded they were ruthless about removing friction in the dev and deployment pipeline because they had to deal with it too. If build times went up, deployment infrastructure broke or someone’s PR broke dev they would roll it back immediately. If someone consistently blocks PRs the manager noticed the trend and would address it.

2. You get a much better sense of IC’s contributions by writing code.

There are ICs who play politics very well and sell themselves but that set is not the same as the ICs who deliver. If you are writing code you start to notice which ICs have written key features, built critical APIs or worked on hard problems because of comments and Git blame.

3. Understanding your codebase.

I hope most managers have solid CS and engineering fundamentals but that is a necessary but not sufficient condition to grasping the full picture. There’s a reason it takes time to ramp up to full productivity on a new codebase. If you work in the codebase and have had to use that one annoying but critical library or dealt with that tech debt from 2 years ago then you know what is hard and what isn’t. I’ve found when a codebase has a quirk that makes developing certain features hard all of the non-technical people keep forgetting why we can’t do that thing and all the technical people have it burned into their brains.

I've seen plenty of managers who increased team output 10% for a few months by coding and completely lost the thread in the meantime. That's how your entire team gets laid off. Leadership doesn't care how much code your team writes, they care that you're working on the right stuff.

An engineering manager's job is: take long-term ownership for the performance of the team. That might include aligning with leadership, marketing your team's work internally, hiring, performance management, team bonding, planning, retros, oncall coverage etc. etc, although sometimes you'll have a PM/tech lead/HR contact who handle some of these.

Every now and then, your bottleneck really is just writing more code (more common in smaller companies). In that case, jump in.

Should they be able to? Absolutely. Should they exercise this on projects they manage? Probably not.

I ran into this problem years ago. It's not exactly good form to be manager that contributes to the team's project, is at the apex of code review, and is responsible for team performance reviews, all at the same time. It can work, but without other people at your level reviewing your work, you'd be asking the team you manage to call out your mistakes. That's the kind of thing that a lot of people might not be comfortable with, so you're really asking for softball and rubber-stamp reviews on your work. This makes for poor optics: your work always goes to `main` virtually unchallenged, while everyone else has a harder time.

At the same time, you need to be technically competent if you're managing a team while in the review loop. To do otherwise is to create situations where you will lose face with your team. So, sticking to review only is probably the best answer here.

There are workarounds though. It makes sense to maintain a pet automation project just to stay sharp while solving real problems (e.g. every manager needs better reporting). You can also negotiate out cross-team contributions where your work may be reviewed by folks that do not report to you.

No. The amount of work that a manager has to handle to do their job right is incompatible with coding at a professional rate. If you have a manager that codes, then they won't have (enough) time to:

- Write and design your packets (if in a corporation), or your career path (if in a smaller company)

- Align with other teams, get consensus, shield you from politics beyond your level.

- Make long term planning and making sure your team and neighboring teams follow it.

- Listen to you and your colleagues and handle conflicts.

EDIT: forgive me for not reading TFA first. I won't change my comment as it aligns very well with the article. I still think that the answer to the "should code" question is no, not maybe... Let's not try to overload and overcomplicate what "coding" means.

People never discuss company size along with this question.

If you're at a giant company, the answer is likely no, there's enough politicing and paperwork where the highest impact thing to be done by a manager is likely not coding.

If you're at a startup / smaller more nimble org in a big company, the answer is likely yes, if you've gotten to the point where you're a manager, in theory you're a very good engineer and you should spend your time coding, but on things that aren't on critical path. Bug backlog, experimental things with no hard deadlines, proof of concepts, all of these are valuable things. Leading from the front is also just generally good with smaller groups.

Also under discussed by people having these debates (typically managers), is not acknowledging how bad most managers are at coding, especially if their job hasn't required them to code in a while. I see all the time that managers look for any excuse not to code, because it would reveal to their team that they're at best an L4 level coder after being in management for 5-10 years.

My best managers were ex-engineers who didn't touch the codebase. They understood how things worked, and could talk architecture & concepts, but they didn't expect to be able to sit down and write code at our level. Maybe they wished they would have the time/opportunity still, but realistically they were focused on leading.

When I've worked as an engineering manager I've found the advice to "stay out of the critical path of getting features into production" to be very helpful. It's difficult to commit to coding timelines as a manager, and it harms your team if you are the bottleneck to shipping something.

But... keeping your hands in the mix elsewhere helps you stay informed and make better decisions. I found writing things like internal debugging tools, documentation, helping out on code review and architectural discussions, building example features against APIs etc were all good uses of my time.

An interesting trend I've observed over the past couple of years is that a lot of my friends who had moved into engineering management and stopped coding completely are picking up more coding tasks now thanks to LLMs - previously spending ~4 hours getting a development environment working and getting back up to speed wasn't justifiable, but LLM assistance means they can now get something small and useful done in just an hour which is much easier to carve out time for.

Generally, no. There’s a risk of unfair competition for work (they can delegate the stuff they don’t want because they have political power) and their code often becomes “untouchable” because few will call it out if the code is bad.

A hobby project to keep current isn’t a bad idea, though.

I've never seen a dual role manager that is actually good. It's usually a senior dev that gets stuck with management duties. They are usually good technically but then lack finesse and knowledge about most managerial issues (budget, employment law, team dynamics, etc). You're now at a disadvantage because the stuff they are supposed to protect you from or have power to help further your career is not developed. If your managers don't have enough management work, then flatten your org and expand their reach so they do.

> Do code reviews. Don't just skim PRs (sorry, reader!), but really dig into them: run the branch locally, test it, think critically about the design and the implementation, and provide feedback. Record a video of your review to highlight things that could be better.

Please no! Most managers want to increase output and engineers are aware of that. It is exceptionally frustrating when your manager tells you during your 1:1 that they want to help move things along and then does quite literally the opposite in a PR.

If you must dive deep into a PR, get the PR unblocked and then follow up with the change. Or stop telling your direct reports that you want to help unblock them.

From a practical point of view, I think anyone who has written any substantial amount of code is aware of the cost of context switching. Now picture a diary where you have 5 to 10 meetings a day with 30min free slots in between, and staying on top of emails on top of that. How much thoughtful coding do you expect to achieve?

I think managers gain at doing some coding to stay in touch, and certainly home projects are very useful. But it is impractical to expect anyone who runs a large team to be productive in term of writing code.

Reading these arguments are funny to me. By now there should be decades worth of research on this topic, all we seem to do is give anecdote after anecdote with no conclusion.

Either no is doing this research or no one trusts the conclusions.

Other topics like this are WFH or RTO, 4 day WW or 6 day WW. We as a profession seem to never come to a consensus.

Steve Jobs on this topic: https://www.youtube.com/watch?v=bVKcxK_tVBM

Summary: Managers who know how to manage, but don't know how to do anything are not the best managers. The best managers are the great individual contributors who never ever wanna be a manager but decide they have to be a manager because no one else is going to be able to do as good a job as them.

"But what does this mean for frontline engineering managers? Is the new normal just about writing more code and doing less of the other things that peacetime managers would normally do?"

"Should they be able to do code reviews? Yes."

There is no standard answer here. At many companies it is the Head Of Product who also oversees the tech team. They may not know how to write code, and will not know how to do code reviews. Some project managers lead engineering teams without knowing how to code. That's especially true outside of the tech industry.

For anyone interested in specific numbers, I interviewed Eric Garside about his experience scaling Freshly from 3 engineers to almost 80 engineers, see here:

https://respectfulleadership.substack.com/p/eric-garside-as-...

I also surveyed several of the CTOs who I know in New York City, about team size and scale and responsibilities, they gave their answers here:

https://respectfulleadership.substack.com/p/a-survey-of-ctos...

My view - as a manager - is that as I get older (simply put) I’m not as good any more as my younger team! While I help them, I also learn a lot from them as well. And I love that.

I have experience: project management, real life, edge cases, dealing with C-levels and customers, risk mitigation, being calm when sh*t hits the fan, mentoring, etc. These are things the team benefits from far more than my fingers typing away.

At the same time, if I’m actually doing my job well, I really don’t have time to code.

I also want my team (senior and junior) to have a feeling of ownership. The company has goals and directives, but what’s being built belongs the team, not me. All the successes are theirs, and the failures are mine alone.

When do my fingers hit the keyboard? When there’s a time crunch (if I’m not willing to work more how can I ask the team to?), when the team is having a difficult time with a particular problem, or maybe when there’s some downtime and I can fix a few outstanding bugs or work on something extremely isolated (I never want to get halfway through something and then have to hand it off to the team).

Short answer: no.

Background: I used to be a developer and have managed developers. Now I work in info sec and manage info sec.

Longer answer: this question has been discussed throughout my career.

(1) Does it help for a manager to understand what a team does and needs to do? Absolutely yes. Some managers can do this without domain experience but it's a lot harder for them and the team.

(2) Should a manager keep doing what their team does? Probably not. I can actually do most of what my team does faster and better, but they need to learn to do it. I don't scale. I can mentor them, and if there's a need for more resource I can get it for them.

Edit: I do actually still write code - but just in my own time because I enjoy it.

A lot of what's in the article resonates with me. Specifically, I've seen cases where managers unwittingly use coding (a problem they feel comfortable with) as a way to escape from facing more serious manager responsibilities (problems they don't feel comfortable with). When you've got 10 years of experience doing X, and you just started doing Y two years ago, it's natural to try to play to your strengths by doing X.

But as a manager, there's a whole category of things that only you can effectively do, because the social environment and power structures are set up that way. In that context, coding is a distraction for a manager. Writing code often takes a lot of mental energy and stays in your head even when you're not at the keyboard.

I don't want my manager getting nerd sniped when they should be coaching a struggling colleague, advocating to upper management, having a tough conversation with a toxic team member, or reigning in the PM.

This misses the forest for the trees.

I understand most people have negative experiences with managers that can't code, but I would argue that the experience with these managers would also be bad if they could code.

You don't need to be able to code as a manager but you need to be highly skilled.

The only problem that arises then is that for decisions based on technical knowledge you have to trust others.

The main problem I see is that a lot of managers are too mediocre for the job and compensate it by trying to sell themselves.

Yes.

Back at my first salary job, the CTO of the company would occasionally jump into the code base to write some code when everyone else was busy .

It was a very small startup, but seeing him do that motivated me to work harder. It was also just a very awesome thing for him to do, since he could have just said no I have to go pick up a pie this evening figure it out .

Below him, the best manager I've ever had, was regularly writing large amounts of the core code base. In my opinion this is really good for team morale. If you want to call yourself a startup, this is how things should be done .

When I think about it, I really would only like to work for startups with around 50 people or less, or mega corporations. I don't particularly like the quasi 1,000 person start up with 800 rules, and a lack of stable funding.

Maybe?

Managers should not be evaluated based on code output -- it's not their job. However, writing code here and there -- to evaluate new technologies, make a rough prototype, or demonstrate a technique to be adopted by individual contributors -- may aid them in their management responsibilities and should be embraced when it does.

I've seen what happens when a manager is also responsible for individual coding duties. He ended up with roughly twice the work, shifting between two mutually incompatible mental modalities all the time, cranky with his subordinates and making a lot of sad phone calls to his fiancée explaining that he'd be late home from work, again. Not a good fate for any worker, even if the pay and prestige are better.

Middle management for software engineering has gone through a paradigm shift. The trend for businesses is flatter structures, manager roles with more responsibilities and this reality largely includes being as deep in the code as direct reports.

Since 2023, most roles I reviewed or interviewed for were player/coach roles, often close to 50/50 split. In my last EM role, I was hired for exactly this.

I found very few EM roles that are primarily managing with only "being in the code"; I can count on my hands out of hundreds of Engineering Manager roles. Probably closer to 95% or more EM roles require hands-on technical work similar to Tech Lead roles.

Yes, managers not coding was caused by ZIRP. It's a dumb idea.

- You need to understand the work to evaluate your team

- You need to understand the work to prioritize and rank what's important

- You need to understand the work to know which roles to hire for

- Good developers typically don't want to have a non-technical engineering manager

- The budget for a non-contributing team member reduces your budget for engineers

- It creates a more hierarchical and less flat org chart, which creates communication scaling challenges

I would only consider non-technical managers for companies with large budgets and a non-technical product. Not only should they be technical, they should be excellent.

This is a good article, and I find myself agreeing with it almost entirely. My current manager is one of the most effective development managers I've ever had in my career, and I think a good part of that is because he is involved in the codebase, but not directly responsible for new features in whole.

I have had managers with no concept of what's going on the codebase, and those were dysfunctional development teams that produced poor results, despite great communication and productive meetings.

I have had managers that were effectively another engineer on the team, and those were dysfunctional development teams that produced decent results, but had poor interaction with stakeholders and no unified direction.

As in many things, the ideal seems to be a happy medium. Someone who can read the code, who can write the code, and is interested in both, but whose duties are not primarily to do so.

One important distinction to make is the difference between a manager who started out as an IC on the project (esp. the case when they actually built the thing from the ground up) versus a manager brought in from the outside. I think the former is more likely to be hands-on and generally speaking the expert in the room. But, sometimes you need the latter. It can be challenging (even counterproductive) for the latter person to try to be technically relevant, when they are often the least technically knowledgeable person in the room when they join.

I say this as someone who has been in both roles!

I once had a job where my job title was officially “Team Leader”, and my job spec was both to do development work and manage a team of 2-3 people doing the same work. (Except my team never existed, because they were all vacant positions and the salary budget I’d been given was too low to actually hire anyone, and then I resigned.) But, in that org, you could manage people without being classified as a “manager”. Whereas everywhere I’ve worked since has had this hard rule, if you have reports you must be a “manager” (or director or VP or whatever), no equivalent concept of “team leader” (although conversely you can be a “non-people manager” who has the word “manager” in your job title but no reports-common for product managers, project managers, etc). But I’ve observed first line managers often de facto act as “team leaders” anyway (actually doing hands-on technical work), but as you go up the hierarchy it gets less common-although exactly where it tapers out is determined more by the individual than the title or management level

As an aside, I like the implied job description in the article:

    - Hiring and retaining great people.

    - Owning the team's strategy and roadmap, and ensuring efficient execution.

    - Making decisions to ensure that the team is working on the right things and saying no to the things that don't matter.

    - Dealing with fires, escalations, and other crises that pop up all of the time.

    - Building a strong culture within the team so that people are engaged, challenged, and motivated.

    - Mentoring and coaching your reports so they get better and can have more work delegated to them, thus increasing output further.

    - Managing the team's stakeholders so they can offer their steer to the team early and often.

    - Actively performance managing the team so that superstars can continue to shine and underperformers can be coached or exited.

    - Building close working relationships with other teams so that smooth collaboration happens across the organization, leading to a better and more cohesive product.

I don’t think you can answer this for all managers of all teams all at once. Most of the time the answer is probably no but sometimes it’s not. Sometimes good managers want to code and they need to in order to stay happy. Sometimes your team is fast moving and small and your manager needs to code to keep up. Sometimes your manager is so busy managing they can’t. It just depends.

From the article:

> It depends on the manager, the team, and the organization. As a senior leader, I would rather my managers be in the code as per the above list, but not necessarily putting themselves in the critical path by writing code, given that they are likely to be interrupted more often, have more meetings, and be pulled in more directions than their reports.

I once had a lead programmer (who was a great one) who insisted that a manager should not have access to the source code of our project. I have often worked with engineers who were nervous about who could check in code so I kind of understood. And this really was a great lead, technically strong and a good leader so I said, fine no code access for me.

It was a terrible mistake for me. So much of my value as a manager was being understand what the engineers were doing. I am not a great programmer but I can do it, and I could usually understand what an engineer was trying to do from looking at their check ins rather than listening to them in stand ups. I was just not that useful. There were times when the team was doing things that I knew I had seen done before with much better before with different methods, but just talking about it was relatively ineffectual.

I'm a good coder, not a great one. I hire great coders to work for me so that if I had to write code I would be the worst and least productive person on the team.

Therefore it is to everyone's benefit that I don't touch code. I do sometimes put on an engineering hat and help with brainstorming how we will build new things, or asking questions about open issues that no one has asked, but that is the extent I will get involved in engineering. I try to only get involved when it's something where my decades of experience will add a lot of value.

And historically all the best managers I ever had were the same -- former engineers who could understand everything we were talking about, but did not get involved in coding or code reviews.

Your manager should never be a blocker to getting things done.

A manager should never write any code that is part of the critical path of the deliverable.

Every single time without fail that I have had a manager who still tries to be an active contributor, one of two things happen.

Either they never keep their commitments as a coder because they are spending too much time on their management duties including meetings, manhole up, and career development for their reports or they are horrible managers who don’t or can’t do what I need from them as a manager - get me the resources I need to do my job, play politics and manage up and especially fight for raises.

Curious why large open source project (like linux) have no manager for all the political stuff but works well but every private company have a huge management chain for less complex stuff.

Entry level managers (supervisors in working class tongue) code. They are often player coaches and own a series of tasks. They work more than the other guys and are compensated more and on the manager path. They are lieutenants or sergeants.

Directors manage multiple supervisors and own a complete function. They are like colonels. They don’t code as their job is aligning with senior management and across teams/divisions/organizations. It’s political.

Larger companies subdivide this more.

I come from the $MegaCorp world rather than startups. We shipped massively successful products reaching billions of people.

Our managers did not write code, nor would they have been better managers if they did, IMO. They represented the team to the wider company, they facilitated communication across teams, kept us informed of changing goals and priorities, etc. Basically helped everybody row in the same direction.

As a manager, I make a point to produce some code that makes it into production every quarter, but my reason for that is not because I can't help myself from writing code. Rather, I want to understand the pain of the sdlc. I want to understand firsthand how long it takes to get something deployed and how many steps it takes. Being able to empathize, rather than just sympathize, with my direct reports has made me a better manager.

undefined

If you can't code, you should not be managing programmers. How much code you write will obviously depend on circumstances.

One of the most regrettable things I did was stop writing code. There were the usual things; to busy to keep it up, not a good look in mgmt circles, keep coding hurts your upward mobility potential etc., all true, but I felt I was on a slippery slope towards effectively managing development.

Tech is in constant flux. While the basics rarely change, and yes, there is a lott of reinventing the wheel in programming approaches, thongs do progress and you feel your grasp weakening.

So now I code again. Not daily, but enough to complete some projects on my own. It feels great, and it definitly has a positive impact on my overall capabilities.

Two things:

- It's dangerous for the manager's career to lose the ability to code. After five years they will have regressed nearly to a Jr engineer's ability and would need significant effort to get back into interviewing shape. I've interviewed so many managers who didn't like management or aren't quite good enough to stay in the management track who can't even remember how an if statement works.

- Small companies and early startups are much more likely to need engineers over politicians, so the manager working on engineering is much more helpful in smaller teams.

In my experience in all pretty small companies, keeping hands on the keyboard was essential for me and the team. This was great because I got tired of being a manager and wanted to go back to pure engineer again.

I find a good manager treats me like a peer and not some superior. Part of that is being in the trenches and having experience with what I'm working on. Maybe they're not taking on entire features anymore or doing the same level of code review as others but I think for a manager who manages ICs - they should be much closer to the code. Maybe this would be the "team lead" in some companies or a supervisor type role but I think I prefer that kind of person to be my actual manager than some other nebulous person who doesn't even interact with me on a daily basis.

A big problem I had with some managers who didn't code was that they sometimes had wild expectations. They were also not good mentors. They couldn't really teach you much of anything because their job was so distinctly different from what you were doing.

Tldr: managers should be in the code, but should not code themselves.

The article suggests managers should focus several tasks that depend on the ability to code, but not produce code themselves. I think that's not sustainable. When a manager stops producing code, their skills in that area begin to deteriorate. I've seen it happen several times. When you stop coding, you'll eventually start missing crucial things in code review, make very poor estimations of labour involved, and assign the wrong team members to tasks. It won't happen immediately, but it will happen within 5 years.

I think a better way is to keep producing code, but at a lower rate. If the project you're working down doesn't permit low-commitment development, start developing tooling or pick up a side project. A coach doesn't need to be a top-player in the coding field, but they do need to remain fit.

undefined

I do managing, coding, design, governance and overall what i call "improve the company" within my areas of expertise and context switching and commitments are the most challenging things. Need to be very disciplined and know the other areas are currently very stable and under control to carve out time for the other things. It is not impossible but it have to be very focused and the problem domain need to be quite good understood before jumping in. Example, writing a internal tool that in worst case get delayed for later is easier to start working on then being part of customer facing product development as all of a sudden i would need to jump into some urgent management tasks or overall just let the governance and long term company quality go down.

It can help but it's not essential. Most coders make lousy managers and most mangers make lousy coders. Sometimes you find people that can do both.

The meta question here is do you need a person to manage you or can you perform at your peak without being micro managed. If so, you might be management material ;-).

I've worked as a employee, freelancer, consultant and lately as a CTO. As a consultant you tell managers how to manage. When freelancing you don't lift a finger unless you are told by your client, who is not your manager (important to be able to tell the difference). If you start consulting them effectively, make sure you increase your rates.

If you are an employee, you might have a career path that may involve you evolving into a manager. And as a CTO of a small company it's my job to make sure shit gets done. And sometimes that means getting hands on and leading by example. And sometimes that means delegating work and optimizing my time use. And sometimes I like having some fun. I hate delegating fun things.

A CTO that isn't hands-on is not a CTO and cannot provide long term meaningful technical leadership. Any technical skills they have don't have a long shelf life. Companies with no CTO and a hands off VP of engineering are probably not technology companies. This is a problem if they are trying to be one. Or used to be one.

I've consulted a few of those. I've also advised startup founders to not hire a CTO and instead find themselves a good VP of engineering. Because they weren't creating a tech company and their technical staff was OK but clearly not management material. In a startup, CTO is a founding role. A tech lead reports to the VP of engineering, a CTO reports to the CEO. Or sometimes is the CEO as well. Big difference. The VP of engineering is not a C level executive typically. They aren't co-founders. They don't hold a lot of equity. Good CTOs that can double being a decent VP of engineering are rare. Very different skill sets. The opposite is more common.

In short, it all depends.

Should managers exist?

We got along better before all these managers arrived, could ship huge projects with 100 people and one ceo. And the ceo wrote code and didn't do any of the nonsense most managers fill their days with.

Yes, because you need to have a feel for the quality and direction of the systems your organisation is building and maintaining, independently of the people you have dominance over. Because that hierarchical relation itself will make it hard for them to accurately communicate to you how it's going and the more time you make them spend explaining things to you instead of cooperating with each other, the worse for the team that actually gets stuff done.

You should also keep using programming and automation to perform managerial tasks and not degenerate into constantly droning away in some mind-numbing office software suite.

Should managers know about what they are managing in general?

I know most people will say yes but many will argue they don't need to actually do it to have that knowledge. But my experience is that with software and IT, things just move too fast for that. Managers need to continuously stay active at some level to keep up to date enough to make reasonable decisions from a management perspective. Of course it always depends what you mean by "manager", if you just want someone to approve leave, check on project status and organise birthdays, farewells and christmas celebrations .... well yeah you are wasting someone who can code on that.

It should be mandatory they code a bit. It should also be mandatory they don't code too much. This may not have been true in the past but times have changed in my 40 years as a programmer.

No thank you. All of the worst managers I've worked with know just enough to offer annoying suggestions but not enough to understand why it doesn't work (usually workplace politics). Sometimes I can't "just" do anything. The best ones I've worked with used to code 5+ years ago.

I don't want to spend my time managing my manager. He thinks I'm an idiot because I can't follow his simple instructions and I think he's naive and ready to fire me. Bad for all of us!

Don’t allow your skills to atrophy. If you get to a high position, it’s much harder to find a job at that level. If you can code, you can always find something.

If you can’t code, and you can’t get a manager role, you’re in trouble.

Of course they should. The volume does not matter as much as the depth. alternatively you could ask "Should managers understand the domain in which they work"?

There should be fewer layers of middle management and more individuals who genuinely understand why they are building the product and for whom. Success is ultimately defined by how effectively your product solves customer problems — everything else is secondary to that goal. I've had brilliant managers that never programmed in their entire lives and had horrible managers that had >10 year experience programming.

Isn’t one of the best gifts a manager can give to their team, the gift of clarity?

A manager able to give their team programmatic tests to pass brings a level of precision which makes it much easier for them to know when the work is done.

If managers communicate requirements in the form of acceptance tests more often, then the problem of projects running over time and over budget will occur significantly less often, because there will be clear finish lines for dev teams to run towards and cross, and reproducible pass/fail outcomes to inspire confidence.

A lot of this really depends on the scale and scope of the codebase(s) that the manager is accountable for. Staying on top of a service with a few very clear usecases is very different than staying on top of several services, jobs, and tools across multiple repositories.

The ideal is to try and structure code in such a way that it "screams" about what is going on, but realizing that ideal in some contexts is extremely difficult.

Big distinction we’re not talking about is: coding just to stay sharp, or coding to actually merge to production?

Strong no from me on (2). Production coding takes tiime and focus and would eat into manager’s management work.

On the other hand, managers at all levels should actively engage with the output of their org. For the direct manager of a development team, that means periodically running the build, maybe coding on weekends (which I do) to explore ideas and prototypes, etc.

This comes at a good time for me as I’m being threatened with a promotion from dev to management. But I have a question: I am being told that for SOX compliance, the manager approving a release can’t have any code in the release. Any devs or managers at publicly traded company that can tell me if this is bunk or not?

I haven’t seen anybody else mention this:

Yes, because in a downturn managers get cut first and you never know when you will need to be an IC again.

I'm manager and contribute every week for clean up, refactoring, add tests, extra e2e testing and overall fixing issues in existing development and production environment that are either too boring or affecting team operations. Sometimes contribute to small features but I make sure Im.not blocker or promise things to anyone (mainly work on experiments with no SLO)

20 years ago I got tired of programming day-to-day and became a manager who still dabbles in development. I don't contribute to production code, but I do architecture, prototyping/research, and will roll up my sleeves to help a stuck programmer do debugging.

A manager of developers should know how to code, but any formal expectation of them coding shouldn't exist. This holds for medium to large organizations. The managerial duties should be enough to require the vast majority of their attention. Expecting development on top of management is a way to add fake capacity to a team to help ensure burnout.

Practitioners need to be led by a practitioner. I worked for a while with a team leader whose only connection to the team was absorbing credit, giving confusing orders and acting like he saved the day. I dont want to work with aggregator managers ever again.

> If you mean being the primary implementer of features, then probably not.

It's my belief that any engineering manager worth their salt should push back on this and argue for a seat at this table, every single time. I don't want to work for someone who is disinterested in new functionality in codebases under their purview.

Keep in mind that many companies don't do non-coding engineering leaders anymore. So from a career perspective, if you take a role that's non-coding, you are making yourself unemployable for large parts of small - medium sized companies, as well as larger companies that don't do this.

I don't code much anymore, the vast majority is reviews the only time I really need to get on the tools is if it's a concept that I need to teach someone, or more likely, something has gone quite badly wrong and it's 2100 and I don't feel like waking anyone.

Absolutely not. Nothing like an out of touch manager slowing down their engineers with random PR comments and questions. They should not go deeper than the design document/architecture level. They are not adding any value trying to put their washed skills on display.

undefined

I found that coding was important, as it helped me to connect better with my employees, and their challenges.

I also found that I could never schedule myself into the critical path. Most of my coding was open-source stuff, on my own time.

I code a little but largely on the periphery of stuff so as to not be in the critical path. I try to fix small bugs snd keep up with the code base but with nine reports I've got much other work to do.

Does a floor boss still do work of the grinders and machinists, YES!

Intimate knowledge is what makes a boss better, not by itself alone.

Have to continually sharpen the scissors (mind) and not just the boss'.

Should managers still code... or... should managers know how to code? They are very different questions. The former answer is probably no. The latter is much more interesting, IMHO.

I've had managers who not only don't code, but never coded. I'd settle for one that had substantial engineering experience in the past.

If they're directly managing coders, then they absolutely must be able to build, tweak, and run their product by themselves, and articulate the reasoning behind its design decisions.

i lead dev teams in the consulting world, I don't code even though i really want to. Sometimes I get pulled in for time sensitive work or the hard stuff but that's about it. Every time I get bored as a manager and reserve some fun stuff for myself a crisis happens that needs my attention and then the code falls behind schedule. So I guess that's the end of that.

i like the analogy of a sports team coach for being a manager. You can motivate, cheer, train, and drive the team to victory but you can't step on the field.

I think reliable managers probably don't have enough time to write codes (unless they work 70~80 hours week), but at least they should be able to review other's code.

One important caveat: I recommend that managers take on coding tasks that are SMALL and UNIMPORTANT. Leave the key work for the front-lune developers (feel free to code review it).

"Administration" and "technical leadership" are two different skill and task sets, and it's a shame they're conflated into one position

I like two kinds of managers:

- Managers who came from the trench and can guide us through the mist;

- Managers who know nothing about the trench but leave us alone;

They should be able to automate their work with code: analytics reports, bots, etc. Committing code to product codebase - probably no.

It depends on the company and how much is reflect in the paycheck

Yes but only a little.

I think managers are often out of touch with what the team puts up with.

Does a football coach need to be in the lineup in order to be great?

Yes, if you can't measure it you can't manage it

At one point in my life, I became a manager that didn't code. Way too many meetings, travel, and non-technical things to do, so no time and most importantly, no headspace left for coding.

Over the course of about two years this drove me to mental issues. I was unhappy, I could not have a technical discussion with people working for me, I felt inadequate and dumb. On a practical level, I was expected to make technical decisions, but couldn't, because all I had was poorly communicated information from people working for me. That is not enough!

I realized two things:

1. It is not possible to do technical management without coding yourself. CTOs that don't code are not good CTOs.

2. I never want to be in a work situation where I am away from coding.

Since then I've been very happy, first co-founding a business where I had a CTO role (and coded), and then solo-founding and running a business where I code all the time.

Big no! from me.

Anyone in a people leadership role has to have a political mindset. You simply cannot be technically oriented and politically oriented at the same time. This is not a negative, the human brain, generally speaking, can't focus on two unrelated priorities at the same time.

I have had a highly negative experience over this. When resolving technical conflicts, you expect technical merit and reasoning to be used. But managers who also meddle in technical affairs use political leverage and tactics to see their desired technical outcome. This is a very unpleasant and toxic affair overall.

Imagine trying to solve a bug, but your manager wants a specific work around implemented which will result in bigger problems down the road. If you disagree with your manager, your performance review will suffer, you will be called disagreeable, hard to work with,etc.. You are only considering the best course of action from a technical perspective. Your technical peers can also review your code and reasoning and you can debate in a civilized manner over the technical merits of the issue. But as soon as a manger is involved, things will get toxic fast. It is nearly impossible to avoid micromanagement as well.

Especially if the manager really knows what they're talking about. Then they'll really be looking over everyone's shoulders and causing drama.

It is such a disorienting thing, having to fight political drama over simple and straightforward matters. This is how I learned what gaslighting is! I would be reaching out to people I consider a lot smarter than myself, asking their opinion on the subject (and they'd mostly agree with me), because I legitimately got disoriented and doubted everything I knew.

I really hope none of you experience this. At least when someone is being mean/toxic for other reasons you can explain it away, but when they use their technical expertise, that's a whole new level.

The whole concept of your management trusting you with the details and expecting you to show result goes out the window this way.

whats a manager? I have an AI for that!

it depends. Last ~3 years i was CTO of small company - with a software team of 7-10. With the idea of - no-more-coding (TM). Just the other 90% - from cleanup repos and jira etc project things - to organise proper reviews and workflows - process things, etc, to.. pushing infrastructure, hiring and pitching to investors/clients. But the codebase was.. not ideal. It was hard and tempting but i managed not to touch it directly, only did rise architectural-and-similar-level issues, and did (somewhat nitpicky) code reviews - like, date-arithmetic IS very important, pretty-please. Still, had to build a crawler over the production-codebase to visualise the event-flows and map consequences - a reverse-engineering live-Documentation, kind-a. Scratching my own itch? pfft. (hint: There was no other docs..)

But then there were some small non-essential projects that would have been complete distraction for the team.. so i took one, then another. In nodejs which isn't my cup of tea - so even better, learned some things the hard way. Though.. lucky they were rarely needing support, or it would have become a chore.

Then, one day, the company was acquired, and.. i am not a CTO anymore, but a (tech and also non-tech) Lead of some future greenfield project, ~re-factor with diff.goals. So here we go - research, architect, code, hire, discuss, demo, eh.. usual tiny-startup thing - except that the politiking and the now-everyday-a-new-policy chase me away.

So.. IMO one has to keep some of the programming skills sharp. In order to stay relevant. But this applies when you are 1 level away from actual coding. If farther.. probably no time, and no point. Although.. too long staying hands-off, and there may be no coming back - and losing grasp of how-it-feels-like to be in the trenches.

too much context switch (and no good answers here)

It's 2025 and the software industry is still debating this :-).

I was a manager for about 9 months before being laid off. The goal was to attain power but also work on the huge backlog of work we had. I wanted to use the power to block outside forces that would push my engineers in the wrong direction or remove their focus for investigating things like platform bugs.

What ended up happening was that the project work even though completed was used to gaslight me for 6 months as breaking the platform and causing bugs. Issues that infra team promised to “fix later”.

Go ahead code when you’re a manager it can be effective, but be careful when working with outside ineffective teams they will put blame on you if they can. Management is political and once you start making moves that outshine other teams, opposition will come out of the wood work to bring you back to their level of unhappiness.

The trick to identifying this is when people start naming you at meetings you aren’t in. It might be good things they are saying but that may shift as the good things well dries up.

No

Ex-Dropbox manager here. I also built BetterHelp (sorry) and I was the VP of Engineering Grooveshark.

Dropbox was the first job I've had where managers are not expected to code although they still go through a small ramp up where they can fix a tiny bug or make a hello world commit and deploy it to production to show that they understand how some of the systems work. Due to some crazy circumstances with the team I took over, I never even got to go through the ramp up. I was thrown straight into the deep end leading a team dealing with an urgent crisis that could end the business.

It was scary as hell, going from what in hindsight had been a "TLM with extra responsibilities" in my previous jobs to a full fledged EM role with all of the same accountability for quality and timely production, but none of the direct control. But I quickly realized that I was surrounded by people who were at least as capable as I was and usually more brilliant.

I think my greatest technical strength was always in eliminating technical complexity, making systems more robust and maintainable. It turns out you can still spot the blinking red lights of unnecessary complexity just by talking about the systems from a high level and asking the right questions, and when you help other talented engineers see those problems, they will naturally want to fix them. No need to jump in and do it yourself.

Once I knew I had a team I could trust and understood the strengths of the different players, I had to shift my focus to learning how to be a real manager. Managing people is a wildly different skillset than writing good code or building a good product, and I realized that I had never really been a manager despite leading teams of people. My apologies to everyone who ever worked for me before this point. Without realizing it, I had always treated the human factor as an annoyance and I probably hindered the growth of past teams a lot by stepping in and doing the high stakes, high urgency stuff myself.

When folks grew under me, especially at Grooveshark when I was young and immature, it was a happy accident and not something I was very intentional about. At Dropbox I really learned the importance of investing in people, giving them opportunities to grow and creating space to allow them to make mistakes. I didn't touch a line of Dropbox code or ever commit a thing, but my teams were high impact and many of the engineers who worked with me told me I was the best manager they've had.

Now I'm a co-founder at my own startup and, of course, I'm writing code again. Yeah, I'm a little rusty with some of the language specifics but I've been talking to brilliant engineers about their work for the last 7 years, when it comes to robust system design I'm probably a better engineer than I was the last time I wrote production code that was used by tens of millions of users. I will of course be in the hybrid role of building and managing folks for a while, but I hope I can keep my manager chops honed and support my team properly as I build and grow it and, eventually, stop writing production code again.

Yes

strong yes for me

flat orgs win.

Managers should never write or review code. This is a fundamental principle of human work, it has nothing to do with code itself. Honestly I don't know what insane HN meme started this trend and made it acceptable, but it really needs to end now.

The whole point of different roles doing different things is distribution of labor [1]. This idea is thousands of years old, this shouldn't come as a shock. As an extreme oversimplification, the people who are very good at a particular thing, should focus on that one thing. The more responsibilities or tasks someone has, the worse their output will be. Specialization enables higher quality, faster work, with less difficulty and waste.

A coder's job is to write code. A manager's job is to manage people. The author's post listed nine different important responsibilities for managers, that has nothing to do with code. But then just brushed it off, like it's easy! People, these aren't easy things to do, nor are they quick! Just doing general management work will easily suck up 40 hours a week on any team. If you've run out of management tasks, you probably aren't managing well.

Almost every manager I have had has not performed well. With the exception of one, they never trained as a manager, nor read or understood the basic yet critical functions and skills of a manager. Some of them used to be engineers and were just promoted up to a job they are incompetent at (the Peter Principle). And some moved into the position from some other job that wasn't management (or engineering management). On top of that, often there is a shortage of project managers, product owners, etc. These are critical roles to ensure high-quality, faster output for a team. In the absence of people filling those roles, and on teams that are not high-performing teams, it's up to the manager to fill those roles for their team.

So now the manager is not only responsible for the careers of their direct reports, they're responsible for keeping the team on track and producing high-quality, fast work. I've only ever seen one or two managers that could accomplish this feat - and that's before we add-on writing and reviewing code. How the hell are they supposed to get the time for all this, much less build up expertise in all these things, simultaneously?

I would argue that knowing how to code at all makes for a dangerous manager. I've had several managers turn into micro-managing freaks, telling me how to do my job, even preventing me from doing my job. They insisted they knew better, because they had written some code in the past, or adminned a server one time... yet I'm the one with the most experience and skills in that field. (Dunning-Kruger seems worse in those with authority, who don't wield it with humility)

On the other hand, the most effective manager I have ever had, had zero idea how to code. Because he was not technical, he focused on his actual job: mobilizing groups of people towards a task, measuring its progress, helping resolve human challenges, protecting his direct reports, and helping them progress in their careers. He never once told me how to do my job. He instead asked myself and my peers a series of simple questions in order to have us explain what we were doing, and through that process, we actually discovered several times that we should do it differently, or solved our problem.

So if the manager isn't writing or reviewing code - who will?!

An engineer's job is to build things. So someone who specializes solely in engineering should be reviewing code, designs, etc. There are many different roles for people who do this - software architect, systems architect, engineering team lead, staff engineer, principal engineer, etc.

A long, long time ago, the whole point of having "Senior" in your title was to convey the fact that you were an expert in your field. If there was a "Senior" engineer, that was the person you went to to tell you if you're doing it right. Hey, I just wrote this algorithm, does this look okay, Senior Engineer? I'm changing this field, can you see anything that might go wrong, Senior Engineer? Now of course it just means a college grad got a promotion after staying for 2 years.

Not having a real Senior Engineer somewhere in the company means you are going to end up making some real turds. Having a manager take the place of a Senior Engineer doesn't help, because they can't do two completely different jobs well. You can't be both a great plumber and a great electrician. You can half-ass both, though, and get half-assed results.

If you're a manager and you want to have a high-performing team, stop writing/reviewing code, and instead do everything you can to implement the suggestions here [2]. There is an enormous amount of work needed to achieve these suggestions, so you will not have any time to look at code, I guarantee you. But your team will end up working much better, producing better outcomes.

[1] https://en.wikipedia.org/wiki/Division_of_labour [2] https://dora.dev/research/?view=detail

Of course they should still code, otherwise they're not engineering managers, they're project/product managers, and good engineers will not follow those people.

Whenever we talk about EM duties, it's always a list of fuzzy, empty words:

> Owning the team's strategy and roadmap, and ensuring efficient execution.

This is project management.

> Making decisions to ensure that the team is working on the right things and saying no to the things that don't matter.

This is product management.

> Dealing with fires, escalations, and other crises that pop up all of the time.

How can an EM that doesn't code deal with fires? The only thing they can do is pull the sleeve of someone else who does code, and then, what, hang behind their shoulder until the fire is put out?

> Building a strong culture within the team so that people are engaged, challenged, and motivated.

This is meaningless. We're not children.

> Mentoring and coaching your reports so they get better and can have more work delegated to them, thus increasing output further.

Mentoring them at what? If an EM doesn't code, how can they mentor in an area that's relevant to the mentee (i.e. a coding engineer)? They can coach them on moving Jira tickets more effectively, or playing office politics better?

> Managing the team's stakeholders so they can offer their steer to the team early and often.

Again, product management.

> Actively performance managing the team so that superstars can continue to shine and underperformers can be coached or exited.

Combination of meaningless and project management. In order to be able to evaluate someone's performance, you need to know their work. If you don't, the only metric you have is number of tickets, or "velocity" or whichever other bullshit metric you use because you're not in the trenches.

> Building close working relationships with other teams so that smooth collaboration happens across the organization, leading to a better and more cohesive product.

Office politics.

The only one that make sense is hiring and retaining great people, but you can't do either without being a technical person.

EMs that don't code making technical decisions is a showcase for divorcing decision making from suffering the consequences of those decisions. And having teams with EMs + tech leads + team leads etc. is just making things worse by diluting the responsibility.

https://world.hey.com/dhh/we-once-more-have-no-full-time-man...

[dead]

[dead]