> Somewhere around 100 people, the ground shifts underneath you, as you realize you don’t know everyone anymore. You just can't. There aren't enough hours in the day, and honestly, there aren't enough slots in your brain
This is definitely a rare exception, but at my first job, my boss's boss's boss (basically the #2 in the engineering organization, although probably only like 2/3 of it was under his purview) somehow seemed to have an unfathomable ability for knowing not only everyone under him the org tree, but details about what they were working on. I think it must have been at least 150, maybe even 200 people, and as far as I could tell he could recall every single person's name, project, and the general status of their work without needing anyone to remind him before talking with them. Maybe he did just really studiously review notes or something before any meeting or even chance of an ad hoc conversation in the halls, but I never really saw him typing at meetings or writing stuff down to keep track of later, so at the very least he'd need to have been able to retain a lot of information long enough to accurately record it later. Witnessing this firsthand for a few years was easily one of the most impressive mental feats I've ever observed.
I recall in the first lecture of some Comp Sci class back at uni, our lecturer had learnt what felt like every student's name and face from their digital profile, some 100 people. Whenever a random student raised their hand to participate he would say say "yes, <first name>". I'm still to this day in awe of that.
Did having such a person in charge make a qualitative difference in the atmosphere of how work proceeded among people there?
If so, do you think it would have played out similarly if the organization had had an equally effective "glue person" who wasn't in charge (therefore didn't have any authority to delegate or divide most tasks) and was required to manage upward [sic] to coordinate things for people?
> I have an idea for a quirky event experimenting with the "minimum viable feeling of community", but need to explain some context first. Bear with me...
> [...]
> So here's the event idea: what if someone ran an event where the 2nd rule was "NO INTRODUCTIONS", but only because the 1st rule was "you must arrive having fully memorized ONLY everyone's name and face". Beyond the strange entry requirement, what would such an event feel like?
> And what strange sorts of intimacy might be created by this minimal scaffold of "knowing everyone"... & being in community together? I suspect it might feel like a warm event full of friends, but where everyone had mysteriously forgotten everything they knew about one another :)
My last CTO was hired after me, the org was hyperscaling.
When i was interviewed I was told that the company is banking on JS and that's what we were doing on both ends [1]
When CTO was hired he made a walk through the office, greeting every team, he stopped at our cubicle and asked what we were doing - I told him basics - and he said "you should be doing that in Java".
Few weeks later he had a townhall presentation. He came to a room full of people, plug in his computer and the screen started playing a pornhub flix.
Is this that impressive? In my high-school there were roughly 200 people in the year group and I'm pretty sure most of them could list off practically everyones names and something about them?
The biggest (engineering) team I've built and led was 150 people, so I cannot speak for 200. I don't feel that the solution of keeping the understanding of what is happening was about a formal structure like an "employee steering committee".
Rather than trusting the same principles used for scaling the doing side of the business. Things like empowering people to make decisions [1] or being clear about what/how/why you make certain types of decisions [2]. Working on staying aligned with your closest team, which then spent energy on staying aligned with their teams, etc. Sample randomly from the whole org, but mostly at your pace.
The biggest mistake I've made was that I've pulled myself (for legitimate reasons, it seemed) from having a true conversation with every single hire before they've gotten an offer, when we were around 70. 30 minutes is typically enough, but I feel you need a singular person as a gatekeeper for the final values-fit check. Partially thinking that 70 is already good enough, but later I've come across people talking about, like, 500 people before pulling out [3] :)
I dont think the solution to not knowing people in your company is to create bureaucracy. Ie - only hanging with 10 executives and a focus group. Get out there and talk to people for a few minutes - at the office or wherever.
Ultimately, it is. The post didn't touch on this, but it's exactly why the world looks like it does - it is, and has always been, recursively subdivided. It's why we have districts and towns and counties and states and countries. Hierarchical governance is a result of trying to cooperate in groups larger than the limit of how many direct relationships our brains can support.
Maybe, but a 200 person company isn’t really that big. The CEO should probably get over themselves if they think they couldnt possibly know everyone at least a little bit.
I think, putting what you're saying another way, just because your capacity might be limited to hearing from N people, that doesn't mean it has to be the same N people all the time. It should include a sampling across everyone so you have a lower chance of systematically missing entire points of view.
Teacher here. Best Principal I had would gatecrash your class once a year, then have a chat giving feedback. Kind of stressful (it could happen with little warning) but whatever.
They knew everyone in the school (ebery teacher and about 500+ student names), and what happened in every class. It took time and talent to do it, but it made them a lot less insulated.
Claiming you can't know 100-200 people - your high school teacher wrote 100 reports. Now obviously they aren't 100% on the ball, but they have some idea (I hope).
There's an old story about how Bill Gates once took a call in tech support. A far larger organisation, and he still was willing to dive deep and see what was going on at the least glamorous part of the coalface.
There's a difference between trying to micromanage everything, and micromanaging enough that you're not out of touch.
Feedback is a two way street. It both let's you know what is happening, and let's the people below know that you actually care. Even if you can't (and arguably shouldn't) be everywhere at once, it has its place.
Now yes, it's drive by management and isn't the main tool that a manager should use, but being overly scared that your trusted expert juniors will be destroyed by a senior checking up on them is maybe a bit silly, and if a senior manager is such a tool that they do cause havoc just by looking over someone's shoulder and giving them a bit of feedback you're already in trouble.
Inulation isn't the answer IMO, just accepting that yes you don't need to know everyone and everything to the same level as if it was a small team.
Our CEO does this. She talks to a lot of people. Once you start talking business, she clearly doesn't care about your opinion, unless you're praising something. If it's remotely critical or a suggestion to change something, you can see in her eyes she's not even processing the words anymore.
I rather have her not talking to me, because it's much worse knowing she fakes her openness, than actually just not showing up.
So I get confused when I read things like "feedback doesn't scale".
Because what am I, if not a self-organizing collective of a trillion cells.
That seems like feedback which scales right there.
This seems like a concrete example of why this logic is flawed.
To me I believe it more useful to start with the premise of:
I'm already communicating and leading trillions, how do I actually do that?
A common issue is that we hold thoughts, logic and language as a type of universal gold standard, while ignoring that most of our communication isn't even verbal to begin with.
It's context, observation, pattern recognition, a self-serving goal which aligns with the collective, because we're all wanting the same things.
What feels good, what's expansive, what's beautiful etc. These are the reward functions for healthy communication in the human body, the more that we align and work with these, the better the results.
> Because what am I, if not a self-organizing collective of a trillion cells. That seems like feedback which scales right there.
You're not, not in this sense. There is no body-wide feedback at all at the cellular level, any single cell is disposable and nothing will notice if it dies. Any meaningful feedback exists between and within functional units of the body.
There is, however, the other, original form of feedback that allows the body to exist - the one that allows you to not have this commonly understood "feedback" in the first place. That is, feedback loops, the control theory concept of systems that self-stabilize or self-amplify. This, not some top-level control, is what's keeping the body together.
The body is a perfect example of a naturally hierarchical system. Society is another. That's what scales.
> So I get confused when I read things like "feedback doesn't scale". Because what am I, if not a self-organizing collective of a trillion cells.
Well, yeah, that feedback scales perfectly because your cells don't have free will.
I think there are plenty of real-world examples of large-scale projects where feedback scaled well, for similar reasons... though I doubt we want to use those as a guide.
Most people who live in a city want the city to function well, and actively do their tiny little bit to see this happen. This doesn't stop them from flipping each other off on the freeway.
More broadly I think you're missing the point of the article. A single person can command a military of millions, but that single person can't ensure that everyone in that military have all of their needs met, personal emergencies dealt with, or just plain care enough to not half-ass it. Much less hear and respond to everyone's ideas on what would make things better, or what's making things worse.
Our individual cells have very simple needs in order to keep our larger structure functioning, and even then sometimes things go catastrophically wrong.
I just wanted to stop by and say thanks for the discussion. There are a lot of good thoughts here, and I've enjoyed reading them. And I'm glad y'all liked the article. Well, at least those of you who enjoyed it.
Though I agree with the larger point, there is a critical way to overcome that. The second line of leadership must own the culture at their team's level. This only works if you have direct access to the larger group. An open-door policy where anyone can schedule time with you is essential.
You might not understand their struggle, but you can hear and route it to the right people. Sometimes the best way to show empathy is simply to listen.
> An open-door policy where anyone can schedule time with you is essential.
In my experience, open door policies are necessary but not sufficient. If the policy is to wait for feedback to walk through your door you will only hear from the set of people motivated, willing, and trusting enough to do that.
You have to go out and ask everyone one by one the appropriate questions and also be willing to listen. I’ve been in some companies where feedback was requested but then the immediate reaction was to argue and deny any feedback given, which is a fast path to ensure people stop providing feedback.
Agreed. Setting up structures for people to provide regular feedback is a must -- I believe this goes without saying. The only downside is that this cannot be done 1:1 with each person. Hence, enabling even the motivated few is a good trade-off. But sure, necessary but not sufficient.
At the same time, accepting and acting on feedback is a skill in its own right.
> You can't know 200 people, but you can know 10 people who each know 10 people
You are still 100 people short to know 200 people, but I got the idea.
The 100 people limit is already know by most of teachers. Having more than 3 classes, it is mostly impossible (very hard) to have a "deep" follow up of each student.
Having more than 6 classes and it is strictly impossible to follow them even in the best conditions.
When I worked at Netflix many years ago, they loved to boast about how they didn't have any "processes". My experience was that process ALWAYS exists, but at Netflix you just had to figure it out and hopefully not step on the wrong toes along the way.
Sounds like an opportunity to give yourself a promotion. Just start sitting with the CEO at lunch, then walk around telling people what to do. (Joke btw)
* that the interaction with a peer _is_ the problem. I know we should all be grown up and able to talk about these things in a mature and effective way, but I can't cope with conflict in any shape or form, so if someone says Boo to me I cave in which doesn't get me any further
* because peers aren't the people that need to hear some of the things I've got to say, it's layers above me that need to hear it
Yup - they touch on proxy relationships where you have a few trusted reporters to break the crowd into cohorts that you can mentally simplify but whenever you do this you need to accept that it won't be complete. You should expect and make room for occasional noise from the fifty people behind your one trusted reporter because the problem could always lie with the reporter themselves.
This piece feels like an AI-generated effort because it’s not so much an exploration of leadership challenges, but rather a series of surface-level observations that lack depth. It's not just a general overview, but a collection of familiar tropes without any original or nuanced analysis, and the sentences aren't just simple, but lack the complexity and emotional depth that would make the piece feel truly human.
> set the expectation that they have strong relationships with their own teams
Good luck with that.
In most cronytocracies (typical, at the top levels of most companies), you get who you get. They may be really good engineers and "first line" managers, but suck at anything else.
A big problem is that companies don't have career tracks that match people's skills. The Peter Principle[0] applies.
Bad managers hire and promote other bad managers. Highly skilled engineers can often be terrible managers, but want to be managers, because that is the position they equate with "success," at an organization.
A Principal Engineer should be just as valued and well-treated as a CTO. Most companies fail to do this, so everyone wants to be the CTO. Establish a career track, where technical people aspire to technical positions.
And hire good managers; not ones that don't make the CEO uncomfortable.
> A Principal Engineer should be just as valued and well-treated as a CTO. Most companies fail to do this, so everyone wants to be the CTO. Establish a career track, where technical people aspire to technical positions.
I spent my whole career avoiding engineering management and trying to grow in the pure technical leadership direction. One day I realized that for every staff engineer there are 10 managers, for every principal there are 5 senior managers, etc.
Turns out management is not so bad and companies seem to appreciate that kind of help a lot more
edit: also as a manager you get to work on all those pesky “It’s a people problem, actually” parts of engineering which is pretty fun. Every time in technical leadership where it felt like “Well we’ve got the plan now we just gotta incentivize doing the plan” you’re the one doing the incentivizing yay!
I hated it, but was actually pretty good at it (I worked for a company that didn't suffer slackers, and they kept me for almost 27 years). I mainly kept it, because I couldn't trust anyone else to do the job correctly.
But my heart has always been in the tech, and I did side projects, that whole time. Since leaving, I ran screaming back to being a technical implementation person, and am almost deliriously happy.
A good manager is actually fairly hard to find. It's been my experience that a majority of highly-talented developers, don't make good managers.
> A Principal Engineer should be just as valued and well-treated as a CTO. Most companies fail to do this, so everyone wants to be the CTO. Establish a career track, where technical people aspire to technical positions.
100%. I’ve had to fight to build a real IC track at my last three companies - I don’t care if HR wants to call everyone a manager, or differentiate with “staff”/“principal” or whatever else, but there has to be a viable promotion track for everyone, and it has to be equitable.
Just experienced something similar working-at-definitely-not-capital-one as a principal engineer. My manager was horrible, and replaced by another bad manager. He incentivized bad behavior on my team and promoted inexperienced engineers, and group think, pushed me out for questioning the status quo.
Feedback does scale if you’re willing to adjust the loss function.
I like the Jason Fried-ism of: If something really matters, you’ll hear it again. If you have to write it down to remember it, it’s probably not important.
> but you can know 10 people who each know 10 people
Theoretically, yes, but in practice those 10 don’t really know 10 people. And if you don’t hire well and don’t have everything you need to keep them motivated, some of your 10 won’t even care about their 10 people or may actively be sabotaging you, peers, and/or their subordinates.
I like the building relationships concept. Ideally, each person in your 20 person group is in a different one of the 20 total teams.
The organization will never be perfect but I think it could work.
> Feedback doesn't scale because relationships don’t scale.
I would not say it this way; it is too simplistic. In fact, I generally caution against the dominant metaphor here of comparing feedback to scaling. It falls apart quickly.
Here’s a counter point. In many scenarios and settings, relationships provide transitive benefits. For example, if a leader builds trusted relationships with other leaders, a significant amount of trust can flow through that relationship.
To build a better understanding, I suggest building diverse models. Try to answer the question: What kind of qualities do relationships confer and why?
There’s also a generational aspect here. I started my career in the 2000 tech boom and bust. I’ve seen a lot of up-and-down cycles in the industry. I’ve seen lots of management styles and organizational cultures. People that had formative years during peak social media and/or COVID often have a different kind of socialization and this affects their default expectations. I won’t attach normative judgments without research, but there are significant differences.
When I think of the most impressive collaborations I’ve participated in with amazing results, relatively few of them involve tech organizations.
Building a scalable culture over various company sizes feels hard in the sense that generalizing prescriptive advice is tricky. A two person start up is cake because you only have to manage one internal relationship (a pair). People know great culture when they see it, but that is nothing like growing it.
feedback fails when it is scaled down as well, the very worst manager to employie ratio bieng 1:1
where for every single consideration an employee turns to there manager, and effectivly reverses roles
vs when there are several co workers who can use each other for feedback and in many ways then become self managing
which of course lands us in the impossible to define world of "team building"
it's mojo, nothing more
> Past a certain size, you have to make peace with the fact that a lot of people in your org are going to be frustrated with you, and you're going to have no idea why, and you may not going to be able to fix it.
I agree this is a necessary cost of being at the top of a large organization.
But I am very suspicious of the kind of people that seem to have no trouble at all being misunderstood and disliked by many. Yes, decent people can be in those positions. But it's often a honeypot for sociopaths.
> Without an existing relationship, it feels like an attack, and your natural human response is to dismiss or deflect the attack. Or worse, to get defensive. Attacks trigger our most primal instincts: fight or flight.
It is really important to recognize that it is the perception of an attack that triggers certain responses. For a counter example,
watch how puppies play. It can very rough at some level but at another the intent is clearly benign.
There are ways to shape and modify perceptions! Culture. Norms. Timing. Technology. Inclusion and exclusion criteria. Information architecture.
Never assume that the technology or protocols you use have been designed for your core values. Often you have to redesign it for your purposes. Please do.
Feedback *can* scale if one carefully defines protocols to suit particular goals. We are not helpless even if it seems we are hapless. Leaders and designers (often social scientists) must step up and show better ways.
Computer scientists and software engineers must show curiosity and intellectual humility here. Better to draw broadly from other fields: social work, negotiation, psychology, anthropology, public policy, and more.
> Somewhere around 100 people, the ground shifts underneath you, as you realize you don’t know everyone anymore. You just can't. There aren't enough hours in the day, and honestly, there aren't enough slots in your brain
This is definitely a rare exception, but at my first job, my boss's boss's boss (basically the #2 in the engineering organization, although probably only like 2/3 of it was under his purview) somehow seemed to have an unfathomable ability for knowing not only everyone under him the org tree, but details about what they were working on. I think it must have been at least 150, maybe even 200 people, and as far as I could tell he could recall every single person's name, project, and the general status of their work without needing anyone to remind him before talking with them. Maybe he did just really studiously review notes or something before any meeting or even chance of an ad hoc conversation in the halls, but I never really saw him typing at meetings or writing stuff down to keep track of later, so at the very least he'd need to have been able to retain a lot of information long enough to accurately record it later. Witnessing this firsthand for a few years was easily one of the most impressive mental feats I've ever observed.
I recall in the first lecture of some Comp Sci class back at uni, our lecturer had learnt what felt like every student's name and face from their digital profile, some 100 people. Whenever a random student raised their hand to participate he would say say "yes, <first name>". I'm still to this day in awe of that.
I had a math prof who did that in a class of 200 and still remembered our names 2 years later. Incredible.
Did having such a person in charge make a qualitative difference in the atmosphere of how work proceeded among people there?
If so, do you think it would have played out similarly if the organization had had an equally effective "glue person" who wasn't in charge (therefore didn't have any authority to delegate or divide most tasks) and was required to manage upward [sic] to coordinate things for people?
LOVE this question! Thanks for asking.
I also like toying with variants of where "essential elements" can live, sometimes in odd places :)
https://x.com/patcon_/status/1963648801962369358
> I have an idea for a quirky event experimenting with the "minimum viable feeling of community", but need to explain some context first. Bear with me...
> [...]
> So here's the event idea: what if someone ran an event where the 2nd rule was "NO INTRODUCTIONS", but only because the 1st rule was "you must arrive having fully memorized ONLY everyone's name and face". Beyond the strange entry requirement, what would such an event feel like?
> And what strange sorts of intimacy might be created by this minimal scaffold of "knowing everyone"... & being in community together? I suspect it might feel like a warm event full of friends, but where everyone had mysteriously forgotten everything they knew about one another :)
counterpoint
My last CTO was hired after me, the org was hyperscaling. When i was interviewed I was told that the company is banking on JS and that's what we were doing on both ends [1]
When CTO was hired he made a walk through the office, greeting every team, he stopped at our cubicle and asked what we were doing - I told him basics - and he said "you should be doing that in Java".
Few weeks later he had a townhall presentation. He came to a room full of people, plug in his computer and the screen started playing a pornhub flix.
He didn't got fired. I was.
I’m not sure it’s really a counterpoint, but still an interesting story (interesting, because it’s not really funny, is it?)
Why did you get fired?
2 replies →
Was he a good leader? And is that capability related to him being a good leader, would you say?
Is this that impressive? In my high-school there were roughly 200 people in the year group and I'm pretty sure most of them could list off practically everyones names and something about them?
On day 1 of your first year?
The biggest (engineering) team I've built and led was 150 people, so I cannot speak for 200. I don't feel that the solution of keeping the understanding of what is happening was about a formal structure like an "employee steering committee".
Rather than trusting the same principles used for scaling the doing side of the business. Things like empowering people to make decisions [1] or being clear about what/how/why you make certain types of decisions [2]. Working on staying aligned with your closest team, which then spent energy on staying aligned with their teams, etc. Sample randomly from the whole org, but mostly at your pace.
The biggest mistake I've made was that I've pulled myself (for legitimate reasons, it seemed) from having a true conversation with every single hire before they've gotten an offer, when we were around 70. 30 minutes is typically enough, but I feel you need a singular person as a gatekeeper for the final values-fit check. Partially thinking that 70 is already good enough, but later I've come across people talking about, like, 500 people before pulling out [3] :)
[1]: https://www.bobek.cz/decisions/ [2]: https://www.bobek.cz/work-principles/ [3]: https://mastersofscale.com/reed-hastings-culture-shock/
I dont think the solution to not knowing people in your company is to create bureaucracy. Ie - only hanging with 10 executives and a focus group. Get out there and talk to people for a few minutes - at the office or wherever.
Ultimately, it is. The post didn't touch on this, but it's exactly why the world looks like it does - it is, and has always been, recursively subdivided. It's why we have districts and towns and counties and states and countries. Hierarchical governance is a result of trying to cooperate in groups larger than the limit of how many direct relationships our brains can support.
Maybe, but a 200 person company isn’t really that big. The CEO should probably get over themselves if they think they couldnt possibly know everyone at least a little bit.
I think, putting what you're saying another way, just because your capacity might be limited to hearing from N people, that doesn't mean it has to be the same N people all the time. It should include a sampling across everyone so you have a lower chance of systematically missing entire points of view.
Teacher here. Best Principal I had would gatecrash your class once a year, then have a chat giving feedback. Kind of stressful (it could happen with little warning) but whatever.
They knew everyone in the school (ebery teacher and about 500+ student names), and what happened in every class. It took time and talent to do it, but it made them a lot less insulated.
Claiming you can't know 100-200 people - your high school teacher wrote 100 reports. Now obviously they aren't 100% on the ball, but they have some idea (I hope).
There's an old story about how Bill Gates once took a call in tech support. A far larger organisation, and he still was willing to dive deep and see what was going on at the least glamorous part of the coalface.
There's a difference between trying to micromanage everything, and micromanaging enough that you're not out of touch.
Feedback is a two way street. It both let's you know what is happening, and let's the people below know that you actually care. Even if you can't (and arguably shouldn't) be everywhere at once, it has its place.
Now yes, it's drive by management and isn't the main tool that a manager should use, but being overly scared that your trusted expert juniors will be destroyed by a senior checking up on them is maybe a bit silly, and if a senior manager is such a tool that they do cause havoc just by looking over someone's shoulder and giving them a bit of feedback you're already in trouble.
Inulation isn't the answer IMO, just accepting that yes you don't need to know everyone and everything to the same level as if it was a small team.
3 replies →
Our CEO does this. She talks to a lot of people. Once you start talking business, she clearly doesn't care about your opinion, unless you're praising something. If it's remotely critical or a suggestion to change something, you can see in her eyes she's not even processing the words anymore.
I rather have her not talking to me, because it's much worse knowing she fakes her openness, than actually just not showing up.
So I get confused when I read things like "feedback doesn't scale". Because what am I, if not a self-organizing collective of a trillion cells. That seems like feedback which scales right there.
This seems like a concrete example of why this logic is flawed.
To me I believe it more useful to start with the premise of: I'm already communicating and leading trillions, how do I actually do that?
A common issue is that we hold thoughts, logic and language as a type of universal gold standard, while ignoring that most of our communication isn't even verbal to begin with. It's context, observation, pattern recognition, a self-serving goal which aligns with the collective, because we're all wanting the same things. What feels good, what's expansive, what's beautiful etc. These are the reward functions for healthy communication in the human body, the more that we align and work with these, the better the results.
> Because what am I, if not a self-organizing collective of a trillion cells. That seems like feedback which scales right there.
You're not, not in this sense. There is no body-wide feedback at all at the cellular level, any single cell is disposable and nothing will notice if it dies. Any meaningful feedback exists between and within functional units of the body.
There is, however, the other, original form of feedback that allows the body to exist - the one that allows you to not have this commonly understood "feedback" in the first place. That is, feedback loops, the control theory concept of systems that self-stabilize or self-amplify. This, not some top-level control, is what's keeping the body together.
The body is a perfect example of a naturally hierarchical system. Society is another. That's what scales.
> So I get confused when I read things like "feedback doesn't scale". Because what am I, if not a self-organizing collective of a trillion cells.
Well, yeah, that feedback scales perfectly because your cells don't have free will.
I think there are plenty of real-world examples of large-scale projects where feedback scaled well, for similar reasons... though I doubt we want to use those as a guide.
"because your cells don't have free will"
They are still independent cells. If they stop cooperating with the rest of the body, they become literally cancer.
5 replies →
Might be the biggest example of false equivalence I have seen in a long time.
Cancer.
Most people who live in a city want the city to function well, and actively do their tiny little bit to see this happen. This doesn't stop them from flipping each other off on the freeway.
More broadly I think you're missing the point of the article. A single person can command a military of millions, but that single person can't ensure that everyone in that military have all of their needs met, personal emergencies dealt with, or just plain care enough to not half-ass it. Much less hear and respond to everyone's ideas on what would make things better, or what's making things worse.
Our individual cells have very simple needs in order to keep our larger structure functioning, and even then sometimes things go catastrophically wrong.
There's a lot of research on this, particularly from Robin Dunbar, who gave us "Dunbar's Number" https://en.wikipedia.org/wiki/Dunbar%27s_number
If you are operating at that scale, it's not a team anymore, it's an online community and should be treated as such.
I just wanted to stop by and say thanks for the discussion. There are a lot of good thoughts here, and I've enjoyed reading them. And I'm glad y'all liked the article. Well, at least those of you who enjoyed it.
> Their struggles are not your struggles anymore.
Though I agree with the larger point, there is a critical way to overcome that. The second line of leadership must own the culture at their team's level. This only works if you have direct access to the larger group. An open-door policy where anyone can schedule time with you is essential.
You might not understand their struggle, but you can hear and route it to the right people. Sometimes the best way to show empathy is simply to listen.
> An open-door policy where anyone can schedule time with you is essential.
In my experience, open door policies are necessary but not sufficient. If the policy is to wait for feedback to walk through your door you will only hear from the set of people motivated, willing, and trusting enough to do that.
You have to go out and ask everyone one by one the appropriate questions and also be willing to listen. I’ve been in some companies where feedback was requested but then the immediate reaction was to argue and deny any feedback given, which is a fast path to ensure people stop providing feedback.
Agreed. Setting up structures for people to provide regular feedback is a must -- I believe this goes without saying. The only downside is that this cannot be done 1:1 with each person. Hence, enabling even the motivated few is a good trade-off. But sure, necessary but not sufficient.
At the same time, accepting and acting on feedback is a skill in its own right.
1 reply →
> You can't know 200 people, but you can know 10 people who each know 10 people
You are still 100 people short to know 200 people, but I got the idea.
The 100 people limit is already know by most of teachers. Having more than 3 classes, it is mostly impossible (very hard) to have a "deep" follow up of each student. Having more than 6 classes and it is strictly impossible to follow them even in the best conditions.
First paragraph already assumes a lot. We're a team of 5, but no, I cannot tell anyone my concerns or problems.
I've read somewhere that company politics is necessary. Whether that's true, I'll probably never know.
When I worked at Netflix many years ago, they loved to boast about how they didn't have any "processes". My experience was that process ALWAYS exists, but at Netflix you just had to figure it out and hopefully not step on the wrong toes along the way.
Ah, the tyranny of structurelessness.
1 reply →
Or those "we have a flat hierarchy" companies
Which is really "We don't document our implicit hierarchy, screw with it at your own risk
Sounds like an opportunity to give yourself a promotion. Just start sitting with the CEO at lunch, then walk around telling people what to do. (Joke btw)
1 reply →
What do you feel you cannot share concerns with your peers?
Not the original poster but:
7 replies →
This approach can't inform you that someone in the feedback chain is causing a problem.
Yup - they touch on proxy relationships where you have a few trusted reporters to break the crowd into cohorts that you can mentally simplify but whenever you do this you need to accept that it won't be complete. You should expect and make room for occasional noise from the fifty people behind your one trusted reporter because the problem could always lie with the reporter themselves.
This piece feels like an AI-generated effort because it’s not so much an exploration of leadership challenges, but rather a series of surface-level observations that lack depth. It's not just a general overview, but a collection of familiar tropes without any original or nuanced analysis, and the sentences aren't just simple, but lack the complexity and emotional depth that would make the piece feel truly human.
> set the expectation that they have strong relationships with their own teams
Good luck with that.
In most cronytocracies (typical, at the top levels of most companies), you get who you get. They may be really good engineers and "first line" managers, but suck at anything else.
A big problem is that companies don't have career tracks that match people's skills. The Peter Principle[0] applies.
Bad managers hire and promote other bad managers. Highly skilled engineers can often be terrible managers, but want to be managers, because that is the position they equate with "success," at an organization.
A Principal Engineer should be just as valued and well-treated as a CTO. Most companies fail to do this, so everyone wants to be the CTO. Establish a career track, where technical people aspire to technical positions.
And hire good managers; not ones that don't make the CEO uncomfortable.
[0] https://en.wikipedia.org/wiki/Peter_principle
> A Principal Engineer should be just as valued and well-treated as a CTO. Most companies fail to do this, so everyone wants to be the CTO. Establish a career track, where technical people aspire to technical positions.
I spent my whole career avoiding engineering management and trying to grow in the pure technical leadership direction. One day I realized that for every staff engineer there are 10 managers, for every principal there are 5 senior managers, etc.
Turns out management is not so bad and companies seem to appreciate that kind of help a lot more
edit: also as a manager you get to work on all those pesky “It’s a people problem, actually” parts of engineering which is pretty fun. Every time in technical leadership where it felt like “Well we’ve got the plan now we just gotta incentivize doing the plan” you’re the one doing the incentivizing yay!
I was a manager, much of my career.
I hated it, but was actually pretty good at it (I worked for a company that didn't suffer slackers, and they kept me for almost 27 years). I mainly kept it, because I couldn't trust anyone else to do the job correctly.
But my heart has always been in the tech, and I did side projects, that whole time. Since leaving, I ran screaming back to being a technical implementation person, and am almost deliriously happy.
A good manager is actually fairly hard to find. It's been my experience that a majority of highly-talented developers, don't make good managers.
1 reply →
> A Principal Engineer should be just as valued and well-treated as a CTO. Most companies fail to do this, so everyone wants to be the CTO. Establish a career track, where technical people aspire to technical positions.
100%. I’ve had to fight to build a real IC track at my last three companies - I don’t care if HR wants to call everyone a manager, or differentiate with “staff”/“principal” or whatever else, but there has to be a viable promotion track for everyone, and it has to be equitable.
Just experienced something similar working-at-definitely-not-capital-one as a principal engineer. My manager was horrible, and replaced by another bad manager. He incentivized bad behavior on my team and promoted inexperienced engineers, and group think, pushed me out for questioning the status quo.
Feedback does scale if you’re willing to adjust the loss function.
I like the Jason Fried-ism of: If something really matters, you’ll hear it again. If you have to write it down to remember it, it’s probably not important.
> but you can know 10 people who each know 10 people
Theoretically, yes, but in practice those 10 don’t really know 10 people. And if you don’t hire well and don’t have everything you need to keep them motivated, some of your 10 won’t even care about their 10 people or may actively be sabotaging you, peers, and/or their subordinates.
I like the building relationships concept. Ideally, each person in your 20 person group is in a different one of the 20 total teams. The organization will never be perfect but I think it could work.
Complaining that it doesn't scale past 100 isn't really that much of a complaint!
I'd say it scales pretty darn good!
Assuming the feedback you need will just come to you might not scale.
Going and seeking out the feedback you want does not stop scaling.
Isn't this more or less rediscovering hierarchies like that of the military? (i.e. "two levels down command & control"?).
> Feedback doesn't scale because relationships don’t scale.
I would not say it this way; it is too simplistic. In fact, I generally caution against the dominant metaphor here of comparing feedback to scaling. It falls apart quickly.
Here’s a counter point. In many scenarios and settings, relationships provide transitive benefits. For example, if a leader builds trusted relationships with other leaders, a significant amount of trust can flow through that relationship.
To build a better understanding, I suggest building diverse models. Try to answer the question: What kind of qualities do relationships confer and why?
There’s also a generational aspect here. I started my career in the 2000 tech boom and bust. I’ve seen a lot of up-and-down cycles in the industry. I’ve seen lots of management styles and organizational cultures. People that had formative years during peak social media and/or COVID often have a different kind of socialization and this affects their default expectations. I won’t attach normative judgments without research, but there are significant differences.
When I think of the most impressive collaborations I’ve participated in with amazing results, relatively few of them involve tech organizations.
Building a scalable culture over various company sizes feels hard in the sense that generalizing prescriptive advice is tricky. A two person start up is cake because you only have to manage one internal relationship (a pair). People know great culture when they see it, but that is nothing like growing it.
What's with the ghostly images? Can't we have one picture that's not some disturbing AI hallucination?
> Note: The photo is of a large crowd gathering for a union meeting during the 1933 New York Dressmakers Strike. That's scaling feedback.
From the bottom of the article.
feedback fails when it is scaled down as well, the very worst manager to employie ratio bieng 1:1 where for every single consideration an employee turns to there manager, and effectivly reverses roles vs when there are several co workers who can use each other for feedback and in many ways then become self managing which of course lands us in the impossible to define world of "team building" it's mojo, nothing more
> Past a certain size, you have to make peace with the fact that a lot of people in your org are going to be frustrated with you, and you're going to have no idea why, and you may not going to be able to fix it.
I agree this is a necessary cost of being at the top of a large organization.
But I am very suspicious of the kind of people that seem to have no trouble at all being misunderstood and disliked by many. Yes, decent people can be in those positions. But it's often a honeypot for sociopaths.
> Without an existing relationship, it feels like an attack, and your natural human response is to dismiss or deflect the attack. Or worse, to get defensive. Attacks trigger our most primal instincts: fight or flight.
It is really important to recognize that it is the perception of an attack that triggers certain responses. For a counter example, watch how puppies play. It can very rough at some level but at another the intent is clearly benign.
There are ways to shape and modify perceptions! Culture. Norms. Timing. Technology. Inclusion and exclusion criteria. Information architecture.
Never assume that the technology or protocols you use have been designed for your core values. Often you have to redesign it for your purposes. Please do.
Feedback *can* scale if one carefully defines protocols to suit particular goals. We are not helpless even if it seems we are hapless. Leaders and designers (often social scientists) must step up and show better ways.
Computer scientists and software engineers must show curiosity and intellectual humility here. Better to draw broadly from other fields: social work, negotiation, psychology, anthropology, public policy, and more.
[dead]
just sent you a note Carter . . . this is something close to my heart :-)