hckrnws

> 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.

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?

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.

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 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.

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.

The same principle holds for quality management. You don't need to inspect every single product. However, if you inspect a small number of products at random, you'll detect a large percentage of the quality issues.

While leaders can't know everyone they should make it a priority to have those random connections outside their inner circle. If they don't, they become in danger of hearing only the info that their inner circle wants them to hear.

> However, if you inspect a small number of products at random, you'll detect a large percentage of the quality issues.

As quality issues become fewer, the odds increase that inspecting a small number of products at random will lead to you thinking that there are zero quality issues. Have your inspections procedure scale and adapt to the relative proportion of quality issues you have reason to believe exist. And if you believe you truly have zero quality issues, then you need to switch to an immediate feedback procedure (such as an anonymous tip line, or a non-anonymous one for customer feedback).

> There's a difference between trying to micromanage everything, and micromanaging enough that you're not out of touch.

I think there's a good point to be made here that this isn't micromanaging, it's bypassing feedback layers that have a tendency to filter out critical or important information. That information may or may not be withheld intentionally, but being Bill Gates and seeing that a crucial tool to help a customer doesn't work very fast, or is missing information, or relies on "hacks" (tribal knowledge on how to bypass restrictions or flags) to keep the support process going would be something that wouldn't filter upwards easily.

Definitely a balance to be had though for sure.

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.

> 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.

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.

> 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.

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

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.

What do you feel you cannot share concerns with your peers?

Not the original poster but:

  * 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

This is a great example of why managers have to exist.

Managers are rarely the right people to take these concerns to

They are political beasts, and unless you have some political capital with the Manager.

Long story short - if the other guy is seen by the manager as more valuable, you speaking up will get you a one way ticket to the door

Ugh, your attitude really pisses me off, but I want to help you because my liberation is bound up in yours, so here goes.

> It's layers above me that need to hear it

Most workers socialized under capitalism feel this way, that the power rests at the top of the hierarchy and IF ONLY THEY KNEW, they could FIX THINGS. Well, guess what? Your job is to keep them from knowing. You, as a leaf-node of the hierarchy, operate "the sharp end of the system" where "all ambiguity is resolved."[0] You exist to DO the WORK, and that includes the "theory building"[1] from which the owners of the business pay you to be insulated.

However you interpret that on a moral level, practically speaking it means that YOU and your peer practicioners are actually the ones with the power and the (sometimes merely implicit) mandate to enact whichever policy you think the "layers above" ought to impose.

If you REALLY need something from the higher-ups, the only real way to get it is to march on the boss and, with sufficient leverage, demand it as a collective. You're going to have to talk to your peers to organize that, or we'll slide further into thisbdystopia in which "we fear our neighbor’s opinion more than we respect our own freedom of choice."[2]

To effect lasting change, one must act with consistent commitment alongside one's peers, rather than waiting for a moment of grace from the "layers above."

"Loyalty, which asserts the continuity of past and future, binding time into a whole, is the root of human strength; there is no good to be done without it."[2]

0. How Complex Systems Fail

1. Programming as Theory Building, Naur

2. The Dispossessed, LeGuin

> I can't cope with conflict in any shape or form

Why would you live this way?

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)

> 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.

Complaining that it doesn't scale past 100 isn't really that much of a complaint!

I'd say it scales pretty darn good!

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.

> 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.

> I hated it, /../ I mainly kept it, because I couldn't trust anyone else to do the job correctly.

I think the best managers are people who will even do management if that's what it takes.

> 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.

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.

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.

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 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.

> 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.

just sent you a note Carter . . . this is something close to my heart :-)

[dead]