I have felt this too and it turns out I leave my startups around 200 people -- when I can no longer remember everybody's name instantly.
(but where's the line? I worked with a guy who wouldn't work for a company of more than a dozen people. I've looked at his linkedin over the years and indeed he worked for some phenomenal startups...but always left when the started to grow).
The Dunbar Number (the number of relationships any individual can maintain) is typically considered to be around 150-200. So there's a pretty good reason for that 200 number.
The number of individual channels of communication in a company hits 100 just past a dozen, which often makes it an inflection point for company culture and might affect why your acquaintance tends to leave at that point.
I have a working theory that these inflection points follow the Fibonacci sequence. At least based on my experience, communication and culture changed dramatically at each stage, most notably around 34 and 55
Not sure if it's the fibonacci sequence per se, but yes, there are inflection points around those numbers.
Up to ~15 people with experience and a common purpose, you can just all work together, little structure and roles, and stuff works OK. Teams form ad-hoc, based on problems.
From there, you need to define more well-defined roles, so people don't step all over each other. Communication starts being important, because everybody talking to everybody has broken down at that point. Hey, look, a wild manager appears.
You're a startup, you try to save on managers, so teams can easily balloon to ~20-30 under a single manager, but then it breaks down again. You realize you need a second manager.
As you approach 50, you want to add a third manager. At that point, you have, want it or not, a "leadership team". There needs to be very regular and deliberate communication. Policies start being set, because teams start drifiting apart.
You'll hit the next inflection point at 100-200 people. Because at that point, your leadership team is starting to hit communication limits at the leader-to-leader level. (Your leadership reaches the same initial inflection point your original team reached).
As that cascades up, you add layers. The problems keep repeating. The next hunk is at ~500 people. The "old hands" feel like the amount of people they don't know really outnumbers the people they know.
Those are the sizes I've personally seen happen, several times the same pattern. And based on that progression, I'd expect the next transition point at around ~1,500 people.
So it's not quite fibonacci - it seems to skip steps - but it's a good enough guide for the first few 100 people.
> I'd expect the next transition point at around ~1,500 people.
I wonder if this is present, at least for the same human-relationships reasons as the others.
It seems like there ought to be an upper bound past which people are basically disconnected and lack personal relationships to most of the organization, at which point those relationships become less relevant to structural issues. I guess the ~200 and ~500 problems will keep recurring within subgroups like divisions or offices, but excluding that it seems like non-social constraints might dominate.
At 1,000+ person companies, it seems like the problems I hear about are often structural (e.g. the contractors are treated as second-class citizens) or logistical (e.g. we can't get the bulk of the company to an all-hands, even once a year). At 10,000+, purely-statistical treatment of employees seems to take over, like IBM's systematic dismissal of older workers. (And resulting individual-level idiocy, like bringing back some of those essential workers as contractors at higher rates.)
I'd wager a medium-sized amount on the 1,500 inflection point. It's based on a vague feeling of being in a 1,200 person office, and feeling like things were likely to change soonishly.
Also, for large companies, you have separate social dynamics for offices. I've seen the 200/500 point happen in offices of much larger companies. You then layer the "too large to comprehend" dynamics on top of that, for added fun :)
(Also, you can get 1,000 people offsite. It's just a lot of work. That might be another of the 1,500 size problems)
Either way, social dynamics at large institutions is a fascinating problem. I wish I could find decent literature around the topic, because it's a fun subject :) (Corollary: All the things I said above are anecdata, I don't know if there's a theoretical foundation)
Interestingly, 30-50 people is the standard range given for 'band societies', the most basic pre-agricultural group. That's been cited as a problem with Dunbar's number, but most people seem to think it's just another major "natural group size" among many. Dunbar has also cited 5 and 15 people as rough 'rings' to which we devote disproportionate amounts of our socialization time.
I have a bit of trouble judging how Fibonacci compares, since the distinctions I've seen are gradual rather than nonlinear. But it does look like moving from ~15 to ~35 might match a switch from close-association to band structure, and exceeding ~55 would match another restructuring away from that.
You might have 30 people who live together; but even just to avoid inbreeding, their social milieu would likely include a handful of other bands with nearby territory. Thus any given persons social ties could very reasonably number in the low 3 figures.
Lieberman ought to more than I do, but his argument that "150 connections" can't be selected for because pre-agricultural bands were smaller seems ridiculous. Most fundamentally, because 150 doesn't have to be directly selected for; it's a consequence of working memory that might be valued for unrelated reasons. But more directly, because "bands of 30-50" is not "only knowing 30-50 people". Bands were usually extended family groups, so the value of maintaining personal relationships outside of those groups is obvious.
Honestly, the best I can work out is that Lieberman is a few decades out of date on his understanding of how pre-agricultural groups interacted.
200 is on the order of Dunbar's number, the theoretical maximum number of individuals a human can keep track of. The most commonly cited value for Dunbar's number is about 150, but there's a range between 100 and 250. Dunbar's number is thus a theoretical upper bound for the size of a startup a single person can comfortably relate to every employee of, but some people just aren't comfortable even with a company that size; presumably they want to leave plenty of headroom in their social network for family, friends, church, bowling league, etc.
I've always been a bit sad that there hasn't been more empirical work on Dunbar's Number. It looks like attempts to validate the modern number range from around 150 to around 250 (Bernard–Killworth, who seem to be leaders here), but of course that variance has a lot to do with exactly what counts as a 'connection'.
You're very right about the mistake of looking only at company size; Dunbar suggested that maintaining the full connection set would take a very large amount of socialization time, and most people have at least some connections outside the office. Even anecdotally, there are lots of people who describe feeling disconnected at work because they spend more time than coworkers on non-work connections. This seems to have been ignored by a lot of people using the number. Wikipedia claims the Swedish tax service limits offices to 150 people based on Dunbar, but that could easily be 50 people too large for a workplace community. Malcolm Gladwell, for whatever his data is worth, claims that Gore-Tex hit cultural problems as offices grew beyond 150.
If those anecdotes are accurate, it implies offices function well with high rather than total connectivity. Intuitively, that seems quite reasonable. Cross-organization connections are valuable, but not every pair of jobs has equal value-from-connectivity. And small teams are infamously hard to build because everyone involved has to get along and agree on how to function. If you aim for 90% connectivity (or 50%, who knows) instead, it's possible to maintain a high-empathy organization while still allowing for personality conflicts.
Why does it matter? If your startup has 12 or 200 people, it has lots of clients and vendors who you aren't friends with, just like people in a different org in a large company.
That's a reasonable question. The simple answer is that I just become uncomfortable when I don't know everybody's name (yet they all know mine).
The other part is that it's a company (a group of people) not just a corporation; while I am friendly with some clients and vendors (ones I worked on, either old ones or major ones) when we happen to hit it off, it's not the same as the company of people.
I do wonder if it isn't necessarily that the 'generalized person' loses empathy as their place in our economic power structures change, but instead that they are more empathetic with the people in their current position in the power structure? It seems natural that a person might do so.
I went on a quick search to see if such a thing as "Winner's guilt" exists, and while it does appear to be a thing to some extent, I found not studies on it in a larger context. It seems to me that relating to people that are like your current state would probably be a good defense against that, psychologically.
> I went on a quick search to see if such a thing as "Winner's guilt" exists, and while it does appear to be a thing to some extent, I found not studies on it in a larger context.
I've read about this before, and I think it's true. But has the phenomenon been examined in context? By that I mean: let's say someone becomes wealthy or powerful, and they are known to be wealthy or powerful. The nature of their position allocating resources means that others (not every one, but a significant number) seek to instrumentalize them; i.e. try to get the resources flowing in their direction. I wonder what portion of the decrease in empathy is a response to instrumentalization.
I feel like power is an issue often solved by education or something. The more you understand about what is going on in the world, what has gone on in the world, humanity, how people work, etc. the more you realize the guy in front of you isn’t shit (and neither are you). Of course reasonable prudence means monitoring your behavior so you don’t end up broke or homeless. But this may include long-term life approaches as much as (more than) what your boss thinks of you (in my opinion)
I can't speak to "most founders", but those who have impacted me the most (like Jimmy Treybig) or even my own experience ... I have not found this to be true. Maybe I just need to get out more...
I think it boils down to 'most founders' not continuing to walk around and talk to people. In my experience the non-empathetic leaders were the ones that were rarely if ever seen interacting with the general staff, either through casual conversation or meeting with groups of people. Point being, if someone is in a position where they are 'steering' groups of people or changing organizational structures then they should be interacting with those people quite a bit beforehand.
One of the things that sickened me most about myself happened about a year ago when I took a trip to Philadelphia as a tourist. My wife and I were walking around, and I make the comment of "there doesn't appear to be any homeless people here", after which my wife immediately pointed to a homeless guy that we just walked past.
Somehow, I'd managed to shelter myself enough and live comfortably enough to where I stopped noticing less-fortunate people. It really depressed me, because I had previously thought of myself as a fairly empathetic person.
To help combat this I donate a fair amount of money ($50/month) to mental-health research and make a conscious effort to pay attention a bit more when I'm walking around.
I'm hardly a hyper-successful person, so I can't imagine how hard it would be to empathize if you're a multi-millionaire/billionaire.
EDIT: I know throwing money at a problem hardly counts as "solving" the empathy problem, but it does help me assuage some guilt.
If you have time, I'd recommend volunteering for local soup kitchens and donating nonperishable goods, hygene objects, etc. to local shelters. Exposure will help prevent empathy erosion and force you to notice/understand the problem space of the less-fortunate. First-hand accounts and documentaries regarding experiences not your own will also aid you.
After that story happened, I actually moved to a less-afluent neighborhood in NYC (for other reasons), and I've (very) recently started tutoring people in math at the local library for people to get their GEDs, which has also helped me somewhat.
A soup kitchen isn't a bad idea though. I will look into that.
I would like to commend your actually going out of your way to deal with something you've identified you wish to change about yourself. Also- I noticed you're in NYC. Would you be opposed to having a coffee chat sometime? (Feel free to decline- no stress. This is HN after all.)
That said, a lot of homelessness happens as a result of mental illnesses like drug addiction or schizophrenia, or degenerative diseases like Alzheimers. It might be a pipe dream, but I would like to have as much research as possible to help treat these diseases, to (hopefully) help reduce these issues, perhaps by developing a cheap medication to reduce or eliminate the effects.
It's not something any of us are immune to; any of us could develop Alzheimers when we get older. I'm fortunate enough to have a lot of really cool friends and family that would take care of me if that happened, but not everyone is so lucky.
> That said, a lot of homelessness happens as a result of mental illnesses like drug addiction or schizophrenia, or degenerative diseases like Alzheimers
Those with mental illnesses who are also homeless are more visible, but the majority of those who are homeless are homeless because of financial reasons. And often, they're entire families.
For NYC[1]:
> Research shows that the primary cause of homelessness, particularly among families, is lack of affordable housing. Surveys of homeless families have identified the following major immediate, triggering causes of homelessness: eviction; doubled-up or severely overcrowded housing; domestic violence; job loss; and hazardous housing conditions.
(but where's the line? I worked with a guy who wouldn't work for a company of more than a dozen people. I've looked at his linkedin over the years and indeed he worked for some phenomenal startups...but always left when the started to grow).