There was once an accidental control group though. Jessica missed about half of the interviews and PG, RTM, and Trevor were left to decide on their own. I would estimate that the error rate without her was about 2x-3x as high.
The other thing to remember is that we keep track of the startups that we interview, and for the ones that we accept, we end up working with them for many years. We make a point of studying our mistakes (and successes), so there is a feedback loop, and I believe that we're getting better at it every time. But of course there are about a million other variable in play, so it's very difficult to be certain about anything (which is part of what makes this business so interesting).
But having the ability to judge if someone is a good or bad person or has a good or bad character is the part I can't wrap my head around. The inherent bias is overwhelming as this is an entirely subjective barometer) and the ramifications of playing G-d or Santa sends a screaming message that is counter to everything the post wants me to believe about YC Culture.
My definition of a good guy is someone who accepts me for who I am and doesn't pass judgement on me. I suspect most peoples judgement of good or bad character is more a relfection of how they see themselves and how the other person fits into that narrative (i.e. people want to be consistent with their beliefs, so the good/bad evaluation is based on how you perceive yourself and if this person fits into that story), versus any real indication if they are good or not.
I guess I'm saying that I am very suspect of the character and judgement of someone who thinks they are good at judging peoples character.
Edit: Just to be clear, everything I have ever heard about JL has been wonderful...but this comment about being able to judge character really rubs me the wrong way.
Edit2: Grammar, and some elaboration in ().
Which is precisely why you have to be very, very, very careful. The matter-of-fact way it was phrased in the blog post didn't give me the impression that there was a whole lot of skepticism regarding her abilities. If one does make such a grandiose claim, it seems like a person conscious that it is a grandiose claim would preface it with a whole lot of qualifiers and such.
> There was once an accidental control group though. Jessica missed about half of the interviews and PG, RTM, and Trevor were left to decide on their own. I would estimate that the error rate without her was about 2x-3x as high.
I'm curious what the sample-size is and what error-rate actually means. I realize with the control group question and this question, I risk coming off as awfully pedantic.
Speaking about unconscious bias, although this judgement can be valid, would you have said the same about say, a (male?) engineer claiming that "I can often tell good software architecture from bad". Which is a similarly subjective, heuristic-based imprecise classification problem.
Not saying you wouldn't, by the way. But I wonder if for plenty of us there are biases in play here as well, since we are talking about skills usually considered less "hard" scientifically (even though they are just as much 'hard skills' as the above example about software architecture [1]).
[1] Unless you are in the 0.00001% of developers who deal with formally specified and verified software.
This is true, but the way people go about the process of sorting out the other elements is a critical part of the analysis (pardon the wording).
I think to say such selection is an 'untractable process' strikes me as an oversimplification, as does dismissing it simpley as "something necessarily done by 'gut feel".
And I think that this (latter) oversimplification is the root of the issue that is worth discussion. This combination of elements is off-putting. It reaks of being arbitrary rather than well considered.
And so its worth more consideration.
Consider the following example.
We are to judge two Atheltes that have equal technical skills and equally strong character traits (in an objective sense), and yet we may still reject or prefer one or the other based on the fit of their respective skillsets.
With the system that is preferred by the (coach, owner, GM) of course being preferred.
That preference for style is not a preference for 'character,' (again, in the objective sense) it is simply a ... preference for the style that fits in with the ex ante preferences of the selection principals.
(not trying to be sexist and use a sports analogy, but its not really any more useful here to use YC or other VC firms as examples).
So, at one layer there is always bias towards style.
Say our team runs a fast-paced offense, that is a style-bias that is layered over technical skill and 'objective' character.[1]
YC seems to have almost made a philosophical point to form its business around some style-bias...insofar as the 'style' of millenial undergraduates cum entreprenuers in primarily male-dominated technical fiels is going to by nature exclude (of course) a wide variey of styles that clash with the 'core style'.[2]
Again, there is nothing with wrong with this style bias.
And I mean that both in the sense that there is nothing wrong with selecting for style-complementary attributes (on the one hand), nor is there anything wrong with dimensionalizing style as an explanatory variable.
Style in this sense is functional to productivity.
The upshot of this is that style is inherently <intuited> before its really comprehended rationally or analytically. This explains why style judgements seem so intuitive and 'snap', but at the same time the phrase 'rigourous sense of style' is not either un-common or innaccurate.
While 'objective' character and Style are both attributes subject to (intuitive, snap) judgement it is really best not to conflate the two.
I think we are doing the topic a dis-service by not calling out style bias as something distinct from objective character evaluation.
-------
[1] What I'll just call 'objective' character here, which has both moral and pragmatic dimensions, but you get the gist, in that its covering things like fundamental integrity, work ethic, and functional social-skills, leadership, altruism, etc).
[2] the core style being the subset of those in the target class deemed qualified enough to be alumni of YC.
I certainly hope so, although I of course can't be sure. I do think I did more than just make a bald assertion, though, by presenting an example of what I would have expected a more skeptical claim about her abilities to look like. In other words, I think I gave a characterization which could apply to either gender. Of course, whether I would have highlighted it to begin with if she were a man is a whole different matter and, like I said, I hope I would have.
My point is only that many claims subject to pretty much the same biases are usually made in HN (about hiring, about software architecture, about understanding technology/futurism) and treated with far less skepticism. That said, I am inclined to agree that the standard should be more skepticism in all cases, rather than less. With the caveat that it might not be practical up to a point.
Although, to be honest, I'd love to go into an interview with "I am X deviations above/below the mean of developer productivity in Java software analysis when compared against a sample of Y developers selected by procedure Z, measured by metric W; I am usually more productive working on teams of size n-m (p < 0.01)."
Perhaps n72 should also consider checking his privilege...
In any case, most of my emphasis was on the bias of discounting "soft skills", rather than gender bias.
The way to communicate from a country to another is really different, same for the culture, ... How do you integrate the 'foreigner' factor in the evaluation?
That said, this part (for example) sounds both realistic and useful: "They'd get excited talking to the applicants about all the things they could do, and afterward think the interview went well, when sometimes all the applicants did was reflect their own ideas back at them. But it was second nature to me to notice these things." and I imagine there are many other insights someone trained to observe people (say a psychologist or anthropologist or someone self-trained in doing just that job) could gather that most engineers would miss by miles.
Edit: Humans also overestimate their understanding of say, software systems (ask anyone who has ever tried to formally verify anything). Yet an experienced programmer can still tell a bad architecture from a good one with far more success than an untrained/novice one. Aesthetics are as hard to define as Ethics, if not more, yet designers do know what they are doing to a far greater extent than your average person. Isn't it likely to be the same for reading people?
As an example, I've had some sorrow watching some DC folks who are "amazingly good at reading the mood in a room"[1] badly misread a group of Native Americans. It was scary broken since they had no frame of reference. I would imagine a person who spent a lot of time jumping cultures[2] would get the feel, but I worry from experience.
1) direct quote
2) "a particular society that has its own beliefs, ways of life, art, etc." definition
@Sekasi, in your opinion, what would be required to test and establish a persons character?
* What they are motivated by [the motivation itself gives you clarity in their personality. Lack of motivation does as well. None are bad traits, they just reveal things]
* Posture and body language
* Eye contact and eye movement when speaking
* How they speak of other people, good and bad
* Lying [everyone lies. some people lie about important things]
* Reaction to criticism/blame
* Evidence of perseverance
* Dealing with silence [some people call them 'awkward silence'. That's pretty revealing ;D]
* Balance of aggressiveness and passiveness
* Ambition or lack thereof [no right or wrong here] * Levels of child-like enthusiasm for passions versus task driven enthusiasm
* Social interests [again, unsocial introverts aren't necessarily a 'bad' trait. Just something that adds to your personality]
* Instability
I just wrote those up then, it's hard to verbalize, admittedly.
The emphasis on avoiding false positives means that errors are more easily identified: bad character stands out in the small pool of YC founders, whereas good character may be common in the pool of unselected applicants. In other words, it's more of a bloom filter than a dictionary.
Think about it: The only way that you could change your view of your own abilities is by comparing your predictions about people's character to their actual behavior over time. Jessica did the same thing, but she found that her predictions were accurate.
JL has the opportunity to make judgements about people, and then spend months or more with them as a part of YCombinator. She can say "I don't think the members of this startup will work well together" or "I don't like the looks of this guy", and then see if she was right or not. She has the opportunity to do this dozens of times per year, every year. That's how she knows.
Also, she gets feedback so she gets to improve.
"I've always been a good judge of character–even as a little kid"
"In fact my cofounders respected my opinions about character"
"I can tell things like when people are BSing, how tough they are, if they get along, .. whether they're good people or not."
"Paul, Robert and Trevor are really good at technical stuff, but it's precisely because they're so good that they often misjudged founders as people. " while "I didn't always understand the details of the questions they'd been asked"
"I just have a very intense reaction to people I think have bad characters. "
At this point my reaction is as it always is when someone belive they were born with special abilities, but the article does end on a better note:
"I now have experience dealing with literally hundreds of startups. So now when I see a problem I've usually seen the same problem ten times before."
Born ability? I would call it earned experience, while I strongly suspect she was as bad as anyone else at being a judge of character before that.
She is one of four people who founded YC. YC is wildly successful. Paul Graham's writing is practically worshiped here on HN.
I am rather appalled at the reception this is getting.
I don't know her, but I will suggest that the substantial success of YC suggests she isn't just fucking self-deluded. Perhaps she isn't explaining it in a way that appeals to the data-oriented mindset of the majority of people here, but the degree to which she is being outright dismissed as utterly full of shit smacks of, among other things, sexism.
Absolutely. I think Paul Graham said silly things about hacking and painting. I think Paul Graham is utterly wrong about philosophy being just a confusion over meaning. I think Paul Graham at time tends to riff on things he just doesn't know much about.
> outright dismissed as utterly full of shit smacks of, among other things, sexism.
Please show me where she's being "outright dismissed."
Please show me how I could make the argument I made without it being interpreted by you as sexist. Or is the mere fact that I made it about a woman proof of its sexist nature?
No, it is not the mere fact that it is about a woman. It is the context here and the framing of the comments. It is the fact that, on the whole, this piece is not being talked about with the kind of respect with which her spouse's writing gets treated here. It is the fact that you and others are not asking how she knows or what it would take to prove it or any other questions that come from a place of "Well, this very powerful, intelligent, successful person has said something I do not understand and I would like to understand it better and learn from it." No, the comments here are mostly dismissive of the very idea that anyone can be any good at this or can objectively know they are far better than average. Most men get asked to show their work. Women get told they are full of shit. There is a difference between the two, a very big difference, and that difference boils down to fundamental respect.
It is in part a contextual thing that is very hard to prove bias on, but if you experience it, you eventually know that there are just no circumstances under which you will be taken seriously and there will always be some excuse and some plausible explanation why no one takes you seriously or respects you while they claim it is unrelated to your gender.
I will just leave this link here on the off-chance that you genuinely would like to understand what I am trying to convey:
http://micheleincalifornia.blogspot.com/2014/03/the-creepine...
The great part about this business is that I can make money by investing in those that others are missing :)
I am not talking about "women need to be liked." I am talking about "Women need to be able to access knowledge that is currently overwhelmingly held by men." If they can't effectively do that, they are going to tend to have lower success rates.
Serious question. I had the same reaction as you to these comments. It seemed like Jessica was getting a lot of skepticism relative to what PG or one of the male partners would have received from similar claims.
But I don't know how I would change things. Any suggestions?
Note: I'm truly baffled that many commentors here seem to think that judging character is i. impossible to do correctly, ii. undesirable, iii. skill in judging character is intrinsically impossible to evaluate
If you are male, some things that you can do include:
Engage women in discussion about ideas and a) don't gush at them just because it's a woman speaking up b) don't give them condescending pats on the head of "encouragement" as if they would be more successful if only their girlie little feelings are helped to be more positive c) and invite them to push back.
I previously had a blog post about "shoving matches." I am not currently finding it, which may mean it was not republished when my blog was moved, likely due to combination of it probably wasn't that well written and I am frequently busy with other things. In a nutshell, think of the scene in "Remember the Titans" where the white team captain gets into a shoving match with a black team member. Initially, the black team member is afraid to push back. That's the kind of thing that could end badly in the racist environment in which they live. But the team captain shoves again and invites him to shove back. He does finally shove back, they engage in a friendly shoving match, and that is the moment where the team becomes a team and race stops mattering.
This can be done to a woman. It is possible to encourage her to argue with you in public in a way that is not dismissive of her. It is possible to invite women to stand their ground and defend their ideas. And that's one of the big things that needs to happen on HN.
I think (hope?) I'm fine personally as far as female engagement goes. I run a forum in my niche, a niche that's slightly majority female.
Having said that, the new moderator Dan G. has been a very valuable resource for me in the last year. I think appointing him was a good move and I think things are already moving in a better direction for HN wrt the participation of women here.
You may be right, maybe I called Jessica out because she's a woman. I really don't think so, but am open to the suggestion I did. I really don't see anything in my post, however, to indicate gender bias was a factor. In fact, I was calling her out for precisely the type of thinking which leads to gender bias.
So, most people with good social radar are not going to be able to readily give the precise, data-driven explanations that the HN crowd values so highly. But the reason I suspect Jessica is being dismissed based in part on her gender, rather than the hand-wavey nature of her assertions, is because of the framing of the criticism she is getting. As I said earlier, a lot of the comments here do not boil down to "please, show your work and prove you are right." They boil down "Oh, you are full of shit and no one can know that and, god, my respect for YC just plummeted" as if she had confessed that her method for deciding who to accept involved casting their horoscope or consulting a psychic.
When I first joined HN, the top three people here all had about 50k karma and there was a much more collegial environment than currently exists. I was never given access to that collegial environment. I never got a mix of some acknowledgement that I might have a point and some criticism that not everyone agreed with me and the opportunity to state why I thought x, y or z. Instead, men on the leaderboard closed ranks. For a long time, if I said anything, you could bet dollars to donuts that two different people (not the same ones every time) from the leaderboard would chime in to announce how wrong I was.
It took me literally years to figure out that this was going on in part because, for a woman, I had a high amount of karma and I was, thus, rocking the boat. Initially, I had no idea because, at the time, my karma was so far off from qualifying for a bottom rung on the leaderboard that I figured I was a Nobody and no one should be noticing me at all. However, I was able to infer (based on that social indicators in the way people reacted to me, in spite of the lack of hard data – in other words, “social radar”) that I was apparently prominent for a woman, in spite of my pathetically low amount of karma. Eventually, I was able to dig up real data proving that fairly firmly (in spite of the limits on the data available to me). Once I understood the position I was in, I was able to quit attracting so much damning attention.
So what I am telling you is that a) Hacker News has deteriorated in some ways generally, which has hurt the experience of most members and I would like to see the bar raised such that the collegial environment returns generally and b) I would like women included in that respectful debate of ideas in a way I do not believe they have been previously included. And one way you can help achieve that, if you so desire, is to change the manner in which you criticize women and their "hand-wavey" attempts to express themselves.
One thing that would really, really help a whole lot is make damn sure you understand that "absence of evidence is not evidence of absence." In other words, just because women are generally going to not be good at defending their positions on terms that the men here will readily accept does not, by itself, prove they are wrong. Just because the explanation is hand-wavey does not mean she is full of shit. Ask better questions. Encourage women to get the data that backs up their position. If they can't immediately do so, genuinely "agree to disagree" instead of pissing on them like they aren't worthy of any respect. Give women an opportunity to grow into the ability to communicate in more precise terms and learn to effectively back up their statements. Because shooting them down and pissing on them is not something that will ever help women up their game and learn to go toe-to-toe with the big boys. It merely excludes them.
And thank you for reading my blog post and engaging me in good faith.
I do however vehemently disagree with this: "In other words, just because women are generally going to not be good at defending their positions on terms that the men here will readily accept does not, by itself, prove they are wrong." I think here you are in fact buying into gender stereotypes and doing women a disservice. Some women can be and are just as analytical as men and some men can be and are just as strong with the soft skill stuff as women. To start thinking of these things in terms of gender just perpetuates the status quo. If anything, YCombinator itself is doing that very thing, with the three men as the analytical ones judging the tech and Jessica, as the woman, as the soft skills person. It's quite possible that it's just a coincidence it worked out that way and they all seem to be happy with those roles, but isn't it a little bit fishy that their roles line up with gender stereotypes so neatly?
I do however vehemently disagree with this: "In other words, just because women are generally going to not be good at defending their positions on terms that the men here will readily accept does not, by itself, prove they are wrong."
I apologize if it sounded like I was promoting gender stereotypes. I have my own personal biases that do sometimes come out that way. But let me be clear: My intent was to make the point that women are often ganged up and dismissed in a way that is toxic and really hard to cope with at all, much less argue effectively in the face of. So, given the degree to which women are underrepresented in STEM and on HN, I think a reasonable assumption is that most women here will have a certain lack of experience of effectively going toe-to-toe with the big dogs. It would be nice if the big dogs did not try to eat their lunch and, instead, engaged in some basic respect for boundaries while women try to get their sea legs as HN gradually changes to become more female-friendly.
I don't hesitate to debate with anyone here. For a long time, it really made my experience here miserable in a way that did not seem typical for what was happening with men who engaged in debate. I think the collegial environment that existed at that time, from which I was largely excluded, has deteriorated some, thus harming the experience for everyone. I would like to see the general atmosphere here improve to something more akin to what it was when I originally joined. But I would like to see it not reserved for the men this time around. I would like it to be a more inclusive environment.
So let me attempt to reframe that as: If a member here happens to not be good at defending their position, regardless of who they are, that shouldn't be reason to simply be dismissive of them or make them look stupid. It is a good general rule of thumb. If it is actually applied consistently, it will help level the playing field so women aren't facing such an uphill battle.
Edit:
As for making you feel unjustly attacked, I will note that my original reply to you was this one: https://news.ycombinator.com/item?id=9147934 You opted to not respond at all to my assertion that people can tell such things. You just kept arguing that there isn't enough data, the example offered you by paul was not sufficient and so on. So, in the context of the general atmosphere, that began to look to me like, no, your inquiry wasn't really about "So, how do you know?" Instead, it was a polite way to be dismissive -- that when push comes to shove, "la la la not listening -- no amount of data will ever be enough." You seem to not be giving much credence to the things that are being said in support of what she said.
So that is where my parent comment of this particular discussion came from in asking why you are being so dismissive.
I do understand that it is often more engaging to argue with people where you disagree and that there can be a lot of non-sexist factors influencing your choice as to what to reply to in this discussion. But having been on HN a number of years, I have seen this far too often where I ever so politely have my character assassinated in ways that would make me look like a lunatic to protest it. It's really maddening. So the context and pattern of behavior here matter. I mention that since you seem genuinely interested in understanding how to do this better.
Take care.
But this isn't something people are pulling from nowhere. There's plenty of research showing that people are lousy at making judgements. She hasn't said how she can make good judgement or how she knows she makes good judgement - there's just a bunch of handwavey stuff and people are criticising the handwavey descriptions, not necessarily her work.
There are well developed paradigms in, for example, police work and courts of law for judging who is probably telling the truth, who is almost certainly full of shit, what the evidence supports, and so on.
I will reiterate: Just because most people are terrible judges of character and overestimate their own ability does not prove Jessica is terrible at it. And just because her explanations are hand-wavey and not as precise and data-driven as the HN crowd would prefer, that does not prove she is simply self-deluded and full of shit.
If people were really criticizing the hand-wavey descriptions, as you claim, they would be making different criticisms from what I have mostly seen here. The door has mostly been firmly closed to discussions of how this works when it is done well. There haven't been questions about how to do this in a more data-driven, less hand-wavey way. It has mostly just been a matter of pissing all over her and dismissing her claims.
All the research I've seen shows that people are lousy at these.
To me her comments are a bit worrying: it sounds about as bad as any other stupid interview technique and it seems like it leaves a big risk for subconcious biases.
Take something as simple as a suit jacket. In one country you leave it on while you're working. Taking it off is sloppy and unprofessional. In another country you take it off and roll your sleeves up. You're working hard, you get hot. Keeping your jacket on means you're having an easy day. These are mostly uncociois and if you're not aware of ot and correct for it you risk calling someone sloppy or lazy eventhough neither is true.
HN frequently talks about bad interview practice; HN sees this as either an example of bad practice or of good practice poorly explained, and so they talk about it.
YCombinator partners get to interact with founders in a variety of situations, over a long period of time. It's not like airport security trying to pick who's a terrorist or a jury evaluating a witness based on formal testimony. It's watching people as they behave and interact with each other, often when they don't know that they're being observed. Hundreds of different people a year, for close to a decade now. The YC partners are in a unique position to develop views that are far more nuanced than "this person is sloppy because he's not wearing a jacket".
Every professional has subconscious biases, from doctors to engineers to investors. What makes them professionals is that their biases are usually correct, and they have sufficient experience to know when to question them. For whatever reason, we react negatively when someone evaluates their own abilities. But it doesn't make sense for everyone to refuse to self-evaluate, or to always speak poorly of their own abilities. When you get to a certain age and level of confidence you know what you're good at and what you aren't good at. People should be able to state the truth as they see it about their own weaknesses AND strengths. And we should be evaluating the validity of those statements, not that their choice to state them.
This is so wrong. Doctors have biases - "this treatment works; I know it works because I see the effect it has on my patients". Doctors have access to evidence bases that tell them about the actual effectiveness of that treatment yet we still find real medical doctors using homeopathy (and not just as a placebo but because they think it works).
I don't doubt that most people suck at it. One of the four people who co-founded YC is hardly "most people." Your general skepticism fails to be a good argument in this specific instance.
That's exactly when you should be more sceptical! You need to work past survivorship bias.
It's odd that you think Paul Graham's writing is only worshipped here. There's criticism, too.
The halo effect around YC is a form of delusion. Person making important gut-based decisions about people's lives gets criticized for publicly announcing her own competence. Why is that appalling?
As one example I can share:
When I worked at a Fortune 500 company, I once got myself moved to a different team very, very quietly because the individual taking it over was going to be trouble for me. He had not yet done anything worthy of bringing it to the attention of HR, but I wanted nothing to do with him as my new boss. A bit over 2 years later, he was quietly fired. The rumor mill had it that he was either having an illicit affair or was engaging in sexual harassment. I believe either or both as plausible stories. A former team member (a woman) and I then talked about how terribly he had treated her, something she had been unable to effectively address while she worked for him. I felt terrible for not telling her sooner why I had been quietly moved.
If you earnestly want a more in-depth write up, I would be willing to put one on my blog sometime.
Edit: Maybe "quietly fired" isn't a good descriptor. More like "disappeared" overnight. One day, he was his usual jovial self. The next, we came in see a manager cleaning out his desk. The many people who bought his smarmy bullshit and adored him were in shock. I spent the week trying to hide my desire to pass out kazoos and party hats amidst a funerary atmosphere at work. Lots of people were completely taken in by this creep.
How do you know you have this ability and it's not just bias? I really don't mean this as snark, but increasingly more research confirms that humans far overestimate their abilities in this area. In my own case, as I've grown older I've come to realize that I'm not nearly as good a judge of character as I thought I was in my early 20s. So, again, not snark, but how do you know? Have you set up a control group?