What's more important, is I don't understand how anyone could possibly improve their 30 second impression.

I guess it could be that you believe this to be the case, and are remembering wrong. Why don't you note this down after 30 seconds, and then check it only after you make a decision to see if you were correct?

Companies would like you to think this, too, and law makers would very much like to think that their rules regarding non discrimination have any effect. The real situation tells another story.

But it's OK, companies can just include a nice stock picture of an ethically diverse group of people in business attire working together, an all is well.

This is not actually very hard: start with good grooming, nice clothes, posture, confidence.

Of course, this doesn't change the fact that the hiring process seems... odd.

Appearance is important, and I've never been in an interview situation where it wasn't? Unless you're insinuating choice based on ethinicity (more aimed at _pmf_'s response to your comment). Which is completely irrelevant, as are accent/sex/other discriminatory factors that make zero difference to how someone is likely to work within a team.

As I've said in other comments, there is a serious element of external preselection based on geography and company sector, there's nothing particularly underhand going on.

This is just plain ignorant. Every single place I have worked has had its share of everyday sexism and racism, and that includes how current employees act in hiring interviews. You claim to not notice these things, but what you are saying is that you don't consider those things valid differences that may affect how someone perceives your team or work environment. How can you ever confront issues in your workplace by shoving your head in the sand?

Given the way you're talking about making a hire/no-hire in five seconds, I find it incredibly difficult to buy the idea that you're not taking that into account. You're glorifying the snap-judgment hire.

Because most people do, and a good hiring process takes that into account and counterbalances it.

Maybe your company sucks at interviewing? Maybe you're lucky and all of your candidates are good enough to do the work so you can cherry pick the ones that "fit"?

I've interviewed a reasonably large number of people, and I've worked with some folks who were good at interviewing, and there were many times where it took most of a day to figure out whether or not someone was worth hiring.

What I've found is that credentials, like a CS degree, are generally garbage and bear no relation to whether or not someone can code or problem solve. Experience and resumes can be misleading as well and rarely tell you for certain whether or not someone can actually do the work. Maybe something like half to 2/3 of the time it's obvious within the earliest portion of an interview that someone just doesn't have the skills. The remainder often takes a lot of work to suss out skill levels. Rarely someone will knock it out of the park and be an obvious hire. More often than not it'll take several interviewers concentrating on a few weak spots to see how weak they truly are and figure out if they're still a good hire. Not to mention the problems of trying to gauge someone's skill level in a high pressure environment where they will absolutely not be at their best.

Increasingly I get the impression that silicon valley is becoming like a giant frat house where companies just hire people they like and then bang out horrible code and fail to execute on projects until the VC money dries up or they get pity-acquihired.

1. The IT world in the [geographic] area of the UK I live/work in is small (this is critical, because experienced candidates have a high likelihood of being known). 2. The company I work at isn't a tech company. 3. The company operates in an unfashionable sector.

Shouldn't technical interviews be different? No, not really. In general the interviewer will give more hints to people he likes and more follow up questions to people he dislikes. After enough hints even the worst candidate will find the solution and get extra points for good collaborative skills. Similarly even the best candidate will stumble sooner or later and gets a mark for being argumentative if pressured enough.

Also notice how everyone says "We care about the thought process and not if the solution actually works!". This allows for their subjectivity to reign supreme. Things like "Your solution is wrong but I love your thought process!" or "Your solution is better than mine but you didn't think things through enough before implementing it" can be said about everything.

It is of course not that hard to do technical interviews correctly, you just need to have hard rules on what to accept and what to reject, what hints to give and when to give them etc. Any leeway for subjectivity will be abused by your subconscious.

So you just admitted that you make your hiring decisions based on a snap judgment. OK. But that doesn't mean you can tell in 30 seconds is whether someone is actually going to be good at the job.

It's usually not a conscious thing, and it's something we're all prone to doing.

But he's also right. There's been studies on this. Avoiding initial impression from dominating hiring decisions is incredibly hard and requires a ton of process that pretty much "nobody" does. Few places do much to handle even much more apparent biases.

It's interesting that you're allowing hiring decisions to be influenced by what you think is best for yourself. What's best for the company can and should be a secondary concern. I've also found it useful to do the following:

1 Telling HR that candidates are flaky. It's easy to convince HR that every single one will leave in a year or less, especially these hipster kids that are interested in the latest startup.

2 Telling HR that candidates will take forever to add value, even if we have a smaller component that needs working on. This is still easy to convince them, given the ramp up to some of the prototypes that we do.

3 If the candidate actually makes it to the interview stage, I just finish my questioning and tell HR that the candidate has an attitude problem. This works even if they get every question right, or if they go into detail. The latter works especially because they think they're showing value, but they actually make themselves easy to label instead as arrogant up-jumpers with a "behavioral problem"- and I can easily convince my HR of that.

4 Telling HR that candidates don't know a specific tech well enough. Even if that tech has a ramp-up of a week or less. Because again, HR will take my word for it.

You're also right about those thirty seconds. It is the best and only approximation that one can make. Forget the interview, forget allowing them to show value, and if they do, they have an attitude problem.

Good luck with your recruitment efforts and next quarter.

I find that phone interviews actually work differently - I've been able to turn around an unenthusiastic interviewer and I've been able to lose someone who liked me. For that reason, I think questions of actual code aptitude are best done on the phone, as that's the only place where you really can evaluate someone's thought process isolated from mitigating psychological barriers that occur during face-to-face interactions.

The icky reality is that face-to-face impressions are pretty much determined by how you smell. I think there are two reasons for this - first, there's the concept that people naturally want to be around winners and if you come in looking like a loser they'll be predisposed to dislike you in ways that escape their conscious thought. To work this one, try to chalk up a win before your interview (or get laid - again, hate that it works this way).

The second is that as people we unconsciously recognize whether or not the candidate believes that they're a good fit for the position. If you go in freaked out, you're going to set off a bad feel no matter how hard you try to cover it up. The feeling a recruiter responds to in the first thirty seconds is "this could work". You can't fake this, and this is why being as good as possible at what you do really does matter. It definitely rewards intrinsic overconfidence, but the more intrinsically underconfident people are generally smarter anyways and will eventually be so much better at what they do than their peers that they'll win in the long run. If you feel insecure in your ability to fill a position, get better at what you do and prove it to yourself by having other people evaluate your work - the feeling will go away eventually. Code review is scary for people insecure in their abilities, but it's easily the fastest way to wipe away pathological insecurity.

You know from the resume whether the person can do the job. It's whether you want to sit next to them for 8 hours a day while they do it.

30 seconds? Yep.

Do you? I remember reading a bunch of articles that asserted most people that come in for programming interviews (even if their resumes look solid) are useless.

It also helps that I live in Melbourne, Australia - the IT world here isn't massive, and if you haven't worked with the person the odds are you know someone who has. :-)

It certainly takes longer than thirty seconds (or a few minutes) to determine that someone who looks good on the surface, who is clean and neat and who sounds like someone who would fit in the team fine, should be avoided like the plague because they can't actually do the job.

I must say I'm not advocating snap judgement at all. I just think it is very common (as Kahnemann states). What he states as an ideal doesn't necessarily make an awful lot of difference in many situations; the snap judgement can very often be the same as the careful one (though external factors that partly force/aid preselection bias things greatly in this respect).

Do you have any evidence at all to back this up? In general, people think they are much better at making decisions than they are, and unless you actually measure things empirically and track false positives/negatives, you are likely to fool yourself.