For what it’s worth, at Facebook the general hiring process is in two steps. There’s a centralized hire/no hire that most people are aware of and then a second phase where team selection happens. This is where all the “hire for weakness” steps that the author mention happen. Teams are looking at this pool of people for speciality, levels... that they need and where the matchmaking happens.
There are also a lot of smaller but specialized longstanding loops like front-end, AI... and one off “reqs” for special roles that teams need.
> at Facebook the general hiring process is in two steps. There’s a centralized hire/no hire that most people are aware of and then a second phase where team selection happens
This is not quite how it works. I was a hiring manager at Facebook for several years. It is made very clear to EMs that team selection is not a second interview or extension of the hiring process. We do not consider team selection part of hiring.
Now, obviously there is a degree of self selection. And if you have multiple engineers coming from boot camp you can, of course, choose an engineer you feel would be more successful. That’s OK.
But I couldn’t turn an engineer down because I felt they wouldn’t be a good ‘fit’ on my team: they’d already been through the hiring process. Conversely, it absolutely can and does happen that people go through the specialist interview loops and end up in a completely different role. For example, you could have spent your entire career writing iOS client code and join a backend server team, no problem.
That’s the theory, in practice of course for some people it does turn into a second phase of hiring, which is very much not the intention.
Frontend interviews are already like this from my experience. They have you implement features or functions in JavaScript. I think I got tossed one leetcode supereasy level question.
In fact, I'd say at least some of top tech companies are doing frontend interviews like this. Can confirm with firsthand experience for Facebook, Netflix, Amazon, Uber. Go a tier (or more) lower and it seems it's still just leetcode interviews though, barely any different from a general or backend track.
Kind of different. Using my example of the frontend interview track as an example, they would ask you something like:
"implement a custom event emitter class in JavaScript" or "implement an image carousel component"...
...instead of something like "invert a binary tree".
Plus you'd get 2-3 rounds of these questions in lieu of the leetcode DS&A rounds. The (usually single) system design round would remain the same (albeit being something like implementing a UI component at a high and holistic level).
This is one of those areas where more companies need to realize they're not a FAANG. As Joel has pointed out in the past, you're not getting the cream of the crop as a normal company. Likewise you probably neither need nor even want the top. However what you do need is to plug a few holes here and there.
As a hiring manager I find it much better to think in terms of a skill mesh. I'm always looking to bring the collective whole forward. A candidate can do that by bringing new skills to the table, by being awesome at existing skills, or some mix of all of the above.
Not sure why you think FAANG is getting the cream of the crop. Looking at stuff like Rob Pike's comments about Go's use for the general Google employee base [0], I think there is quite a lot of misalignment between the aura of mystique from the outside and the situation as seen from the inside.
[0]“The key point here is our programmers are Googlers, they’re not researchers. They’re typically, fairly young, fresh out of school, probably learned Java, maybe learned C or C++, probably learned Python. They’re not capable of understanding a brilliant language but we want to use them to build good software. So, the language that we give them has to be easy for them to understand and easy to adopt.”
> Not sure why you think FAANG is getting the cream of the crop.
The best coders can demand the highest salary. If you can afford to pay the highest salary you are in an advanced position position to get the best coders. FAANG can afford to pay the highest salary.
Anecdotal evidence to the contrary does not cancel this mechanism.
Not really. This is one of those things that sounds correct but when you actually sit down and examine the claim not only is there sparse evidence to support it but it's not entirely clear what the claim even means.
That assumes the selection process is effective, but FAANGs are all monopolies so there's very weak selection pressure for them to distinguish candidates correctly. They don't get punished for their mistakes.
I'd argue the people at FAANGs are the people who are best at hitting the requirements of a FAANG. That's still a competition and they'll brain drain from smaller companies, but it creates arbitrage on unrecognised ability and other opportunities may exploit that (e.g. starting your own venture relies much more on genuine ability than where you stand on a FAANG progression matrix).
I was contributing to an open-source FAANG project recently and the oversights they'd made in the codebase were surprising to me. There were very basic errors and mistakes and the people managing the codebase didn't understand what some of the functions were even supposed to do. The API wasn't consistent. Tests were testing behaviour that was outright wrong (and, somehow, passing). Feature prioritization was way off. There were half-hearted implementations of some features that were bad enough that they weren't useful. They weren't talking to other departments working on the same basic area (who are also open-source, and hold the state of the art), so there were important features that weren't ported. The project was small, too (but not unimportant).
I don't have high market value, but I was fixing some of these bugs in minutes. The guys managing that project are obviously a lot better than me overall, but I think it betrays a flaw in the institutional structure. Those weaknesses in the codebase would be cheap to fix, the problem is the FAANG standards are such that they are able to go unrecognised.
> [...] FAANGs are all monopolies so there's very weak selection pressure for them to distinguish candidates correctly. They don't get punished for their mistakes.
But they do. You see that happening every time a big company fails. It just takes a while longer because of the inertia the company has built (for example by having made more good than bad hiring decisions in the past).
I understand that good people fly under the radar all the time. Also I understand that there are startups that certainly have a higher skill level than Google does (anything else would be shocking, considering Googles immense work force).
Not all hiring decisions are good. Maybe most aren't. There's just no reason to believe that Google would be any worse at it than Random Small Company, considering that more money and expertise certainly helps when trying to hire the best talent.
As a community are we still stuck pretending that being the real deal will be better for your career than being able to convince other people you are the real deal?
I’m not sure how you confirm this, but I don’t see how it’s not true. Do you (or other people on this thread) genuinely believe that the best software engineers aren’t demanding the best salaries? Who doesn’t want to make more money given the opportunity?
>Who doesn’t want to make more money given the opportunity?
Tons of people remain at jobs they enjoy/companies they like even if they suspect they could probably get a better offer elsewhere. There are plenty of people elsewhere in the country who aren't interested in moving to Silicon Valley and working for one of the large firms.
More generally there's probably some relation between quality (which is an elusive variable to measure and can be dependent on managers/teams/etc.) and salary--but it's far from a perfect correlation.
The FAANGs, and I figure about a couple dozen other major tech companies paying similar comp levels, now have offices spread out across the country. You won’t make as much in say Austin or Denver, but it will still be dramatically better than other local options.
>The FAANGs, and I figure about a couple dozen other major tech companies paying similar comp levels
I'm skeptical about the couple dozen. There are a bunch of reasons someone may choose not to work for Facebook or Google. But all the anecdotal information I see--including the migration of people from and to jobs--suggests to me that it's hard for most companies to outbid FAANG (and probably a few other smaller companies) in general.
Just go to levels.fyi and see what people are reporting per company and region, and of course what is discussed on blind if you're familiar with that. Maybe not outbid but coming within 10-15%, which is going to be dramatically better than other regional options.
By these other companies, I mean more established ones like Linked In or pre-IPO like Slack/Stripe. There are quite a few of them and they are indeed competing with FAANGs for talent. But the FAANGs are going to have larger national spread.
My point is, if someone lives in a 2nd tier tech hub they should have a handful options to make dramatically more than average from one of these companies; but the interviews are tough and whether or not they want to have a conversation in the first place are gating factors.
When we're discussing high salaries there should probably only be one 'A' in the initialism. Apple famously does not have particularly high salaries for engineers. Even if they're above industry average, they're nowhere near the Facebook/Google level.
If the industry demands weird or suboptimal practices or has strange evaluation criteria then the people making the most will be the people who are best at the game.
I think the popularity of untyped languages is one example of that. Clearly worse, you're removing an entire class of embedded test and nuking your ability to refactor safely, but huge chunks of the industry don't see that as a bad thing. I'm a worse programmer because I've spent so much time in Python.
Sure a big salary is nice but especially Google hasn't done much in the last decade to make it look there is exciting, or even just satisfying, engineering happening there.
Not to mention, people like to work on things with meaning or value.
Google cancels so many things, so much work and code tossed, and with entire projects, the point of entire projects axed, how can anyone there feel there's a point to their work?
Lmao we're not particles in a simulation. Not everybody optimizes for earnings. I optimize for not giving a fuck, so I'll never work at a FAANG. Not everybody wants to trade ten years of their life span for... a modest home in the tri valley and a horrible commute.
The loop variance is very high, so there's a lot of error, but in terms of the overall quality of my co-workers, the FAANG i spent five years at was definitely the best.
I do notice a substantial difference in average quality between them and less successful Bay Area companies, which does suggest that they are doing a reasonably good job at getting pretty good people in the door.
I think passion comes into play here. You've got to have at least some amount of drive beyond "clock in at 9 and leave at 5 sharp to collect a paycheck" to get into these top tech companies.
Even if at the end of the day, it's still a job, I don't think you can get into these companies unless you dedicate a way-above-average amount of time and effort to do so.
Ignore all of the above if you are a natural leetcode wizard; yes I know there are some people like that.
> You've got to have at least some amount of drive beyond "clock in at 9 and leave at 5 sharp to collect a paycheck" to get into these top tech companies.
I can't say how true this is or not, but I know there are plenty that _want_ to get in exactly so they can clock in & out without doing much
Dunno, I got in relatively early without any leet-coding at all. To be fair though, I work in DS and was in the greater sales org, so there was more of a focus on what I could actually do.
That being said, I think that it was that they (used to) hire people with drive, give them autonomy and a choice of what to work on from the large set of hard problems, and watch the money roll in.
It sounds like OP was talking about the same process, although they skipped mentioning bootcamp and only mentioned team selection. They also accounted for specialist pre-allocated roles. My FB experience was different than many because I was hired to a specific team. I was also the first member of that team to go through bootcamp. It was a weird exception because many of the team members came from the Oculus acquisition, and at the time Oculus had somewhat separate culture. Afaik AR/VR still hires for specific teams and roles externally, although roles in that org are also regularly filled by bootcamp grads with the right experience and/or interests
lol this is not how interviewing works at Facebook, at least not as a general practice. I’ve seen fewer unthinking superficial snap judgements elsewhere than at Facebook.
>Instead of “how can we find the smartest people?” think about “how can we find people who will make our team stronger?”
I would have thought that was obvious? Why would you hire someone unless you had a need for their talents?
I guess I'm probably thinking small-business. At the biggest companies, they probably just hire and hope to fit them in because hiring for exact needs doesn't really scale.
My experience is this obvious perspective vacates with any level of moral maze. The hiring processes kind of require it --> getting headcount approval requires defining a role and arguing how it will lead to better outcomes and then finding/interviewing candidates takes months of time. If your priorities shift in that time...well now you have to shoehorn whatever candidates you get into the role definition you gave before even though you know at that moment they'll be working on something different.
And then priorities shift again before they start. Leaving you not knowing what they'll be working on, but also certain you'll need extra hands on something sooner rather than later...
AND SO: "give me a smart generalist that specialized in these 8 technologies that make up the entirety of our product" becomes the only viable hire.
> And then priorities shift again before they start. Leaving you not knowing what they'll be working on, but also certain you'll need extra hands on something sooner rather than later...
> AND SO: "give me a smart generalist that specialized in these 8 technologies that make up the entirety of our product" becomes the only viable hire.
I mean, unless you were going to fire this person when the project you got head count for is finished, that's pretty much what you should have hired for anyways?
I suspect it's a surprisingly easy trap many of us fall into. It's not about hiring someone without a need for their talents, it's that interview processes can amplify team weaknesses if the signal chosen closely aligns with the current team strengths, and as a result is less likely to fill skills and knowledge gaps.
And this is a sort of natural trap, since the skills we're best at judging are things we're already good at. I suspect these biases can happen at large and small companies. It's constantly on my mind that the signal we use for my interviews are too closely aligned with my own preconceptions, especially when it's a problem space I've spent alot of time in, and may be foreign to the candidate.
Your mindset is modeled after a simple selection process. Big companies often have pipelines for selection processes. HR does screening, leader does interview, manager approves.
Each step of the pipeline, the responsible has to optimize for what the team needs. In practice, each one is evaluated by what the layer above thinks. HR optimizes for CVs the leader likes, leader optimizes for profiles the manager likes.
Specially at large corporations, managers are removed from the daily needs and don’t have the context to evaluate if the person is the optimal fit for their future tasks, so they concern themselves with evaluating “raw talent” (obviously in a very subjective and biased way). Same thing happens with HR doing the screening.
So “talent” is defining both at the beginning and the end of the pipeline, it’s no surprise the end result is an inconsistent hiring pipeline.
Along those lines I'm getting shades/hints that people hire to fit a workforce that resembles something like how they think their software stack works?
Sounds good in theory but in practice people aren't actually programs/cogs and stacks are messy?
Another aspect is making it impossible to assemble parts any way but the correct way. For example, you can write C code like this:
for (i = 0; i < 10; ++i);
do_something();
A colleague of mine, a very good programmer, was stymied by this for a full day. The next day I added a warning to my C compiler for it. As time went on, this warning became commonplace in other compilers.
But in D, I didn't make it a warning. I made in an error. A ; cannot be used to create an empty statement, that can only be done with { }. I still can't believe C/C++ have never made that an error.
Another one is:
if (a < b < c) ...
That doesn't do what one thinks it does. So in D it's an error.
D is better because so many unnecessary things in C and C++ are minefields, and instead of warning the user, they're just illegal. (All have ways of doing the equivalent if one really needs to.)
Oh it took a while until I noticed the extra ';' :- ) It'd likely have slipped past code review o.O
> impossible to assemble parts any way but the correct way
Hmm make me think about database constraints & foreign keys :- )
> As time went on, this warning became commonplace in other compilers
Nice that different languages can help each other become better :- )
> D is better because so many unnecessary things in C and C++ are minefields
I remember long ago when coding C++, we had to add a bunch of macros in each C++ class, to remove dangerous-by-default C++ auto generated things (like the copy constructor that copied pointers). I guess you know a lot about such things. D and Rust seems nice :- )
Hmm there's something similar in the SRS book by Google, they call it "failure domains" (I haven't read all of it though).
> Monitors: If the output is outside some preset bounds, the system is shut down
Maybe in software, becoming read-only can be a similar good idea, when something looks weird
> Deadman: A deadman is a hardware timer switch added to a computer system that shuts it down if it isn’t regularly reset.
This is something I'm planning to add to the software I'm developing :- )
It's forum / blog-comments software, and, in case the admins have been away for too long (maybe vacation for some weeks), the forum would become read-only, maybe even retroactively hide some risky comments & discussions, until they're back — so there's always humans around that can remove toxic troll comments and such things.
> Improving the quality of that component by a factor of 10 will get us there, but at a cost explosion of 10 times the price. But suppose we add in a backup component B, that also has a 10% failure rate. The odds of A and B both simultaneously failing are 10% of 10%, or 1%. This is achieved by a mere doubling of the cost instead of an order of magnitude increase
I think it's interesting that this at the same time, doubles the attack surface, for hackers? Although the failure risk gets down to 1%, now the hackers can try to break in into both A and B? Hmm. I wonder if there're any ways to avoid this tradeoff, ... Maybe there aren't, in the same way as it's going to be 2 x expensive, too
You only need to read a few dozen job adverts to see that tech companies are focused entirely on technical ability and don't hire for other skills. How often do you see "must know Typescript, expert in Docker, passionate about algorithms" compared to "must understand writing documentation, expert in sprint planning, passionate about user value"?
Skills that aren't tech are rarely valued by tech companies despite the fact companies need developers with those skills in order to deliver the best tech. Sometimes the best hire for the team is the second best developer but the best person-who-can-explain-things or person-who-can-run-meetings if explaining things or running meetings is holding the team back.
Lmao, this is such a vague and useless statement that it is barely worth commenting.
What the hell is "passion", how do quantify it and how do you use it? What happens when person you hired for "passion" loses it?
> Sometimes the best hire for the team is the second best developer but the best person-who-can-explain-things or person-who-can-run-meetings if explaining things or running meetings is holding the team back.
You test for that during interview. Do you think algo challenge, system design are just writing some stuff on the whiteboard and you're done? You need to explain stuff in such a way that interview wants to hire you.
Sprint planning is a collaboration between dev and product. If your devs aren't good at it they'll agree to far too much stuff in a sprint, or they won't spot what things are blocked, or they'll fail to report what happened in a sprint in the retrospective, which impacts everything from planning the next sprint to writing release notes to when the company can do marketing about a major release.
Hiring developers who are amazing at algorithms but terrible at everything else has huge, far-reaching consequences for everything a business does. Being a good developer is about so much more than writing good code.
> If your devs aren't good at it they'll agree to far too much stuff in a sprint, or they won't spot what things are blocked, or they'll fail to report what happened in a sprint in the retrospective, which impacts everything from planning the next sprint to writing release notes to when the company can do marketing about a major release.
You're describing common sense, not sprint planning.
> Hiring developers who are amazing at algorithms but terrible at everything else has huge
I like this argument from every anti-algo advocate. Because if they're good at algos they're certainly bad at everything else, like you need to sacrifice one to get another.
> Being a good developer is about so much more than writing good code.
Indeed, writing good code is a basis. You can't be good without writing good code, and you can't write good code without knowing basics of algorithmic design and thinking.
> What the hell is "passion", how do quantify it and how do you use it? What happens when person you hired for "passion" loses it?
It's harder to quantify than 'expert in typescript' and maybe ill described and abused, but there are a diverse set of needs when it comes to engineering, it's not just deeply technical, it's also broadly technical with strong planning, entrepreneurial thought process etc.. my, perhaps naive view is the world definitely needs both of these so keep and open mind and stay respectful.. 'lmao' I'm certain you wouldn't do in person, debate is good but cmon, always put the extra effort in to understand and communicate your own gut reactions
Is this really true? In my (limited) experience, the best coders are the ones who understand the technology, understand the hardware, and understand good program structure. "Communication" and "teamwork" are totally unimportant and they're usually organic side effects of understanding the domain. I've done open-source stuff with a total asshole and it wasn't pleasant but the end product was good because he was good, and I learned way faster than with most people.
But I've not been exposed to that kind of corporate environment. Is it really that important? It reeks of corporate buzzwords to me.
> that was obvious? Why would you hire someone unless you had a need for their talents?
To me, that sounds like you think of weakness in terms of "Oh, our UI sucks, let's hire a javascript developer." But is that really what the article means? Could weaknesses be more abstract things like "low trust", "failure to speak up", "inability to change according to market needs"? Do you know many teams that really know their core problems and actually try to solve them?
If you’re hiring junior engineers, you can’t pick them for talents because they don’t have any. You hire them expecting to be negative for a little bit and learn on the job. If you don’t do this you’ll run out of senior engineers.
On the other hand, if you can work with them it’s good to get people who will contribute things outside the job description.
I had an interesting call once asking me to interview for marketing job at a women's online fashion business.
I said to the person "not sure IM the right person, I'm male and have zero interest in fashion". To which she replied "That's great because I've got plenty of that fit, I want someone that looks at this more as a business than personal interest"
I never interviewed as they salary was low, but I really liked that approach. I find that 'must love and be passionate about our company products' a little overdone, especially in the US. Its healthy to have people with different points of view, and staff that use alternatives and can discuss why.
That's rare in my experience. It seems every startup wants you to think their area is the most exciting in the world and you are full of ideas of how to move the product forward. It's like wanting detailed specs means you are unmotivated, just because you're not interested in making decisions in that space, whether to use one color over another or think ahead about the behavior of the feature, for development it is technically irrelevant, as it's all just translating to code, but business people will have none of that attitude lol.
Being "passionate about the business" is overmuch, but for the employer, it's important that candidate respects the business, in the sense that the candidate believes that what the business is doing matters (quality product that will make the world better, boss recognizes good work and rewards it, customers deserve to be fleeced, or what ever gets you out of bed), so that you want to accomplish progress at work and not just do the minimum for the paycheck.
Two personal counterexamples, whether right or wrong:
Management consulting
- you want to be perceived as hiring the best
- project work that is not known in advance, you need very versatile people
- looking for demonstrated desire to jump through the hoops and do the marginal work that makes you the "top"
Software Engineering
- some overlap with the reasons above
- adds to the moat: if they're happy working for FAANG et al, competitors are priced out of getting the best
- the economics of many technologies are now defined by whether they can return enough to pay people competitively vs big tech. E.g. there are lots of cool deep learning applications that are tougher to try when you need to pay 300k / year for someone really good to run with them
- I'm interested in the sibling post referencing moneyball, this would be very cool to see applied to software engineers
a) You don't need really good unless you're doing cutting edge work at massive scales. There's lots of deep learning talent floating around that is good but not aligned with the FAANG interview process.
b) If you interview process is a clone of FAANG then, well, you're not going to have a good time. See previous point and also you probably don't need a FAANG interview process unless you're at FAANG scale.
c) If the company has other issues then, again, you're not going to have a good time but that's not due to how much you're paying but everything else.
FAANG doesn’t have a single interview process - it’s pretty different at Apple Netflix Google as far as I know. Are there shared features they’re thought to have?
When I interviewed at all the local big tech companies the interviews were fairly similar down to asking identical questions sometimes. Very heavy on leetcode, some system design and a single behavioral interview. Generally looking for the same things in each interview across companies.
They weren't identical but much much closer to each other than the interviews I got from startups.
> Management consulting
- you want to be perceived as hiring the best
This is not really a problem when all four of the companies in your industry use the same short list of universities for hiring. The reality would be even less of a problem with a broader selection of universities but the system works fine for the companies involved.
> Software Engineering
...
- the economics of many technologies are now defined by whether they can return enough to pay people competitively vs big tech. E.g. there are lots of cool deep learning applications that are tougher to try when you need to pay 300k / year for someone really good to run with them
Why the hell are you insisting on someone really good to even start? The pool of people in any given domain with even hobbyist level expertise in ML/DL/AI and some domain knowledge is not usually shallow but non-existent.
That is not an accurate description of management consulting companies hiring process. If you wants to know more look at Lauren Rivera‘s research. It’s about practically nothing else except professional firm’s hiring practices. It is not based on “top schools” as defined academically. and It is based on a certain set of two schools which are very good, and having at least reasonable grades. So what they look for is a certain minimum of grades and class markers. Getting a First will not get you much of anything over a 2:1 in hiring undergrads. Better to be on the crew team or fencing.
Hiring as Cultural Matching: The Case of Elite Professional Service Firms
Lauren A. Rivera
This article presents culture as a vehicle of labor market sorting. Providing a case study of hiring in elite professional service firms, I investigate the often suggested but heretofore empirically unexamined hypothesis that cultural similarities between employers and job candidates matter for employers’ hiring decisions. Drawing from 120 interviews with employers as well as participant observation of a hiring committee, I argue that hiring is more than just a process of skills sorting; it is also a process of cultural matching between candidates, evaluators, and firms. Employers sought candidates who were not only competent but also culturally similar to themselves in terms of leisure pursuits, experiences, and self-presentation styles. Concerns about shared culture were highly salient to employers and often outweighed concerns about absolute productivity. I unpack the interpersonal processes through which cultural similarities affected candidate evaluation in elite firms and provide the first empirical demonstration that shared culture—particularly in the form of lifestyle markers—matters for employer hiring. I conclude by discussing the implications for scholarship on culture, inequality, and labor markets.
> Drawing on scores of in-depth interviews as well as firsthand observation of hiring practices at some of America’s most prestigious firms, Lauren Rivera shows how, at every step of the hiring process, the ways that employers define and evaluate merit are strongly skewed to favor job applicants from economically privileged backgrounds. She reveals how decision makers draw from ideas about talent—what it is, what best signals it, and who does (and does not) have it—that are deeply rooted in social class. Displaying the “right stuff” that elite employers are looking for entails considerable amounts of economic, social, and cultural resources on the part of the applicants and their parents.
You can get a lot better than that, and you actually reference it in a later comment -- you see how well they perform on an actual task. It's true that high standardized test scores and IQ test scores often come along with high performers, but it's at best a necessary but not sufficient condition. That's like saying the only thing you need to make great wine is good grapes. Plenty of awful wine has been made from excellent grapes.
It's well considered by a lot of people in the industry that a prospective's grades and even which school they went to aren't a good predictor of skill.
Is the school thing a strictly American phenomenon? In New Zealand for instance all that tends to matter is which city/town you want to live in and whether the institution offers the field you want to major in (medicine for example is only offered by Otago IIRC).
I've only been asked for grades once in the previous 12 years where at which point I withdrew my application.
Yes, America has more school inequality than other countries (although it’s not the only one, Oxford and Tokyo U also exist). Part of this is because they’re so good they suppress top international schools from forming in other countries. But also they teach in English and I assume it’s harder for international students to attend a French or Chinese school.
In some cases, school is a legal way to discriminate. The top tier companies incorporate recruiting into their culture.
If you recruit at Brown, you’re going to get a different candidate than some CUNY school. They all tested well in high school or have parents with good networks, and they tend to look and quack alike.
"Culture fit" is just another way to discriminate a candidate based on race, age, or gender.
HR from my company has explicitly told all interviewers not to use "culture fit" as a reason to disqualify a candidate presumably to avoid potential lawsuits.
The best programmers I know didn't go to college. It's been my experience that degrees and standardized test don't tell you jack about how well someone will do a job.
for startups, new grads with good portfolios and ok grades > new grads with good grades and ok portfolios
likewise, new grads with a story of upward trajectory & energy > new grads who coast on what their parents/school setup
the bet being made is they'll put in the hard work and achieve the personal growth to progress quickly, and they'll roll with the uncertainties of startups. grades do hint at the hard work, and youth suggests dealing with uncertainty, but not as well as succeeding at difficult and fuzzily-defined projects. if they've done that before, better chance they'll do it again.
outside of startups, priorities change. big orgs need people who won't risk the cash cow. consultancies need people with better credentials than their customers. etc.
Different people have different ways of learning and working. Some people build a bunch of stuff. Others build toys, others may just focus on their primary jobs and don't have a portfolio of things they can share.
most students don't work on classified fbi projects, see in the thread on some of the projects I've seen undergrads involved in that are more demonstrative than grades. a project portfolio is just a cv section that focuses on the projects and your involvement, and hopefully ways for the interviewer to see it and the code, but even descriptions and screenshots go far for the interview conversation.
most good cs programs have significant project components, and students not putting in their project time there are often doing so elsewhere.
if not, they may have the potential to be good at projects, but tbd. without other strong signals, probably fewer startups are in a good position to take that bet.
also, going beyond students, the top programmers I've worked with have a history of great public and private projects - papers, $100B companies, top frameworks, launched top products, record breaking results, news articles around their impact, etc. The main exceptions are 'lifer at apple' types and 'works at secret bank', but most hop jobs enough that it's not an issue. A top group I see that really struggle a lot are 'lifer at secret gov agency x', which gets many of them stuck in the defense contractor round door.
we may have different experiences for what top means and what they help achieve. Who is this silent majority of top 1% programmers that don't have any referenceable projects? And more to the point, top intern candidates?
if they don't have impressive projects.. maybe not so good?
when i think of the best cs ugrads i've known, and would be great at startups:
* cool projects outside of classes, especially self-initiated, like doing their own startups
* cool libraries or algs that got popular or, not quite as good, published
* hardcore mode attempts at class projects, e.g., well beyond extra mile for OS's + graphics, or took grad-level courses for the same (=> w/ accompanying projecs), or new to CS and are proud at winning some class competitions despite that
>if they don't have impressive projects.. maybe not so good?
lol, you realize that for I'd say almost all projects the interesting part/core is small % of the whole, meanwhile the rest is just boring code that needs to be done?
ofc it's not the nicest thing that you arent completing your side projects and just doing the interesting stuff, but you may have experience in industry, so you know how to do that boring ass stuff anyway
Startup conditions are chaotic and every new person takes management + training overhead. Internship applicants who have well-done and challenging projects have shown the energy, self-management, know-how, & self-learning to get projects done in mini-versions of these conditions. So every hour spent on them goes a lot further, and both sides benefit more.
Conversely, if someone can't get their projects done, the projects are too low quality for where it matters, or their skill are on too-different of a thing, at best, it's non-evidence. So you'll have to fall back to evaluating weaker signals like grades. At worst, lacking or failing at projects is a red flag that they'll struggle on the adult-sized projects as well.
For 'interestingness', probably more relevant is 'tied to the mission', but I think being youthful covers 80% of that for intern-level folks.
Failure to figure out fit may mean everyone loses. Interns are more about recruitment pipeline for companies, and I'm not sure that's the wisest way to solve that for a lot of startups if that's the goal. Likewise, startups not ready to invest in their interns (management overhead, training, etc.), or fail to screen for a good match in skills, can easily be a bad experience and opportunity cost for the students, who may have had a more pleasant, productive, and rewarding experience at a more mature company.
I'll speak from perspective of average CS in E. Europe
people who were the best at my year were mostly focused on jobs / real world software development and just wanted to get degree, so just pass the exams and projects effortlessly
but I've been studying at weekends, so I may be biased.
These phrases are mainly intended to flatter the employees, to make them grateful for being "chosen", and maybe to increase trust and confidence among them.
Choosing who to hire (once basic requirements are met), is much less important than the morale and attitudes among employees. Much better to say "top talent" than "randomly picked".
But, the hiring process is a huge factor in the perception of the identity of the employees.
To me, much more important in that phrase is expectation who will you be working with. It's fantastic to be surrounded by smarter* people and keep learning, but at some level it's hard and if you like what you do, then this seems to be the most important thing about the job.
You have your basic financial needs covered when you're a decent coder, even boring projects can have great and interesting implementations, it all drops down to people who you are working with.
Also, not just in IT, true masters tend to be interesting people in my experience. More often than not, they have humility and curiosity that allowed them to get that far.
But of course it often tends to be an empty phrase. If they do hire top talent you are more likely to hear specific names rather than this phrase.
1. smarter / more experienced / more knowledgeable / sometimes even having your level of expertise but coming from a different background
Knowing to hire for weakness instead of "top talent" implies a level of reflection almost no medium+ sized company actually has. Most companies don't need top talent, and the talent they have isn't anywhere close to the top.
Exactly. Some traits are always good, some always bad and yet others depend on context.
For example I currently work with a team with a lot of very hard working technical people but not enough concern for questions like "what's valuable for us to work on?". So I need to hire people who will ask those questions.
On the other hand, if my team was full of big picture pontificators, I'd be looking for someone who's shut up and code for a change.
This is also great personal advice. Every month or two look around and see what is missing on your team. Do you need more shared vision? More code? Does your team lack expertise in some area? Do you need more support and acknowledgement from upper management? Are bug reports piling up?
this whole thing reads: don't hire top talent, hire special talent, to us, whatever. The latter sounds even harder than the former. It's hard to see your own weakness, it's rarer to have people to fill an exact particular gap at that particular point in time.
Maybe do away with the mindset of people as off-the-shelf products and think about them more as raw resources and focus more on upskilling. Universities are hopeless on this so the responsibility must fall on to someone.
Good mindset for big corp, not so for small biz. Waiting for someone to upskill and gain momentum for 6-12 months means a lost lump sum of money that can kill your business.
Hiring a candidate who already knows what you need often takes a lot longer than hiring someone who can learn what you need even if you account for the on-boarding time, especially if it's a slightly esoteric thing.
Difficulty in hiring people who know your stack is also a good reason to choose the simplest and most common tech to run on rather than something niche you feel is technically better.
I guess it's something to balance. There is no reason to believe that the special talent will show up in 6-12 months nor that they will find your small biz a good place for them. I have a hunch that many of the people who can fix the weaknesses for your org are gonna be top talent as well with lots of choices
Amazes me how little data-driven recruiting is. Just as the author puts it, a good hire is for culture ad, and that means being explicitly aware of what's missing. Tough stuff to really look inside and see what's missing.
Most hires I see are either out of desperation or because the recruiter had a good chemistry with the candidate. A 1h interview means 0,3% of the time I will work with that person in one year. Talk about a small sample size.
I've both taken and been at a company that administered personality tests as part of the interview process. I found the results fairly useless to be honest. Although I did manage to score so low on one part of the assessment that an interviewer seemed to question my basic ability to function in society.
Personality tests are broken - there is no way a functioning human being will answer them truthfully. Not for the lack of trying... our subconscious will try to protect us and we'll answer in a way we imagine it is socially acceptable.
While I have little doubt they're indeed broken in the way you describe, as long as the results are consistent they could still be useful to hire for diversity.
You would care more about your team members having a wide variety of results, rather than the qualitative interpretation of the results themselves.
@johnsmith4739 You've arrived at home after a long day at the office. As you take your wallet & keys, you realize that you've accidentally taken home a granola bar that you obtained from the snack room at work. You had intended to eat it at your desk, but for some reason you forgot. On a scale of 1 to 10, how guilty do you feel for having unlawfully stolen property from your employer?
I'm a founder of a rectech startup. I see four big problems in most companies:
1. Turtle speed. The hiring process takes 45 days, and most of the best candidates get hired by someone else, usually before the hiring manager gets around to interviewing the candidates.
2. Arbitrary process. Candidates get a completely different set of questions, completely different interviewers and completely different resume screenings as different recruiters and hiring managers do things completely differently.
3. Overscreening. Screened by the job board (knockout questions). Screened by the AI in the ATS. Screened by the recruiter. Screened by the hiring manager. Then a phone screen. Then a first interview (which is really another screening). Only then can we try to hire the half-of-a-candidate that made it through the process. Most of this screening isn't cultural, it's people trying to read between the lines on a resume to figure out why Lucy left a job six years ago after being there for four months.
4. Assessments applied after steps 1-3 to confirm executive bias instead of early in the process where the data from assessments could actually be used, you know, as data to compare and fit candidates.
Insanity, right? Let me add another one - have you seen how they advertise job openings? With the "technical specifications." It's like they actively try to do a bad job.
We only worked with founders that are involved in the recruiting process. 2, maybe 3 "dates" and you should be able to decide with clarity.
Yeah. Job ads are problematic because they are mostly mix of trying to get bad candidates to self-select out, old-fashioned keyword packing search spam and utter BS compliance disclaimers. It leads to eight pages of garbage, and ironically piles of bad candidates to screen out. Since most recruiters spend 80% of their time screening, it's like a self-fulfilling prophecy.
We get so much more pop out of job ad spend when we limit the job ad to about 150 words and just say what the job is, and what really matters to you. "Software developer with proven track record of building web and mobile apps with React. You must have at least two years of professional development experience and be able to pass a Javascript coding test." Describe pay and benefits (about a 20% boost in applications for including pay). Done.
3. is, I think, one reason personal contacts tend to work well. A top candidate for a position that may not even be actually listed yet can easily end up screened out by someone or something just flipping through a pile of resumes because they are missing some keyword.
Any websites which irritatingly change the page title back and forth to "You have 1 message waiting" (or anything else, really) get immediately added to my blocklist.
The article spends most of its words trying to convince the reader to do it, and only a few paragraphs on how to do it. I was convinced in the first paragraph.
My own experience shows that hiring for your team's weaknesses can make it more effective. In 2013, I went from an operations-heavy role to a new team with little operations experience. Our TL designed and coded a new service. He was not happy with the feedback I gave through the design & code review process. Before deploying the new service to production, we met with one of the company's top engineers (Sanjay Ghemawat) for a "production readiness review". The reviewer was impressed and found only a couple of minor issues. He said that it was rare for him to review a service and find no major issues. Many of the questions he asked during the review were things that I had brought up earlier and convinced our TL to fix. Although I had not written any code on the service, I contributed operational knowledge and filled a weakness of the team.
Our TL did not acknowledge my contribution and even gave me a bad review, costing me about $80k in compensation. Now, years later, I believe that he did that to get me to leave the team. I had often pushed to fix the major problems in our project. I think he wanted to keep control and continue working on fun things and ignore the problems. Our manager went along with him, building his own empire. I left and the person who took my place had the same problem and left as soon as they could. Honesty and respectfulness are important qualities of teams and team members because they develop trust. Trust makes teams more productive. An internal study at Google concluded that teams with members who trust each other are much more productive than teams with high technical ability and low trust [0]. Lack of trust is a team weakness that we can hire to improve. We can also intentionally train and practice skills that improve team trust.
I want a systematic way to identify the strengths and weaknesses of myself, my teammates, and potential teammates. I want a questionnaire like the one in Essential Enneagram which was tested in a study with 1,000 participants [1].
I wish the large tech companies would invest in this kind of work and publish it.
More of a perception change rather than an actual change I think. FWIW, I've always hired to "fill gaps" and not to "add stars". Now, do I want a great person who can fill that gap? Sure. But the priority is the gap, not the person.
There are those mythical people for whom "if you interview they will 'make' a position so they can hire you" is said to be true. In my experience, people who are hired because they seem like super stars rather than for a particular role, do not do as well. Why? Because it usually automatically puts them in conflict with the person who is already doing that role, and second, what they should be doing isn't immediately apparent leading to some meandering around which is perceived by upper management as being 'non-productive.'
And I think the same goes for talent development, too. Career ladders, carelessly applied, incentivize everyone to do all the same things to check the boxes for the next level. This can lead to a lot of mediocre impact.
I would suggest that we really want is well rounded teams, where individuals are playing to their strengths. It takes a more nuanced approach to evaluation, because not everyone is doing the same things in the same way.
This seems like an oversimplification of what it takes to make a company grow and succeed. Is it either a) hire top talent or b) hire for weakness.
The question this article is posing is the wrong one, it’s neither a nor b. Companies need to hire based on what they need most, sometimes it’s profit, sometimes value, sometimes all of the above.
Moneyball's point was subtly different than this, and I was hoping the article would actually touch on it. The article suggests hiring the best candidates that have skills that mesh well on the team; Moneyball explicitly searches for people whose flaws can be overlooked. This way, you get a team member whose performance outpaces the open-market salary.
In other words, you don't hire candidates whose strengths match your weaknesses, you hire candidates whose weaknesses you care about less than the competition does. Like, the best example from the movie was the pitcher with an effective but highly unconventional delivery mechanism.
If you want a competitive advantage and having a better team than the competition helps with that, then you are looking for a team that is better than the competition's. This is what you hire for. If the competition is very good, you need to perform better - that does not mean every member of your team needs to be better than every member of the competition's team, except when it has to. For example, take professional tennis players: they need top trainers, top physicians and top managers to be #1. Hiring someone that is not experienced just because they can ask the right questions... will not work. Even asking the right questions require very good knowledge, otherwise they will ask random questions. You don't want to fly in a plane built by someone without experience, but asking smart questions.
Except that business is not like sports. There is no "level playing field" that matches teams 1 for 1 against their counterparts.
In many cases most parts of the company just have to be "good enough". If you have a great marketing team and 90% of the market already then you don't need a world beating tech team.
Even when you are competing against equal and similar companies there are areas there are points of differentiation. One Company's product might be cheaper a easy to setup, Another is good at Enterprise Sales, a third has great tech and features.
There can be luck, too. Billy Joe BigAccount decides to give your company a massive contract because his daughter's diaper wasn't too dirty that morning.
I recently discovered that, in a topgrading-style interview, it’s possible to filibuster!
They expect to be able to ask 4+ questions about every job from the past ten years. But I will happily ramble for 20 minutes about minutia for each of them. It was wonderful
From a single interviewers perspective: this does not work. At the company I work at, and in my own interviews, being succinct and articulate in the points you’re making goes a long way. Ending early is not a bad sign.
The article assumes that most teams know and openly talk about what their weaknesses are. But is that actually a common thing? Don't we have a natural drive to ignore our weaknesses and blame them on others?
It's a little funny, because the alternative phrasing also includes that already.
One hires the best of the best without specifying the best of <what>.
"The best people will solve all the problems!" without any inclination about what the problems are.
Doesn't that imply that one doesn't actually know one's own problems and hopes that if one can just hire a people smarter than oneself, they will figure it out for you?
I think that managers should have to take the GMAT as part of the hiring process, even if they got their MBA 20 years ago. They need to know number properties and have basic critical reasoning skills for the roles that they are interviewing for. Why give them a free pass while requiring principal engineers with 20 years experience to answer leetcode problems? The low quality management at large organizations is an obvious sign that they're not screening for the best talent, who clearly scores at least 700 on the GMAT.
First we must ask the question: how can we demoralize people the most effectively? An obvious place to start is to just ignore all their prior work. Since we're starting from scratch we need a baseline to gather where their technical skills are at. So you should ask them to do some kind of coding assessment task. Tell them not to spend more than an hour on it. Chances are good the most desperate will spend way more on it because they need the job. If you're lucky they'll even do it without asking for payment.
Spice it up a little! Why not throw in some real work that no one in your company wants to do? You can even ask the applicant to read your organizations repo code and write technical reports on it. That way, later if you do decide to hire them (unlikely) it will reduce your on-boarding costs... and if not... then fuck it, tbh. This isn't about making people feel good. It's about making sure you do what everyone else does. 'We've found that the assessment task is a good signal.'
Speaking of signals, how about you ignore the whole resume and focus on what matters. We both know resumes are largely a waste of time and people only get moved up in interviews for making plenty of eye contact. So make sure to fail everyone who doesn't stare at you the whole time. This is how to assess technical skills the most accurate way possible and its known the vast majority of managers fail people in interviews for not staring at them enough (staring increases productivity.)
Finally, you don't want to forget the computer science tests. At least four years in school and tons of debt is a lot to stomach compared to people who learned practical skills in their own time at a cost of almost nothing. Though chances are almost zero you use computer science knowledge in your job it's best to make sure you fail every self-taught developer who can't immediately balance a binary dick tree. It shouldn't concern you how many developers are self-taught and won't fit in your talent pipe line after doing this. Your hour long algorithm interviews are sure to make sure that only people who want to feel smart get hired.
If you put these measures into place you're sure to build the most mediocre organization possible. But there are other ways to improve the process. Experiment with calling your company 'remote' but only hire from certain timezones; Make sure the interviews happen in person. Hold as many meetings as possible. If there's one thing engineers love it's meetings that break up focused concentration time in their day. Even better if the meetings are only done at certain times in the day as that will be as hostile as possible to a remote working environment.
I want to make sure you don't have any funny ideas in your head about remote working. It might seem to offer flexible time management but people are ultimately children and its up to you to put in place measures to prevent this. If you do this right you can even manage to exclude people with chronic illnesses that would have otherwise been able to fit in your talent pipeline through remote work which I think is really something.
I hope this post has been helpful to other recruiters and hiring managers out there. Peace
That is what most companies do. When you have an over supply of backend engineers but need frontend engineers no one looks for more backend engineers. Their stack has a weakness of frontend engineers and they find the one that strengthens it the most.
I don't think that's how impostor syndrome works? By definition, someone with impostor syndrome is a high performer but they attribute their success to luck, thus they think they are a fraud.
You're probably bring downvoted for the attitude, but you're right.
I've worked for a couple of low-mid tier companies and their core was always mediocre talent who is there because it's the top of the world for them. Top talent will always gravitate towards money, prestige. Sure there are some outliers that will work for food and change world, but if you can earn more why wouldn't you?
I was just riffing on a joke, but much as I love Hacker News, it's a bad place for that. Too many people who take everything too seriously.
But yeah, there was a pinch of truth in what I said too.
Thanks to the rate of bitcoin appreciation it doesn't matter where I work. I can make the same or more than the top companies. I can opt out of the crazy hiring process
Talent is everything. Without Elon Musk, or Von Braun there are no key rocket program break throughs.
Not that I know much about rockets, just that there are key talented people that push the world forward. You need the worker bees. But someone at the top has to be the visionary genius.
If was the manager of a soccer team that came in last in the League. One way of improving would be to replace my weakest players. And the team would probably marginally improve. But it doesn't usually happen that way. Chances are your best players are not talented enough either, and you should replace those too. Or Perhaps, trading two of your good players for one more talented. Which is typically what happens.
So not sure if I would agree with the premise of the article.
Also, If I was building a stock portfolio, one not terrible way of doing it would probably be around how many PhDs the companies have on the payroll.
If he succeeds reusable rockets, privatizing the space program, commercial flights into space, starlink, and the tech he develops around his quest to going to Mars.
Elon has a unique combination of business skills and technical skills.
My contribution is in technology development. But even the dumbest ones of us can make a valuable contribution to the world by properly raising the next generation of human beings.
There are also a lot of smaller but specialized longstanding loops like front-end, AI... and one off “reqs” for special roles that teams need.
The world of hiring is very fascinating :)