Oxide Hiring Process
rfd.shared.oxide.computer
rfd.shared.oxide.computer
> There is no guarantee they will read your responses with the same care you put into writing them
Multiple people review every single submission.
> almost certainly they won't, since they have actual work to do an honest reading
This is why there is such lag time in even getting a response; you are correct that it takes a tremendous amount of work, but also that there's a tremendous amount of non-hiring work to do. We are not always perfect at balancing these things, and the queue can grow.
> so at some point they will be forced to change it.
I am in full agreement that as companies grow and change, their processes must also grow and change. That doesn't mean that the initial processes are inappropriate. That's the nature of building something new.
We will see at what point changing this process makes sense. It's worked decently well from the first founders (who themselves underwent this process, with each other) up to our current size of about sixty. And speaking purely for myself, the problem as I see it is that we have to say "no" to some wonderful folks, not that we struggle to find people to apply.
Have you gotten more feedback in other industries? To be honest, I thought that general standard is essentially silence. But it's been many years since I've worked in not-tech.
It does seem a little bit suboptimal to ask so much upfront from people who you would never hire. A company with a similarly large early screening form, Reservoir Labs, does a first resume screen before sending an invitation to respond to their screening questions. I would suggest that you do the same out of politeness to candidates.
Honestly, it feels suspicious.
It’d probably be better to invest the (substantial) work done in reviewing the submission in a more structured interview process where the traits can be better assessed.
Yes you do, implicitly.
This is why I asked if you can detect any interesting signals in your candidate’s profiles that diverge from the overall candidate pool (hard to get stats on the people who don’t apply, but I think you understand where I’m going).
For the record, I’d LOVE to see that kind of data.
If you make the candidacy process a significant burden to the candidate, I’m weary you might make it a similarly significant part of the selection criteria, just because otherwise good candidates who are less eager to work for the company might not have the time to apply to it. That the process rejects a large number of promising candidates is already predicted in the article.
Can you observe any signal on the candidate background that’d indicate they could be consistently eager to leave their present companies?
(All companies should make their hiring process transparent and public).
So you should probably be spending _more_ time on the feedback for rejected candidates. Be specific, tie it back to your evaluation criteria that you wrote down before asking for the exercises to be completed.
I get the sentiment but it’s very tricky; most places do a screen and then an interview and so review takes place during said interview, we are heavier allocated on “the screen” in a sense. It would be even more work in our circumstances to do so, since we have more people in that stage of the process than others.
If it's "too much work" for you, might be time to re-examine how much work you're asking for.
I also think these sorts of things select for candidates that have lots of time for "applying to places" (which will naturally select against people already in jobs, or who already value work/life balance).
Your process doesn't look as onerous as some I've seen for sure -- but perhaps you _should_ add a "screen" portion and thereby not ask for the "long form" work from candidates you'd reject anyway (also buying more time for meaningful feedback to those you do ask the work of).
This puts you in a vulnerable position to attacks with generative AI. Assume a person using such a tool could generate a relatively large number of credible submissions in addition to their own as a way to limit your ability to assess possibly real competition.
Nothing is guaranteed in hiring. Demanding that another person “waste time” sitting in front of you before you’ll answer questions is strange. Wouldn’t you rather answer the questions on your own time without someone else putting you on the spot?
It’s not very rational to avoid applying to a company just because you think they might not read what you wrote.
> Until then, I think the better approach if you're interested in a place like this is to be so awesome they have to call you first.
So answering a few questions is too much work, but becoming so “awesome” that you’re top 0.01% of industry talent who maintains a public profile so prevalent that they seek you out is the better way to apply?
Applying to a company and investing a few hours interviewing is accessible to everyone. Becoming a high-visibility, industry leader with active public profiles that attract companies is very, very hard.
No, because if that's the process, for all you know they're requesting answers from 1000 people for one position. With synchronous interviews, you have reasonable confidence that there's a limited number of people they're going through this process with, that you have a decent chance at the position.
Why would any company engage with 1000 candidates and request these samples from them with no expectation of reading them?
I don’t understand this desire to assume the absolute worst, even when it’s illogical for the company to do so.
Because there's little downside to do doing so. And of course they would read some of them, they do need to hire someone, but they might not read all.
> I don’t understand this desire to assume the absolute worst, even when it’s illogical for the company to do so.
It's not necessarily illogical. They can just blast out "give us samples" to a large number of people, and then filter out most of them based on whatever they feel like later on.
It also means that later on they can brag about they're so elite that "they only hire 0.1% of the people they interview", if the company culture bends that way.
Honestly no, as you might spend loads of time crafting the perfect response, only to find that they hadn't really worded the question in a way that meant your answered what they were really thinking. This happens often enough in in-person interviews as it is, but atleast you can hope to course-correct by looking at how they respond. The lack of time-bounding is a real problem for those who have genuine time-commitments outside of work
Also, if "everyone" has a 1 hour interview (or whatever) then you're more-or-less comparing equal things. If some people use a week to craft their "asynchronous" responses while some have an hour to do so (because they're already employed, say, or have a family) then you're not comparing apples to apples.
Similar for "take home code exercises": it's impossible to compare them meaningfully because you have no idea if the better-looking one took the candidate an hour or several days (and yes, I've been on both sides of this).
That's definitely the better strategy. Or networking can replace having the public profile. That being said if you haven't gotten there yet, it might make sense to do all this work anyway because that puts you in the 5% of candidates willing to jump through all these hoops to apply, but you'll definitely have a lower chance of getting in as just some rando applying on the internet than someone they already know.
I wouldn’t consider myself anything near a high-profile candidate for anything, but I routinely turn down tempting offers from interesting companies.
The positive side is that I have lunch with a lot of cool people and, sometimes, even get the chance to show them my current city a little bit.
I liked the Amazon, Google hiring process circa 2015. Come in for a day after passing some basic loops and see if we can work with each other.
Seems like a lot of work relative to just applying to jobs.
I am still kinda salty about it; I was proud of the writing and presentation samples I sent in, and was hopeful I was a good fit.
One anecdatum, but which might not be that unusual:
When I applied (while working a very time-consuming and stressful key role), I had a small amount of time to catch my breath and context switch, that I could passably focus on thinking about other companies or interviewing.
So I decided that I had to choose between applying to Oxide and spending the same time prepping for a certain FAANG's Stanford fratbro shibboleth hazing rituals. I decided to go with my heart, and with the opportunity that seemed more inspiring, rather than a bit jerky.
After suffering through the "tenure packet" assembly headache, I waited, and waited, and then got a rejection email, without even a screening call. And a key sentence in the rejection email actually came across as snide, given that the writer seemed strong verbal. Which made me doubly regret not spending the same time practicing the FAANG's Leetcode whiteboard performance art.
(I ended up taking a comparable principal engineer job at an even more exotic hardware&software company, where they could hire people just by having very experienced other engineers, managers, and execs talk with them, no nonsense.)
The bad experience with the bespoke tenure packet also biased my decisions on my current search. One very interesting place I looked at, the extensive screening with a very knowledgeable person went exceedingly well, but when they presented me with the hiring group's systems programming take-home (huge time investment, with a dozen cleverly-spec'ed hard problems to nail, basically solving most of their hard software systems problems), and I didn't have time for that, I demurred.
I also now have even less tolerance for the Leetcode cargo-culting, and any other obnoxiousness, while having more appreciation for wise and intuitive hiring teams.
One factor, put loosely and jokingly, but there's a lot of truth to it: "If you like me, you must be good people."
Lol, Do you have a blog? I would subscribe to your substack.
What work have you done that you are particularly proud of and why?
When have you been happiest in your professional career and why?
When have you been unhappiest in your professional career and why?
For one of Oxide’s values, describe an example of how it was reflected in a particular body of your work.
For one of Oxide’s values, describe an example of how it was violated in your organization or work."
I feel like these are a test of creative writing skills, storytelling, believable exageration, and imagination more than their stated purpose.
> "In other cases, it’s even more nuanced: there have been many later-in-life converts to the beauty and joy of computers, and such candidates should emphatically not be excluded merely because they discovered their calling later than others. For those that concentrated in entirely non-technical disciplines, further probing will likely be required, with greater emphasis on their technical artifacts."
I love this section!
The stated purpose being to assess shared values:
> Values
> Values are often not evaluated formally at all in the hiring process — yet shared values are essential for long-term success. Some questions should be asked to suss out a candidate’s values, e.g.: [the questions]
I suppose that a candidate that fundamentally doesn’t have shared values could B.S. their way through this part, but I’m struggling to understand why they would.
Normally, I do ask these story telling questions. Tell me a story, then I will deep dive into them, dissect them, drill into the technology, relationships, communications, budgets, time frames and more.
This is in lieu of me coming up with a set of rote questions that can answered in a sentence or two, then move to next topic.
If the candidate is just telling me "believable exageration [sp], and imagination", it will come out real quick.
If companies ask for longform answers, people get mad.
If companies ask for code examples or public repos, people get mad.
What I learned over a decade of running a company is that everyone gets mad about hiring practices, just do whatever you think is best and adjust over time. And stop taking advice from people since the Internet is full of many factions whose slices make up 100% of criticism over hiring practices, so it's impossible to get any signal out of it.
True about many other things. Advice is largely useless. But especially for hiring.
https://chat.openai.com/share/1607e7d6-ecd4-4024-bf25-9bd8bf...
There are some hiring practices that really suck, of course. So probably don't do those, like choosing people randomly or based on illegal discrimination. But beyond that, just do whatever and iterate.
1. The expectations and details in the job were specific to the actual exact role I would be filling and were clearly laid out in the job posting.
2. The "interview" was really just a single long conversation on those specific details and a discussion over whether I fit them and can show/discuss examples of how I fit them.
A lot of hiring nonsense is because the actual job requirements and posting is so generic as to be practically meaningless, so you create these pointless hoops to jump through in order weed out an individual from a pool of other generic individuals. The actual employee you pick ends up simply being really good at jumping through hoops rather than a good hire for the specific role. Crazy multi-step hiring processes just seem like selecting for endurance over any ability as a future employee.
[1] One example: https://home.ubalt.edu/tmitch/645/session%204/Schmidt%20&%20...
1. IQ tests work real good.
2. Structured interviews -- which are not IQ tests -- work about as well and strongly correlated.
3. Work sample tests -- which are _definitely not_ IQ tests -- work as well and in ways NOT correlated.
Meaning you can combine these two and get a reasonably effective hiring screen.
It's also pretty keen to recommend Integrity tests which I think the authors play a role in studying.
edit: I guess it's "operational validity" which maybe isn't "percent that did well later"?
If I wanted this I’d have stayed in academics, with its 20-page CVs and multiple statements of purpose.
At least there if you win the tenure-track position lottery, and then the grant lottery, you have a good shot at winning the tenure lottery and getting some job security doing something you (probably) love.
Oxide is a startup with a finite runway.
Hard pass.
Many people stay at a company for longer then they stay at college. So that doesn't seem that crazy.
I have nothing but the strongest distaste for LeetCode interviews but this is another flavor of absurd that still takes a huge time commitment from the interviewee which inherently discriminates against those already employed, those later in life, and those interviewing at more than one company.
More evidence for my big binder of evidence that interviewing culture in our industry is entirely some grotesque form of hazing ritual that builds upon itself, with each company and team deciding that the specific form of pain that they had to go through is the ideal form to subject candidates to in order to justify the pain inflicted on them.
As you can tell I take a dim view of most of our industry practices around hiring and it's one of those things that can rile me up. I think the vast majority of interview structures are heavily weighted toward favoring fresh grads (more time to dedicate to prep, fresher knowledge of the type of rote CS stuff that you learn in college but rarely used in industry, more willingness to put up with onerous interview requirements). I think it's likely because less experienced employees tend to be easier for companies to exploit, and so companies allow it to continue by feeding into my above-mentioned theory of hazing. Interview practices like we have impact people in negative ways small and large, yet I have seen no evidence that they produce better results than more humane methods.
I hope that as Oxide grows and succeeds (and I do wish y'all the best!) that you're able to either streamline your interview process to be more accessible and/or offer alternate paths.
Given the level of detail requested here, in writing, this feels like it’s requiring proprietary information from a previous job.
I don’t think many people do “analysis [...] of a system behaving pathologically,” or “perform[] [analysis] of a [sales] prospect and their fit with the product before reaching out to them” on their own.
> I don’t think many people do [snip] on their own.
It is true that many people need to do new work as part of the application process. I wrote a large portion of my packet from scratch.
I hope you realize as well that approximately everybody is the former sort of candidate.
———
We are so humbled by your application to join Oxide Computer Company. At this stage of the company we are hyper-focused on certain areas of the stack and when we need specific domain space experience such as yours, please engage with us. Our roles will be updated as we need them.
We are grateful you took the time to apply and put so much thought into the candidate materials, we loved reading them. We would absolutely love to work with you in the future and cannot wait for that stage of the company!
All the best, The Oxide Team
> Interviews shouldn’t necessarily take one form; interviewers should feel free to take a variety of styles and approaches
I guess we still don't know.
> Author: Bryan Cantrill
> the act of writing — like so much else that we do — requires not only the ability to create wholly new material, but also the ability to reflect, correct, revise, and polish.
Oh Bryan, you could have spent your time doing some actual work instead of being pretentious to the tune of 4700 words and still not saying anything useful to anyone.
A bad article is bad regardless of what the author has achieved during his career as a whole.
This style of writing gets in the way of actually communicating. It's harder for me to see what the actual process is.
Though, maybe they want to filter for people who enjoy writing this way, and to filter out anyone who wants more straightforward comms. In that case, it'd be low key brilliant.
"They must be able to build castles of imagination, and yet still understand the constraints of a grimy reality: they must be arrogant enough to see the world as it isn’t, but humble enough to accept the world as it is. Each of these is a balance, and for each, long-practicing engineers will cite colleagues who have been ineffective because they have erred too greatly on one side or another."
Ooof... where are you going to find these paragons of balance? It almost sounds like they're searching in Zen temples or something.
This reminds me of the impractical and absurd hiring process of Canonical.²
No regrets on taking a pay cut, personally. It is true that I could make a larger salary elsewhere; I also would not be working on this product with this team.
They're still basically a startup / pre-stable revenue, and most folks at startups are not risk-averse. So I'm not sure this bias really affects them at this stage of the company; even if they didn't have this hiring policy the people who want to work at startups have the same characteristics.
> 9+ hours of interview process for a salary of $201,227 USD. Isn't this bit too much?¹
And we have tons of folks who are not speakers or are even active on social media. Being well known is not an important factor in being hired, having skills that fit with what we need is the most important thing.
If it's 9 hours in one day that's a lot, but if it's over 3 or 4 days it seems reasonable to me.
I will say the worse interview I ever done was for Capital One. It's literally an entire 8 hour day where you get ask the same questions with different groups of people, scheduled with no breaks (no 10 minutes between sessions is not a sufficient break), getting ask LC-esque questions twice, system design, and 5 different behavioral sessions. Worst of all I never spoke with a single person who would be on my team, to make matters worse Capital One doesn't exactly pay well. It's a very average company, with average salaries, and average benefits.
My interview at Meta was easier and more straight forward, it felt like they actually cared about my time as a candidate whereas Capital One puts you through the meat grinder and forces you into their terms.
We ask candidates to give us a schedule of when they are available, and then schedule against that. Some folks would prefer a long gauntlet all in a row, some folks prefer things to be spread out. This lets them do that.
Some of it quite poorly, which is why Sun lost to Linux and other hardware companies, and why basically no one used Joyent.
>And that's who runs Oxide.
It's basically "rewrite it in Rust" Sun Microsystems. Best of luck, but I don't see them being anything besides a niche player with some interesting tech that doesn't beat the competition enough to matter.
Every interview stage which is designed to reject bad candidates (true negatives) will also accidentally reject some good candidates (false negatives). The more eagerly you filter out bad candidates, the more you'll also filter out good candidates; sensitivity decreases as specificity increases.
This means that building more pass/fail tests into your interview pipeline may produce worse results - even if you have an enormous number of applicants, all with infinite time and patience! At each interview stage, there's a risk of accidentally rejecting the best remaining applicant, the one who would have outperformed all others if they were hired. For example, the article mentions "values mismatch" as a particularly good reason to hard-reject a candidate, but several of the company's values seem like they ought to be optional. Surely some truly excellent software developers lack courage, humour, and thriftiness? Could those qualities just as well be taught on-the-job? Could a non-courageous hire have useful things to teach you about caution?
Good hiring processes should draw a clear distinction between "must-have" qualities and "nice-to-have" qualities. Beyond a certain point, the problem stops being "filter out bad candidates", and starts being "out of several great candidates, make sure we choose the best one"; to solve those two problems, different tools are required.
[0] https://en.wikipedia.org/wiki/Sensitivity_and_specificity
The hiring process in most German companies is much more humane and efficient
chunk-Z7TDCR3L.js:6 Uncaught Error: Minified React error
#418; visit https://reactjs.org/docs/error-decoder.html?
invariant=418 for the full message or use the non-minified dev environment for full errors and additional helpful warnings.
at ki (chunk-Z7TDCR3L.js:6:4780)
at Ba (chunk-Z7TDCR3L.js:8:45229)
at Ua (chunk-Z7TDCR3L.js:8:39459)
at Af (chunk-Z7TDCR3L.js:8:39432)
at Fa (chunk-Z7TDCR3L.js:8:34497)
at Sl (chunk-Z7TDCR3L.js:1:1717)
at MessagePort.hl (chunk-Z7TDCR3L.js:1:2107)If you have the qualifications for this type of job (high skill high pay) then life is pretty much choose your adventure mode.
There are people in this world who would willingly play a modded version of a game which has a level harder than hardcore. On the other end of the extremes, you have people actively looking for "Rest and vest" gigs.
The people working at SpaceX or Apple under Jobs are one of the most highly sought after talent in the world.
They are no forced out of circumstances or coerced into any of this.
(I currently work at Oxide, having taken a pay cut.)
In reality, people (especially the exceptional kind) have all sorts of seemingly unintuitive motivations.
It bears repeating that these are not low skilled unsophisticated people in untenable situations vulnerable to coercion. They are the opposite of that in every way possible.
> As a company, we believe that good writing is found in the best practitioners, regardless of role: the act of writing — like so much else that we do — requires not only the ability to create wholly new material, but also the ability to reflect, correct, revise, and polish.
I would like them to loan me one of their stacks for a while and maybe pay me to come up with some neat stuff for it to do.
And bullfrogs want wings, so they don't bump they little butts when they hop.
I went from application on 14/3/22 to rejection on 4/4/22.
- their process works
- they skip it for the right candidate
I would advise Oxide that if SteveKlabnik is not in the PR field with experience in handling HR style topics such as this thread, then to have Steve refrain from doing so, because this isn't helping the community perception of Oxide. You're getting destroyed in the comments.
https://steveklabnik.com/writing/today-is-my-first-day-at-ox... ("Steve Klabnik is the Community Team Leader for the Rust team at Mozilla")
2. Publishing the exact hiring process is a refreshing and welcome step in the industry where interviews are a black box. I also welcome the work that's been done to reduce as much bias as possible. However, this all begs the question: Do we really need this much evaluation in trying to hire an employee? I believe the pandemic and WFH has shown that people want to prioritize their personal and family lives more than "giving it all" to the workplace. The sense I get from reading the Oxide hiring process is one where you truly have to believe and be in love with the Oxide culture. Why? Why can't we just treat work as a means to an end, and that you can still good quality work with that style of thinking? It does not have to be an either/or, but the hiring process document makes it seem like it's "you either fit into our style and values, or you don't". Why the gatekeeping?
2a. The amount of hoops software engineering folks have to go through to land a job is just mostly made up stuff, based on no studies or research on workforce hiring, and at a startup, HR is just really payroll, with no formal processes on handling grievances, human process, etc.
2b. Having said 2 and 2a. Why do we, as an industry seem to mired in spending so much effort on weeding people out? What are we scared of? All of us have worked with people we don't like, or that we feel were not good enough (caliber, skillset, professionalism, etc.), but this and other process isn't helping. What if we changed our mindset to actually trusting and believing in people to do the best they can rather than send them so much homework, have them sit through panel interviews and then conclude with "hire, do not hire" and then provide the candidate absolutely no feedback on what went wrong. Why can't we seem to disrupt this. It's tiring, and it shows in this thread that people are quite honestly sick of it.
3. Personally, when I attend future interviews, I want to have a hard boundary for the type of questions that get asked. If I get thrown a "reverse this integer" style question or "lets do TDD on the white board", I think as a collective we need to walk out or stop the interviewer and put them on the spot and go "You really want to waste an hour of my time asking basic questions like this? Don't you want to know how to solve real problems that you're currently facing or are you not facing any challenging problems so you want play with puzzles and waste my time?" I would bet if the masses resisted and took this course, interviews would change drastically and quickly.
EDIT: @Oxide, bcantrill isn't helping in the comment section either. This isn't going well for you and I strongly advise you to hire HR or PR to handle commentary and writing about these things rather than engineers/founders.