Rebooting (something like) early Triplebyte
otherbranch.com
otherbranch.com
I got roped into the TripleByte funnel through a Reddit ad, which eventually culminated in moving out to SF for a YC startup. Several years later, I had a role at FAANG and reached a level of professional $ucce$$ that was orders of magnitude better than where I had been ~4 years prior.
I wish TripleByte was still around. I remember interviewing.io doing a study on whether there was any signal from LinkedIn profiles with “skill badges.” TripleByte was the only badge that had predictive value for ability-to-receive-an-offer, but the flip side was that recruiters negatively associate these badges with profiles of people in early-stage careers, which means that you’re better served by not having any badges on your profile.
If you want one reason I wanted to start this company in particular, it's that it seemed like a good idea to try to make something everyone wants to see exist.
Triplebyte was my last ditch effort. I failed the online test the first time I took it, came back and then failed their interview. Came back again, got five in person interviews in the Bay Area. Didn’t get offers on the first four (one place even kicked me out halfway through the set of interviews). But I managed to get an offer from the last company, which I then spent the next four years at.
Had I not gone through Triplebyte, I probably would’ve given up on working in tech. Instead, I’m an L5 at Google.
What motivated you to push through so much failure?
I’ve always had sterling reviews as long as I’ve been in a job. I’ll leave it to you to wonder how much of that is competence and how much is successfully pretending to be competent.
It still sets a baseline level of competence, even if it’s not quite as amazing as it appears.
People always think I’m competent, but it’s really me just starting on ‘what I expect people to be doing in half a year’ now, so by the time they start doing it I’ve already gone through all the pitfalls while they weren’t looking.
It might feel better if you realise they "let you go home" rather than "kicked you out". If a candidate is clearly not qualified why waste their time (or ours) going through interviews that we already know won't lead to a hire?
BTW we always consider this a failure of our process (phone screen in particular), not failure of the candidate. We should never have brought them in in the first place.
When one gave me an offer, I knew I had to decline but Triplebyte's recruiter told me on the phone that they were having a company party that night and they wouldn't drink unless I accepted the offer. It felt really manipulative and weird. I decided to completely drop their service after that.
https://www.otherbranch.com/practice-coding-problem
Under step 2, the description says:
> If the selection is not a mine,
> but has adjacent mines in one of the squares next to it (including diagonally),
> change that square to display the number of mines adjacent to it.
So either the "but has adjacent mines" clause should be deleted from the description, or an additional "otherwise change that square to display zero" clause should be added, or the example output should be changed.
No, they don't. In fact, the opposite is usually true: established recruiting firms will keep the company name pretty quiet, and only reveal it on outreach to specific candidates.
Just because someone doesn't list that kind of information on the homepage does not automatically mean "dishonest". It's very common for "allow my company's name to appear on your website" to be a negotiated item in a partnership contract. Getting it approved can require involving people all the way up to the CEO. The juice often isn't worth the squeeze, and the companies you can get to agree to it are the ones you least care about putting on the website anyway.
Particularly for recruiting agencies working on contingency, I can easily see why clients wouldn't want their partners to plaster their open reqs + brand on the front page of the website.
On first look, though, unfortunately now basically any website content could have been made by GenAI. That looks to be what we're living with from now on :-(
Since you asked: we have two roles currently hiring, and one with a contract out (although I think it's <50-50 to actually get signed, they've been extremely difficult to get responses from). We work primarily with engineers and roles in the US (both because that's where we are located and because that's where the revenue needed to make this model work can be found. However, we're not closed to engineers abroad. One of the roles hiring right now is open to remote in most of the world (they prefer not the EU for time-zone reasons), the other is in-office in New York City.
If your criterion is "I don't want to sign up for this unless it'll get me a lot of jobs right now", well...don't. It won't, not yet. The pitch right now is "if you want to see this thing exist and would like a job if we can find you one, give it a shot, you've got nothing to lose".
Maybe some that would have given you their CV would be put off if they see only listings for jobs they don't match - but they don't match them, so they're not going to help your business any time soon anyway. So: the ones you gain by not listing don't help you, and the ones you lose are the ones that would.
Having only a couple of jobs just makes you look small, not weak. But there are many small recruiting forms that do well - for example, one guy who I've used, Mark Ashton, only listed a couple of roles on his website[1]. But when I was a hirer he was still a great recruiter who found candidates that were matches, because he knew the market. As an individual, I still go to his website to see if there's anything of interest. (Okay, he seems to have changed his website and it doesn't show any jobs, so not as good a piece of evidence as I thought. But probably that's because he doesn't have a client right now, rather than because he wouldn't list them).
Because that makes Otherbranch no different than an ordinary recruiting firm!
IE, when you're starting a business, not only is it important to be selective about what you do, it's also important to be selective about what you don't do.
Years ago, Triplebyte had a blog post on the front page of this very site where we mentioned, offhandedly, that we'd gotten jobs for "hundreds of engineers". That was true - the number was roughly 800 that we could track, probably into the low thousands counting ones we didn't - but there were extended threads analyzing that number to prove that we never actually got anyone jobs and our whole brand identity was a lie.
Had you done the numbers, you'd have found that 800 was totally consistent with TB's stage as a company. At 30k a placement, 800 placements would amount to 26m in revenue; at 10m or so of annual revenue and rapid growth that's a totally reasonable lifetime number. The complaint was, I think, totally unfounded. But that didn't stop it from being the top thread on that post (IIRC) for some hours.
That's the kind of thing that makes it hard to be transparent. As much as we want honesty from companies, we (or at least a lot of us) also want to go with the winners, and we often have unrealistic expectations of what "winning" looks like (because our intuitions are tuned on inflated numbers that are either lying or selected). In fact, just down this thread, you'll find someone arguing we can't really be serious about recruiting because our website was not made by a graphic designer.
> Maybe some that would have given you their CV would be put off if they see only listings for jobs they don't match - but they don't match them, so they're not going to help your business any time soon anyway. So: the ones you gain by not listing don't help you, and the ones you lose are the ones that would.
The problem is that, in the limit, this results in no one ever signing up for anything. Because candidates want us to already have the job they're a match for, and companies want us to already have the candidates they're looking for. And even if clients were willing to go for it even without candidates in the pool, it's not like candidates regularly check to see if new ones are posted on a small site. (If you did do that on your recruiters' site, I can tell you - with the backing of considerable data - that you are in a rare minority.)
One person we have at an onsite literally right now signed up a month ago when we had no roles for him. We had a role that was a fit come in a few weeks later, matched him with it and then with another, and now he's got a 50-50 or so shot to get a job in the next few days. Had he just bounced day one, that wouldn't have happened.
There's an analogy here in terms of chemistry. Trying to get both candidates and companies to both be around at the right moment is second-order kinetics, proportional both to the rate at which candidates check your site and to the rate at which companies have open roles they want you to fill. Trying to get one side to sign up provisionally - and it makes sense for that to be candidates, since job searches tend to last longer than open reccs and candidates aren't having to pay for it - is first-order kinetics, proportional only to the rate on the other side. Since at small scale n^2 is << n, it's really important for us to encourage the first-order state of affairs rather than the second-order one.
Of course, none of that gets at whether not posting all our open roles is consistent with honesty. But you're making a strategic argument here, and I think it's incorrect.
One question in my mind, though, is why is it that that most recruiters solve the coordination issue the other way - first finding companies, and then hunting for candidates to match them. IE, there must be some reason why it's more effective for them to do it that way round, and theoretically this presents an obstacle for you. But perhaps Triplebyte already figured out the answer.
I only know of one established recruitment company which does it the other way round - they don't hunt for candidates - and that's ecm selection. They operate in science-heavy areas mainly, and the chip industry, where there's a particular need for the recruiter to have a good grip on what the role means, otherwise they end up flooding their client with inappropriate candidates. But they are probably not as scalable as you want to be.
Companies burning their goodwill is such a painfully negative thing about the modern world. Like just.. don't do it, even for the money. It's shameful.
The privacy thing is a legitimate criticism. It was not my call (I wasn't part of the leadership at the time) and I disagreed with it both then and now. It was a screwup, and it is a big part of why I want to make aggressive honesty one of our major cultural points - because I think once you start with spin it's easy to roll down a slippery slope to "well it's in our pragmatic interest so sure do whatever".
There was so much value to be captured that these companies could survive on the sidelines just eating the scraps of tech compensation, and those scraps fed them well. Now the scraps are far too small to support a company.
I never used TB, but their model appeared broken. Instead of a tool used by recruiters, they aimed at replacing recruiters.
I'm stating this more as a sort of universal economic axiom, not taking a hit at recruiters specifically as people. They're nice people, but a lot of jobs are replaceable by tech.
Interviewing is a massive information efficiency problem. Companies don't know if I'm good, I don't know if the company is good, so we spend 8 hours on Zoom calls trying to hash it out, and the resulting signal-to-noise ratio is still extremely low in both directions. The interviewer's rubric doesn't give a shit if I was in physical pain that day and had a brain fart or if my ability to code a binary heap in CoderPad is or isn't representative of my ability to code a real world app in VS Code on a real task; I don't know if the interviewer had a work emergency that made them sound like an asshole even if they usually aren't, etc.
Anyways I wound up getting a job not-through triplebyte, and was competent enough that I was promoted to senior dev within a year.
I got job interviews from the original Triplebyte that I almost took just from investigating it as a source to hire engineers!
I have thought about having one available just for nerd-sniping growth-hack purposes, for the exact reason you say, but using it as a filter feels like it's just going to select for what is now very easy cheating. (Not that cheating on the TB quiz was that difficult, but it was nonzero effort.)
Pardon the grammar, no AI used ;)
Certainly the market is one likely cause if we do fail, though.
My hypothesis is that real-world collaboration and communication skills, as demonstrated through open source work, are more indicative of a developer's capabilities than typical coding quizzes. (I tried OtherBranch's sample coding problem mentioned at their post, https://www.otherbranch.com/practice-coding-problem, and got this opinion.) With the rise of AI tools, I believe the ability to effectively use these to enhance one's contributions is becoming increasingly valuable.
For those who hire or work with other developers:
1. How much weight do you give to a candidate's open source contributions?
2. Do you find that strong open source contributors tend to be better collaborators?
3. How do you balance assessing technical skills vs. communication/collaboration abilities?
I'm working on a platform to facilitate this new assessment with personalized LLM support. So I'd love to hear your experiences and thoughts!
(IE, the article talks a lot about how they are different from Triplebyte, but I never used Triplebyte.)
FWIW: I really like the idea of the recruiter doing in-depth technical screens and matchmaking:
When I'm in the US job market, the most frustrating aspect is dealing with recruiters. Most have so little "tech" experience that they can't evaluate companies or me. I often feel like I'm talking with a salesman, and it's a crapshoot if the job is a good match for me, or if I'm really what the company wants.
When I'm hiring in the US, it's a lot easier to work with overseas contracting firms. They all screen candidates very carefully, and only present candidates that are probably capable of doing the job. In contrast, recruiters for US-based employees perform very little filtering, so we end up interviewing a lot more poorly-chosen candidates.
IE, as both a candidate and someone who's hired: I want the recruiting firm to do more "work:" Better filtering, better matchmaking, and better technical knowledge. I don't want to deal with "salesman."
What you're talking about in the context of overseas contracting is essentially what we do for (primarily domestic) full-time roles.
We are much more technical than most recruiters. I'm the founder and the least technical person involved with candidates right now, and I would pass most whiteboarding interviews (I'll take yours if you want, just to make the point). Our interviewers are senior engs who've done a lot of interviewing before, most have FAANG experience, engineers generally seem to think highly of our screenings, and our matchmaking is built around technical traits (as a simple example, our matchmaking tools will infer you know JS if you know React).
As for the salesperson thing...I mean, at some level trying to pitch you on why I'm not a salesperson is a bit of an "I am no orator, as Brutus is". But I think if you talk with me, you'll get a sense that that's just not what we're about.
Really? I really enjoy interviewing candidates. If you visit my profile there should be a way to get in touch with me.
(I turned off all GitHub email awhile ago due to SPAM.)
A few differences. We do mock interviews. Other Branch doesn't (at least not yet). Otherbranch has one very good standardized assessment that they use (like the early Triplebyte one). Our way of assessing candidates is by aggregating mock interview performance (and doing some stuff to correct for interviewer strictness etc). Basically people come to us for mocks and stay for the jobs (if they do well enough in mocks).
We both do hiring, but OtherBranch focuses on a different type of employer: earlier stage startups, and we tend to focus more on hiring for later stage ones/FAANGs/FAANG adjacents.
However the 20% commission is non-starter for us. We only used Triplebyte because it had a fixed annual cost ($35k for up to 5 hires).
Even regular recruiters, who aren't running a high-touch interview on top of selling matchmaking on top of a candidate network, charge more than that!
- How many fully remote jobs paying over $200k base do you have in your pipeline? Basically I have no interest in making a profile to be a sales object to help a startup, I want the company to show value to me first [Now maybe somebody who has no current job is in a different situation]
- My current employer is largely hiring non-US. What countries do you hire out of?
Basically why before I give away hundreds of dollars of my time, what am I getting and how likely? It's not even clear to me to me what types of roles they place (e.g. SREs?)
Lastly I want to throw in a tidbit that it's my personal belief, at the later stages of a long career in tech, that technical skill is really not that important to most employers nor even most roles nor getting promoted (as compared to things like Jira-fu, making your manager look good, handling conflict). As a thought experiment consider otherbranch itself -- other than the interviewers (who could be consultants) how many engineers if any would such a company even need? Seems like it could be wordpress + coderpad + salesforce.
But to answer your question:
> How many fully remote jobs paying over $200k base do you have in your pipeline?
Depends on your standard for "pipeline", but we have one such role active right now. It's a performance-heavy low-level (Rust-based) role for a startup doing real-time video processing.
> Basically I have no interest in making a profile to be a sales object to help a startup
Sure, that's fine. You don't owe us anything, but I'd encourage you to think about the incentives that creates for founders. What you're saying, in essence, is "you should lie to me and tell me you're bigger than you are if you want to get off the ground". Which is what most startups - I believe early TB included, by the way - do, and something I'm trying not to do.
> My current employer is largely hiring non-US. What countries do you hire out of?
We're primarily US-focused, but it's not a hard limitation. The open remote role mentioned earlier is open to remote worldwide except for close-to-GMT+0-timezones.
> It's not even clear to me to me what types of roles they place (e.g. SREs?)
Primarily product-eng roles, although by chance the open roles right now skew towards low-level optimization. Our interview is definitely not optimized for SREs - you could sign up if you wanted, and we could even try and match you, but we wouldn't be able to make very strong guarantees about skill for an SRE yet (and we'd tell employers that). FWIW, Triplebyte didn't have a dedicated SRE track for four-plus years; we haven't even been doing business for four months.
> Lastly I want to throw in a tidbit that it's my personal belief, at the later stages of a long career in tech, that technical skill is really not that important to most employers nor even most roles nor getting promoted (as compared to things like Jira-fu, making your manager look good, handling conflict).
There is some truth to this, but it's less true in the startup world in which we currently operate (and will be for quite a while). If there's not demand, there's not demand, and we'll just fail as a company, but even in this market there's so far been enough demand to bring in some clients.
That sounds really prudent.
> What you're saying, in essence, is "you should lie to me and tell me you're bigger than you are if you want to get off the ground".
I'm not asking you to do that, I'm glad you're being so honest. You're right the network effect is challenging and fomo/dishonesty is a common attempt to try to circumvent it.
There's probably no easy legal way to do this, but if there were a way to give like $50 bucks of stock to each serious resume uploaded, or double that to anybody who recommends you to their current company [not sure what's involved with that] then it might create more of a sense of "rooting for you"
https://www.otherbranch.com/for-engineers
(I am asking because I am more of a data science / research person.)
The post is of a somewhat candid, high-information concentration blog post, by a bootstrapped company... and the stereotypical HN response is to criticize them for not using the latest purely stylistic glossy marketing brochure Webrogrammer resume keyword on their Web site? :)
Maybe just like tech companies are still stuck in VC growth investment scam kinds of thinking, where the quality of their software developers didn't matter so much (since Potemkin Villages don't have to be structurally sound), and now companies institutionally have little clue how to hire or even recognize great software developers... So are we stuck in forming opinions of companies based upon our VC scam era criteria?
"You're not investing in the most frivolous image BS, and any good VC scam company knows it's all about what image you present to the marks, so you're not even a legitimate business."
Did we swap the definitions of legitimate and illegitimate?
If so, what would that serious look like?
I was able to find this study:
We're a bootstrapped startup in a hostile environment being funded by me throwing a chunk of my savings at the problem. Do you think that's the most important thing for us to spend money on?
Nobody is forcing you to use your savings for this, I can understand that you’d take any feedback personally but you should try not to.
I do still think you're expecting something somewhat unreasonable from a company at our stage, but I shouldn't have said so and my mindset replying earlier was not in the right place.
But to each his own I guess....