Hacker News Reader

I got my big break through triplebyte but once you switched from the “personal” site that connected me to companies I was actually a good match for and moved towards being just a jobs site I’ve had zero luck so I got my last job via workatastartup (which is also going downhill.)

It’s a shame because I know there’s people like me stuck in a dead end job somewhere despite being a senior level engineer (I was working construction!) and these people will have nowhere like triplebyte to help show the industry their skills.

Yeah. You weren't alone. My favorite was some guy who'd been packing boxes for Amazon for like ten years and couldn't get anyone to give him the time of day, who ended up getting a like 200k a year job at one of the bigger SV companies. That kind of story was something I couldn't shut up about during my first couple years at TB, and it was definitely something lost when we changed models. (We made that change for a reason - discussed elsewhere in this thread - but it was still sad to see that model go.)

The issue of "talent but no experience" is a glaring inefficiency, but it's the kind of inefficiency that (so far, at least) it's been very hard to make a business out of. (And to be clear, this isn't just employers being evil, a lot of their behavior is rational and it's not like candidates never lie or cheat either.)

I have some thoughts on this that I'll probably be batting around with my new bosses a bit once we're settled in with Karat's teams. Since Karat focuses more on screening than on trying to be a marketplace, it's a more natural fit for their (our, I guess I should say now) existing business anyway. Hopefully we can make it work.

TripleByte once told me it was aiming to becoming “the hiring division for Google, Apple and all of big tech”

As a mere jobseeker in the pipeline, I wasn’t sure if the representative understood the reality of recruiting. To me, the ambition sounded ignorant; companies do not want to outsource and centralize their hiring in a third-party that is also hiring for their competitors.

This wasn't as much of a barrier as you'd expect. Yes, recruiters mostly didn't like the idea of framing a recruiting process as "you're getting people X other company didn't hire" (even though, frankly, that's most candidates you get anyway and hiring processes are not that strongly correlated), but that interpretation usually didn't come up in the first place.

The biggest single barrier we faced for early TB was pricing (we were charging 20% of first year's salary) and the demand that companies skip candidates straight to onsites (which we were promising them at the time).

Later on, our biggest barrier was differentiating ourselves from Linkedin and the like: companies saw Linkedin as the primary sourcing tool, and because they were hiring non-engineers there, it was one they usually wouldn't get rid of. So we were always the secondary tool, which meant (a) we tended to get the roles they couldn't hire for elsewhere, which selected for the hardest roles, (b) we had to justify ourselves in a way linkedin didn't, and (c) we were the first to be cut when the economics turned against us.

I was in early TB and had a good experience at first that later went sour. You may have explained a little of what happened. Oh well.

Thought you were onto something back when pilot helping mega-enterprises with talent.

You were the only way to get well-assessed devs in the door despite non-technical hiring managers and an HR department asserting any assessment was discriminatory.

Government can't buy open source since lack of support fails due diligence, so wash OSS through vendors offering enterprise support. Similarly, talent can't be hired well directly, so wash great talent assessment for enterprises while they stay compliant, it should be a gold mine.

Most every enterprise in America needed (needs) that, and any of them running a Transformation had someone that would have been a sponsor.

(Btw, looking for folks interested in setting up a new kind of systematic person-role matching, and while still somewhat marketplace-like, profit could be on the order of $500K per match. Know anyone interested in trying that?)

> We tried some new tactics near the end that seemed to be bearing fruit, and I think in a world where tech hadn't crashed in mid to late 2022 we would have had a reasonable shot.

Is this referring to Magnet? I thought that was neat product idea.

Yeah. (For those unfamiliar with it, Magnet was basically an attempt to automate sourcing. The way it worked was that companies would complete a questionnaire about what they wanted, we would message candidates on their behalf, and they would review only candidates who indicated interest. We applied some throttling and quality thresholds to try to make this a good experience, because taking control of the messaging let us actually do that.)

Magnet grew out of four main observations:

* Our users got much better results when we helped them search rather than leaving it to them.

* What company users fundamentally wanted was to - as our head of finance put it during a discussion last January that ended up leading to the creation of Magnet - "swipe a credit card and get candidates".

* The most interest we'd seen on the candidate side was when we could tell them what companies really wanted, and bringing that in-house meant we could show that information very accurately and without incentive for companies to be dishonest, and

* We had a fundamental lack of activity problem. We talked often about how nice it would be if we had Linkedin's spam problem, because we had lots of ideas for how to solve that problem if we actually had it, which we didn't.

So we built Magnet, which ended up increasing activity on platform by a factor of like five and made companies way happier. We scaled it up over time, and right as we really turned it up in June, the market fell out from under us. And it's a testament to just how well it worked, at least on paper, that our number of connections between candidates and companies actually grew through much of fall 2022 despite the collapse in hiring.

That's not to say Magnet didn't have problems. It did. It had issues with fraudulent profiles (there turned out to be quite a few of these - a lot of job sites are dealing with this problem at the moment in the remote-work world), and we were definitely hitting a point of diminishing returns as we scaled it to some extent. But given how quickly it was built, and how bad the market was, the fact that it took off at all suggests that there was probably something there.

This is model-adjacent to what we're looking for someone to help try again, but with a return per match that makes it easier to make a company out of.