Clojurians Log v2

Hey, San Francisco! Anyone else still in the city?

It's certainly a quiet channel these days 🙂

Last post here was in March before today according to Zulip's archive.

@paul.legato are you actually in SF or "just" in the Bay Area? It does seem like Clojure usage here has gone down a bit over the last few years...

Hey Sean! I’m in the city.

It does seem like that to me, too. I wonder how much of that is Clojure holding flat while Go, Python and JavaScript have surged, onboarding many new developers with large scale commercial support.

Yeah, I think several of the Clojure hotspot companies imploded a while back or got swallowed up by Large Corp and so a lot of the excitement and activity died down. I'm still out in the East Bay and still doing Clojure full-time, at the place I've been for a decade now (longest I've ever worked anywhere). My teammate is hiredman and he's been with us about four years.

It’s a chicken and egg problem. Many times, I’ve pushed to hire more Clojure devs, and I get variations of, “Sure, Clojure is great technically, but the available talent pool is very small compared to $OTHER_LANGUAGE.” Meanwhile, when someone starts becoming a software engineer, they naturally gravitate to the languages with the most jobs and largest communities.

Even places like Climate Corp only had a few Clojure devs on staff.

Meanwhile, entire generations of programmers come up studying specifically for a Google job, which requires you to use Python, JS, Go, Java, or C++ in the interview.

It seems like there were a lot more small startups willing to make esoteric technology choices 10-15 years ago, too. FAANGM have swallowed up most of that talent pool.

I'm very cynical about FAANG (and Silicon Valley in general) 😉

Yep, me too!

They have irrevocably altered the technology ecosystem.

And perpetuated an awful interview process 😐

uggh. Algorithms trivia challenges.

Huge companies need highly repeatable processes that scale to very large numbers. Medium and small companies then cargo cult them.

A large false negative rate is acceptable at Google scale (if not good), because monopoly revenue and way higher than market compensation means they have a line around the block of people who want in. The error rate washes out.

I’ve also heard the theory that algorithms trivia is functionally just a legal proxy for an IQ test, which would be illegal in the US as not directly related to the job in any sense.

The good news is that Clojure and its ecosystem did survive the sort of mass extinction even that’s happened. The environment has changed, and Clojure is adapting to the new system.

On a broader level, I continue to see slow, steady growth in the global Clojure community.

Google pestered me for years to go interview and eventually I gave in and agreed to a phone interview with an engineering manager. It did not go well (which was what I expected). I stopped the interview after about ten minutes and we went on to talk about interview practices instead. Google have since admitted that not only do they have a large false negative rate but they also end up hiring a lot of engineers who aren't really as good as Google liked to pretend they were. But as you said, they have lines around the block so their lousy interview process "doesn't matter" 😐

I've interviewed with Microsoft and Amazon too. The former wanted all core engineering staff to relocate to Richmond at the time which was a deal-breaker for me. The latter went really well but, ultimately, I hear a lot of bad things about how competitive their career ladders are and I'm just too old to want to deal with that s**t, so I politely turned them down (and they still send me job reqs every few months, asking me to interview again).

haha. Sounds similar to my experiences.

I think Google mainly hires engineers as an area denial tactic — mainly to keep them from working for competitors, not to actually produce anything of business value. Anecdotally, all the engineers I know at Google tell me they’re put on make-work dead end projects that get canceled after a year or two. They don’t care (at least for a good long while), because the pay is immense. I see similar stories on Hacker News, etc. all the time.

They have people with PhDs scraping hotel web sites for room rates. Stuff like that.

And their much-vaunted "20%" time only kicks in after a certain period of time -- and for quite a while after joining, you have pretty much no say in what projects you work on. I learned that during my phone interview with them. Which seems like a real bait'n'switch tactic to me.

Yep. And many people deal with it — the people who would’ve started companies and used Clojure 10 years ago are writing hotel scrapers in Python now, for 500k a year.

I’m in Oakland, but a couple of my teammates are in SF. So we’re around.