That's likely why 18th and 14th street are so similar: It only takes a minute or two to get from 18th st -> 14th st.

I'm only half-surprised.

I had a great experience, we learned a ton, and if anything was going to set us up for success it was our time there. Companies are mostly doomed to fail, but our time in Seattle was pretty transformative for me personally. It allowed us to refine our focus and dedicate ourselves wholly to building something people wanted (which, turns out, wasn't that many people obviously). The support and guidance we received from Aviel, our Amazon partners, and fellow cohort members were unparalleled. Say what you want about VCs or Amazon or startup founders (and yeah there are many things to be said), but I really have nothing but great things to say about all of the individuals from our time there. Admittedly, my opinion doesn't carry any particular importance.

On the other hand, I'm not surprised at all when I reflect on the actual Techstars program. Techstars, as an organization, seemed totally in the periphery for much of the program. The lion's share of valuable advice and resources came not from the organization itself, but from everyone else.

Echoing another's sentiments, the value of Techstars seems heavily influenced by location and the mentors involved. We were lucky to only be thinking about two great programs, Seattle and NYC. If we had ended up somewhere else maybe I'd be completely unsurprised.

A friend and I tried to start a company around this, basically resolving three big issues we saw with take-homes:

1) No feedback. If you spend a couple hours on something you should get meaningful feedback.

2) No clear criteria. Is the hiring company going to slam you on correct syntax? If so, you should know that upfront!

3) Time guidelines that had no enforcement and thus led to an arms race where candidates would spend ever-increasing amounts of time and skew the standard of what "good" is.

So we fixed those things:

1) We provided the feedback, not the hiring companies, so legal liability was non-existent for the hiring companies. We double-blinded the process as much as we could (evaluators didn't know who the candidate was and vice-versa).

2) We told candidates upfront what they'd be evaluated on. Not down to the level of "you must implement this problem using a max heap", but we would say something along the lines of "The company is looking for an academic algorithmic solution to this problem" or similar. We would then only allow evaluators to evaluate them on these axes and nothing else.

3) We also strictly enforced time limits by basically telling candidates "hey you'll have 2 hours to submit from the time you hit start and see the prompt, so please make sure you have two hours from when you hit start." -- not ideal, obviously, but the best we could come up with to resolve #3 above.

As you can probably imagine, the market just wasn't really there for this. I think candidates generally enjoyed it in comparison to the vague, unending slog that most take-homes are but:

1) The value prop just wasn't really there for most companies: They mostly use these types of evaluations on more junior candidates, and unfortunately the hiring market for junior candidates is highly skewed towards the employer.

2) More surprisingly, we realized the time their current engineers and managers spent evaluating these takehomes just wasn't really a consideration for them. We tried to frame it in terms of "here's how much it costs you to evaluate these take-homes wrt time spent vs. us", but it was a difficult sell regardless.

We actually had the most success evaluating candidates from more non-traditional backgrounds upfront ourselves and then charging a placement fee if they were hired, but we ultimately didn't really want to continue that.

Dark, Babylon, How to sell drugs online, and 1899 (RIP) are all fantastic series.

They migrated away from it at some point, but some of the people who handled that migration were still around when I was there. Didn’t sound fun at all, sounded like a horrific nightmare.

For an e-commerce platform that allows returns: Yeah that's kind of an important sticking point and was a big part of the problem.

The other problem was that spells were effectively a service and you can imagine that we didn't want to get involved in determining the validity or accuracy of services being performed virtually that had zero tangible artifacts.

> What if I were to ship a complete spell in a bottle to the person?

This is probably what people actually do now to get around it, since there's an actual item being sold now.

The exact details might've changed, but the symptoms look the exact same. I remember we had to deal with a lot of people trying to skirt OFAC regulations, for example, by doing all sorts of things to hide the country of origin for something.

Optimistically I will say this is a very hard problem to solve. Pessimistically of course the company benefits from it, so I'm sure it's not a problem anyone besides maybe a few diehard advocates look too deeply into.

I've "prematurely" split up code across multiple/classes functions so many times before these huge god functions became set-in-stone and unwieldy.

I had a co-worker once who was extremely heavy on YAGNI--and I don't mean that as a pejorative, it was a really helpful check against code solving non-existent problems. They once called me out on a refactor I'd done surreptitiously for a payments flow to split payment states into multiple classes. Serendipity had struck, and I was able to point to a submitted (but not accepted) diff that would've broken the entire payments flow (as opposed to just breaking one state of the flow).

I always think about that every time I question whether I'm isolating code prematurely.

Sure, we definitely used to eye-roll at some articles. But my ass would've been fired in 5 seconds for even sending them an email, let alone anything beyond that.

Population density means zip codes are extremely concentrated in NYC, especially in Manhattan. Traveling down the length of Manhattan might have you cover ~25 zip codes in 13 miles. Basically, a zip code isn't a meaningful measure of distance like it might be elsewhere. A radius might also place you into NJ, which could be an issue depending on what you're trying to accomplish: Delivering something from Queens might be significantly easier than from NJ, even though Queens might be further.

That's my guess anyway, though it's odd because nowadays generally looking something up via zip code + radius is actually easier in Manhattan (with a map you can easily see whether it would require coming from Jersey, crossing the park, etc. etc.)

I'm not going to pretend we didn't roll our eyes from time-to-time when a new article came out that seemed exaggerated, but the thought of retaliation never crossed anyone's mind.

I'm pretty sure we would've been summarily fired if we had even made contact with Ina, let alone any of the psycho shit outlined here.

Doing a cursory google search, I can see most used cartridges going for ~$25-40. Is an unopened cartridge worth many more orders of magnitude? Possibly, but it at least raises a lot of eyebrows that there’s something else at play.

EDIT: An analogous collectible is probably more something like a black lotus card, which as far as I can tell has “only” procured 500K at auction for a flawless one, and as far as I can tell inferior versions cost an order of magnitude less (10K or so) but nothing near to the spread here.

Whether that's true or not I have no idea, but it's certainly plausible.

It's been awhile since I've taken stats, but people semi-frequently overestimate the necessary sample size needed for even a 99% confidence level.

Depends on the standard deviation for sure, but some quick napkin math suggests that a sample size of ~700 people is more than enough to draw inferences about the entire western hemisphere at a 99% confidence level.

I'm mostly kidding, but I do think it's worth pointing out that most (if not all) of those games have been accused of being extremely unfriendly to new players and generally being miserable experiences until you acquire a baseline level of knowledge.

I think the more important part of those games you listed is that they're designed around team play (and thus friend groups), with the obvious exception of Hearthstone (which compensates for the burden of knowledge by having a lot of randomness and a VERY forgiving ladder at the low end).

Getting destroyed in League is fun if you're playing with friends; with strangers it usually just results in a lot of angry messages.

> it is a pretty terrible game to play if you just want to unwind after work.

I think this is really the crux of why games that encourage team play will always win out: It's fun to play with friends, win or lose. SC2 tried this a bit, but at least when I used to play it was very clear that the game was completely unbalanced around team play.

Most of these games, if you want to really succeed, are mentally taxing. You have to pay attention to a minimap, you have to constantly be on the lookout for other people, you have to remember exactly what cards they've played so far.

If you can't counterbalance that with at least some socialization and lightheartedness, it's just work.

Sometimes it felt like a chess match where you had executed your strategy perfectly, but randomly three of your opponents pawns turned into queens.

I get that you're talking about something more subtle than that (and maybe more akin to just, well, card draw RNG) but it's a delicate balance because it can also feel really bad for players being on the receiving end of randomness.