With open source and hackathons, I worry a little that the chance of this coming to programming is creeping up over time.
Quite a few startups are trying to put together spec-work marketplaces for tech. For a while Kaggle was probably the most successful, operating as a marketplace for data-science spec work, but it’s transitioned a bit. The basic setup is that someone posts a data-analysis problem they want solved (usually a predictive one), along with a bounty, then people have N weeks to come up with solutions, and the top one is paid while the rest aren’t.
In Kaggle’s case they’ve transitioned more towards things that look like research competitions, though. In the earlier days there were a lot of $200-$1000 bounties from companies who needed a specific system built, which is classic spec work. At the moment I don’t see any bounties under $10k, and the problems are mostly interesting research challenges rather than purely a company bidding out a solution to a system they want built. You can still think of that as spec work of a sort, but subjectively it feels more like a competition with a prize than the $500 problems did.
I mean, yes, but art and writing competitions have always been a way to disguise spec work.
Yeah, I think there’s quite a bit of that. I think there’s a continuum, though, where at the other end you have something like the Nobel Prize in Literature, where it would be a bit of a stretch to call that spec work. The main features I see spec-work having are: 1) the work is strongly tailored to a specific buyer, so if they don’t select your submission, you’ve spent time to produce a now-useless product; and 2) it’s frequent enough that you can plausibly end up relying on it as a source of income, getting stuck in a job with fairly bad payment conditions. Hence a marketplace full of a ton of $500 data-analysis “challenges” feels more paradigmatically spec-work to me.
Your point (1) cannot be emphasized enough. Prizes for ‘best x’, where doing x will still get you paid anyway are an entirely different beast than prizes for ‘best x’ which are the only expected income from doing x.
The thing that comes to mind for me with the Nobel Prize is that it’s nominating work that was published for reasons not related to the prize. And the recipients retain ownership (or, more likely, their funders do, but that’s a separate topic).
Lots of contests, even the losers have forfeited ownership by participating. Lots of bounty sites, the work wouldn’t be useful to anyone but the original poster. There’s certainly a gradation of how exploitative it is. I agree with the line you’re drawing: it’s the lock-in effect that needs to be opposed strongly.
With that said, I don’t think it being a sustainable income on bad terms is a requirement for it to be exploitative. Buyers almost certainly don’t care at all whether it’s sustainable for workers, and that’s not a fundamental feature of the manipulation. It’s sustainable from the buyers' end as long as they can keep feeding the pool of people who don’t know how to value their work properly.
Getting stuck definitely sucks worse than failing fast, though. My personal experience is with the latter. :) Hard-line negotiation is not something I have any desire to become good at, you know?
I had a lengthy example here with the economic and emotional trade-offs of professional athletics, but frankly making that comparison would have been kind of a jerk move, so I’ve deleted it. :) It wasn’t really supporting anything salient anyway.
Everybody’s personal line should be that they won’t do work which involves risks they can’t absorb. Spec work is a shorthand to help us talk about that, and doesn’t really need a rigid definition.
So… Just say no to bug bounties?
Interesting that Apple (among others) is roundly criticized for not offering bounties (It’s 2015! Get with the times! Stop being greedy!), but what is a bug bounty if not a spec work security audit?
If software was only capable of having one bug and it was a race to find it, that would be awesome.
Like say, the tarsnap exploit competition… https://lobste.rs/s/nemgwt/tarsnap_1000_exploit_bounty
From the vendor’s perspective, in addition to the direct value of it, it’s also a way to calibrate their sense of how many vulnerabilities their stuff has. The latter is arguably an external good, and it’d definitely be nice to have a way to achieve it without this issue. :/
but what is a bug bounty if not a spec work security audit?
I’ve always looked at bug bounties as a company purchasing a dangerous vulnerability so it can’t be used by others. Rather than a competition, it’s a simple exchange of money for labor.
But now you’ve got me thinking. Microsoft pays well for (some) bugs – and I think they consistently pay on the high end for bounty payouts – but I can’t help but think that “the bad guys” would be willing to pay much more.
Money for labor doesn’t account for all the people who scan a site and find nothing. I’m pretty sure I could poke at Facebook or Google for a week and come up empty. Do they owe me anything for my time? There are a lot of unpaid losers in bug bounties too.
Probably the people who complain about spec work are not the same people who complain about not having bug bounties.
@Irene’s comment about professional athletes is entirely on point. There are some other situations like this.
The classic dollar auction is one: you auction off a dollar bill, initial bid 5¢, with the proviso that you have to pay your bid if you come in either first or second, while only the highest bidder gets the dollar. The price can get bid up to several dollars. This is highly profitable for the auctioneer. All the bidders lose.
There are lots of dollar-auction situations in life. Most would-be professional athletes never make it out of the minor leagues, or even college football, and a lot of them end up with permanent injuries. Freakonomics talks about drug dealers living with their mothers because they’re in a tournament to move up the gang hierarchy. Accounting firms and law firms with up-or-out policies do this too: they consume a constant flood of fresh meat, who put in the long hours that are necessary but not sufficient to get promoted. A few win; most lose. Grad students are in a similar tournament for cushy academic positions. Many drop out without their doctorates. Most spend years in postdocs and adjunct posts, hoping for a tenure-track position; of those who get one, many are denied tenure. Not a bad life if that’s the kind of thing you like, but it does involve a lot of material sacrifice.
There’s another dollar-auction tournament in town, too: startups. If your startup is among the 90% (or, if we believe, YC, more like 75%) that fail, or the even higher percentage that take a down round or acquisition, the VCs get the money due to liquidation preferences, and you get nothing. (YC is, to their credit, trying to improve that situation.) The chance of getting rich attracts some people, but it’s not worth it. You have to love this life or it’s hell.
(For what it’s worth, this isn’t exactly sour grapes, even though I’m not really doing a startup right now. I’ve lived for years off startup acquisition money; the down payment on my first house was from a startup IPO bounce. But it’s totally a dollar auction.)
You know what’s not a dollar auction? Open source, free software, and decentralized services. You get to keep all the software you write, and all the software everybody else writes, too. And so do they. Everybody wins. The only people who lose are the would-be parasites who want to charge people money to share with each other and get mad when those people just help each other instead. It’s exactly the opposite of logo contests, dollar auctions, and startups.