News for hackers. Stuff that matters.
Re:Algorithmic bias bounty challenge by slg on Friday July 30, 2021 @03:48PM (#28011642)
I never really pay attention to these bounty challenges. Are those rewards reasonable? They seem incredibly low compared to the work involved. I have seen locally sponsored hackathons with higher total prizes.
Re:Algorithmic bias bounty challenge by noxer on Friday July 30, 2021 @04:36PM (#28012066)
It's PR nonsense. They dont want anyone not working for them to work for them.
Re:Algorithmic bias bounty challenge by potatoman22 on Friday July 30, 2021 @04:43PM (#28012128)
3 grand don't seem like nonsense to me.
Re:Algorithmic bias bounty challenge by daenz on Friday July 30, 2021 @04:46PM (#28012169)
For a week's worth of work, for someone skilled enough to be capable of winning this contest, and there is no guarantee for winning? It is nonsense.
Re:Algorithmic bias bounty challenge by 55555 on Saturday July 31, 2021 @07:00AM (#28017021)
But to Twitter, it does.
Re:Algorithmic bias bounty challenge by version_five on Friday July 30, 2021 @05:43PM (#28012747)
If they hired a contractor to work on this, the $3500 prize would get them 1-2 days work. At the same time, I can easily see someone interesting in this area investing time in it for the challenge. Doing a hackathon as a source of income is probably not too common.
Re:Algorithmic bias bounty challenge by shoo on Friday July 30, 2021 @07:47PM (#28013824)
Seems like a hackathon. s/hack/work/ .

Compared to the $1m netflix prize from 2006, $3.5k seems like a meagre bounty.

There may be the case where someone has already been thinking hard about solving similar problems for the last couple of years, and this challenge falls fairly easily into a framework they've already got in place, so it is relatively quick and easy to crank the handle to generate a challenge submission (or a paper or whatever).

Re:Algorithmic bias bounty challenge by ryankupyn on Friday July 30, 2021 @04:46PM (#28012156)
I think the challenge is that if the rewards were high, Twitter employees (with the advantage of inside information) might be tempted to "tip off" an outsider in exchange for a cut of the reward, rather than just reporting the issue internally.

At the same time, there isn't much of an outside market for algorithmic bias info in the same way there for security vulnerabilities. Probably the biggest effect of this reward will be to pull some grad students who were going to study algorithmic bias anyways towards studying Twitter specifically - after all, there aren't any rewards for studying the algorithmic bias of other companies!

Re:Algorithmic bias bounty challenge by codyb on Friday July 30, 2021 @04:45PM (#28012152)
I figure it's kind of like a musician busking on the street corner.

It's a no risk way to add a few constraints to the practicing you're probably already doing and if you're lucky maybe you'll earn a few bucks too.

Re:Algorithmic bias bounty challenge by qPM9l3XJrF on Saturday July 31, 2021 @04:20AM (#28016391)
https://aipaygrad.es/

Average total comp for a novice AI engineer is about $300K. That's over $1000 per weekday.

If they made it so the expected value of entering was comparable to the amount someone could make in their day job / as a contractor, I'll bet they could attract more & better submissions.

Re:Algorithmic bias bounty challenge by jb775 on Friday July 30, 2021 @05:10PM (#28012411)
Seem insanely low considering it's really just a for-profit business asking the general public to figure out a very difficult problem they aren't able to figure out on their own.