I think its important to understand that Silicon Valley practices are not representative of societal trends. The examples the author uses are about going to a welfare office and interviewing for an administrative position at a cosmetology school. Just because one trend doesn't exist amongst a class of wealthy people in San Francisco doesn't mean it doesn't hold elsewhere.
So its fine to disagree with this article but I dont think that the experiences in the current tech age in Silicon Valley do anything to invalidate her experiences at welfare offices in the Mississippi delta.
I actually dont think you could do that because teams need to randomize amongst plays or else defenses would crush them. That is one of the coolest (in my opinion) parts about football is the amount of strategic depth that goes into it. For a quick reference: http://www.advancednflstats.com/2008/06/game-theory-and-runp...
We're definitely testing out some new things. We never want to show any ads on our sites (we hate ads) so we're looking for alternatives. Based on some user research, paying a little bit for the potential of saving a lot by making better offers was at least interesting to buyers.
I have no proof for what I'm about to say but here it is anyway.
The only information that bidders will practically care about is the current top bid. So the information that you are charging people for is the same thing that most auctions already display (that is, the current top bid). As the seller, I want people to see this top bid because it will raise my revenue. Charging people to see the top bid will result in less people seeing the top bid and lower revenue overall. It makes money for the site but less money for the seller. Then I might want to go to another site and make more money for my item.
You might want to look into using a 2nd price auction (like eBay) does. There's a nice book called Auction Theory by Vijay Krishna that might be of interest to you guys.
I appreciate the feedback and can give you some of our thinking. Our target audience of buyers generally has no access at all to this type of information today for local transactions. You just send an email into the void and cross your fingers. Providing this information at all is an improvement to this audience.
For the seller, showing the top bid means that people will bid no more than they have to. Witholding the top bid (except to those who are interested) means buyers offer what they think the item is worth - which could be a lot more than top bid + 1$.
It's definitely a different approach and may fail miserably. At which point we'll pivot and come up with a better one :)
Witholding the top bid (except to those who are interested) means buyers offer what they think the item is worth - which could be a lot more than top bid + 1$.
That's not true. I'm not going to put down what I believe the item is worth because then you will charge me that amount. If I believe this item is worth 100 and you charge me 100, I get a final value of 0. What I actually do is bid under my value and its not exactly clear what I do.
If you sell the item to the guy who bids the highest but charge him the second highest bid (known as the 2nd price auction), then people are incentivized to bid what they think the item is worth.
I'm not claiming that you should necessarily do any particular auction format because the theoretical results are for scenarios that typically dont hold in practice. However, I do recommend reading a bit about auction theory to help guide your decisions.
Who is that? It looks to me like it's local-transaction buyers. This could definitely make craigslist selling a lot better.
It wasn't clear to me what you did -- obviously I read "auction" and I think eBay. But you're much different. The marketing copy didn't explain it well to me, but the video gave me that 'ah-hah! Nice!' moment.
We've found that most users won't read much on a home page so we kept the copy short and focused on the video. We'll try and clean it up, but getting people to watch the video seems to be working out the best.
I'm not sure machine learning algorithms are the right approach. Crowd curation mechanisms (i.e. the upvote/downvote systems but one could imagine other systems) have the advantage of quickly adapting to the ever changing tastes of a community. I don't think machine learning approaches can be simultaneously good and flexible in such a diverse environment. The Netflix analogy isn't the best because Netflix deals with a pretty structured problem in comparison to the "recommend me things on the internet that I will like" problem.
I will also claim that subreddits are by no means a hack. They are THE key feature of Reddit that allows users to join together on the basis of common interests. There's no replacing subreddits because they are a killer feature in their own right.
That depends on field. In most areas of computer science, conferences are more important than journals (journals generally just publish extended conference papers).
I am not aware of any CS field where conferences are in general more important than journals. Though there are a few conferences whose rigor approaches journal publication.
I think the standard ordering is: books, chapters, journal articles, strongly refereed conferences, lightly refereed workshops, unrefereed works, dissertation. And numbered! Good grief.
So its fine to disagree with this article but I dont think that the experiences in the current tech age in Silicon Valley do anything to invalidate her experiences at welfare offices in the Mississippi delta.