How Jetpac Built a Photo Quality Algorithm for $5k in 3 Weeks

http://www.readwriteweb.com/archives/how_two_startups_used_games_to_beat_the_developer.php

by datageek • 14 years ago

41 14 14 years ago

waitwhat• 14 years ago

Spoiler: The algorithms actually work by doing text analysis on the captions and other metadata, and no actual image analysis.

datageek• 14 years ago

the algorithm needed to execute quickly. image analysis takes too long.

kabdib• 14 years ago

What's the bottleneck? Bandwidth for downloading images? Understanding of an algorithm that would do the job?

lwat• 14 years ago

Technically I wouldn't call it a 'Photo quality algorithm' in that case

Drbble• 14 years ago

Spoiler 2: They achieved the low cost by exploiting Kaggle's army of machine learning practitioner suckers/volunteers/competitors.

socialist_coder• 14 years ago

Pretty disappointing. The image analysis was the only reason I clicked.

brown9-2• 14 years ago

Is 30,000 photos a large enough dataset for a ML exercise like this?

It might just be journalistic simplifying but I wonder if they overfit their data with such specific "good" and "bad" words:

Among the best words: Peru, Cambodia, Michigan, tombs, trails and boats. What photo captions are the most likely to signify a bad photo for a travel magazine? San Jose, mommy, graduation and CEO, Warden says.

humbledrone2• 14 years ago

The 60 second challenge thing strikes me as extremely short-sighted. It is impressive to be able to solve a series of tiny bugs with just 60 seconds for each one, but I don't think it's the right metric to find the elusive "10x" programmer.

It's my opinion that "10x" programmers are an order of magnitude faster than their peers not because they are lightning-fast at spitting out code (although they may be). They're faster because they can analyze the problem, drill down to its essence, and come up with an effective plan for solving it. Contrast this with a poor programmer, who might code circles around the problem, wasting time writing code that could just be taken from off the shelf. It doesn't matter if the poor programmer somehow is brilliant at punching out correct code if the code they're punching out didn't need to be written in the first place.

It's like selecting creative writers by holding a contest to see which ones can spot a grammatical mistake the most quickly. Sure, that's a helpful skill for a writer to have, but it doesn't mean that they'll write anything worth reading.

kkowalczyk• 14 years ago

Or we can assume that the company that does the hiring via 60 second challenge is not ran by complete idiots and they do keep track whether those hires end up being effective employees for them. Let's give them the benefit of a doubt.

Your argument seems to be that there are only two possible programmers: effective designers that are slow to code or fast coders that are bad designers (designer in terms of code architecture, not graphic design, of course).

I claim that there are no good architects that are slow coders i.e. all good architects are also good coders.

When you look into research of skill acquisition, or even common sense, in order to become good at something you first have to master the basics before you master higher-level skills.

In the field of programming, being able to implement a piece of highly constrained code are the basics of our craft.

Once you're really good at that, people move up to tasks that have less constrains and are therefore high-level: choosing an implementation language, designing how pieces of the whole app fit together etc.

But you won't ever get to do that if you were not able to quickly and confidently implement a function that someone else specified, just like there are no chefs who can't cut vegetables very quickly or basketball players that are brilliant tacticians but can't run very fast etc.

Mastery of basics is not a guarantee of mastery of higher-level skills but it is a prerequisite.

humbledrone2• 14 years ago

> Or we can assume that the company that does the hiring via 60 second challenge is not ran by complete idiots and they do keep track whether those hires end up being effective employees for them. Let's give them the benefit of a doubt.

You don't seem to be aware of just how disastrous a bad hiring decision can be. The direct financial cost is easily in the tens to hundreds of thousands of dollars, depending on how long it takes to identify their poor performance, warn them, and finally fire them (all the while documenting their performance rigorously to avoid frivolous lawsuits later on). Beyond that, the opportunity cost of having a poor programmer in place of an excellent one is even worse. It is far, far better to avoid a bad hire in the first place than to catch them later.

> Your argument seems to be that there are only two possible programmers: effective designers that are slow to code or fast coders that are bad designers (designer in terms of code architecture, not graphic design, of course).

I'm sorry, but you wrote a lengthy rebuttal to an argument I never made. Read my comment again. When describing the "10x" programmer, I never said that they had to be a slow coder. In fact, I explicitly said they "may be" lightning fast. My actual argument was merely that the most dramatic improvements in software construction speed are due to making better decisions about what to code, what tools to use, how to approach the problem, etc. And, in my opinion, the benefits of making good decisions about these things absolutely dwarf the benefits of being able to slam out the code in record time.

Iroiso• 14 years ago

The essence of the post is about using crowd sourcing to alleviate start-up challenges, I think they did a good job of exposing these opportunities to founders, (BTW, its also great PR for kaggle), Good Article all in all.

coderdude• 14 years ago

I'm not sure why you were down-voted at least a couple times for this comment. You've done a much better job of summarizing what this article was about than the "spoiler:" above (which is more based on the title than the point of the article).

How Jetpac Built a Photo Quality Algorithm for $5k in 3 Weeks

Related Stories