Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
How Like.com Shut Down A Competitor–And Broke Up Its Funding Round (blogs.forbes.com)
122 points by churp on April 28, 2011 | hide | past | favorite | 32 comments


Like has always been a problematic company. I've heard from many people that they didn't even use computer vision or any algorithms for most of their matching, and a lot of it was done by hand in other countries. I don't really know why Google bought them, to be honest. Even the product they launched at google, Boutiques.com, had a 50% decline in traffic after launch: http://siteanalytics.compete.com/boutiques.com/. Good domain names only get you so far.


While I do not agree with the patent bullying especially because the said patent was very general, I would not go to the extent to say "they didn't even use computer vision or any algorithms for most of their matching". I know a couple of smart computer vision engineers who worked there, and probably the stuff done "by hand in other countries" was getting the training data and getting a lot of it, which I think is fair.


I was an early user of the technology back when it was Riya. It recognized the people in my photos with a 95%+ success rate. The riya team was a dozen solid AI experts from various US and Indian schools, together with Munjal. Google made an offer to buy them before they even launched, apparently just for the tech. So there is definitely some technology to it, I think the only difference today is that the technology is not as unique or special as it once was when Riya first launched (~5 years ago).

As for this story, I don't understand the investor reaction. You would think that finding that Google fear you would be even more a reason to invest in the company and to raise a boatload. They should have raised the 500k with a view of raising a larger round a little later on - who knows, part of the patent settlement may have been that Google buys the company. Not unheard of

ps. software patents suck. i wish there was a mutual agreement amongst valley companies not to enforce against each other. a bit like mutually assured destruction with the soviets. most people here understand it is the execution and not government idea protection that makes you successful. google has their own history of this with adwords and adsense, so i am sure they could be bought around :)


Having been bullied by a bigger company in the past, I feel really bad for Modista founders.

My best wishes to Arlo and AJ. What comes around goes around, the valley is a small place and these things don't go unnoticed.


The problem with cases like this is that the only reason Like.com was able to get away with it is because of how much money they had.

The patent system, in cases like this, is merely camouflage for anti-competitive drive-bys.


What concerns me is that the patent troll mindset has filtered down from the usual big players to the little guys who are starting out. I think if you've done something unique like come up with a cure for cancer you deserve a patent, but on the other hand so many patents seem just a bit too vague.


>I think if you've done something unique like come up with a cure for cancer you deserve a patent

even in this case i'd say it depends on whether it was done on your own dime or public money


Unless the public money came with a string that says, "no patents" then its fair game (its not unusual for the university, for example, to have first right at patents if done there). Everything is done partially on public money. You use public money to drive the paved roads. Public money to build and research the foundation of the web. Public money to have police and a judicial system so people don't simply take your ideal completely with no attribution. Public money that funded prior cancer research that you read.


>Unless the public money came with a string that says, "no patents" then its fair game (its not unusual for the university, for example, to have first right at patents if done there).

that just shows wide-spread mismanagement of public money. At all the private companies i've worked, i signed the document that all my patents done on the company's dime, are to belong to the company.

>You use public money to drive the paved roads.

And thus the contractor-builder of the road can't charge the toll on the road (well, except for the cases when he got a sweet deal what normal people would consider mismanagement of public funds )


that just shows wide-spread mismanagement of public money. At all the private companies i've worked, i signed the document that all my patents done on the company's dime, are to belong to the company

But private companies are very different than governments. Private companies generally want to make sure that no one else has this IP, but themselves.

A government, at least a capitalist one, is generally happy that you're incentivized to invent and create patents. Although they may want to also have use for themself, but it needn't be exclusive (like your employers desire).

In US we have the Bayh-Dole Act which gives you ownership of your invention, but of course, the US has a "non-exclusive, non-transferable, irrevocable, paid-up license to practice or have practiced on its behalf throughout the world".

I don't understand how you can consider that mismanagement of public funds. It seems to satisfy both the capitalist angle and the greater good.

And to be clear this only applies if you take direct funding from the government, typically in the form of a grant, where you sign a specific agreement. Using public goods, such as the roads, street lights, food inspection, don't impose any obligation on your part -- except taxation.


>I don't understand how you can consider that mismanagement of public funds. It seems to satisfy both the capitalist angle and the greater good.

As a shareholder(taxpayer) of this corporation(government) i think it is a mismanagement of the funds to allow for the uncompensated re-assignment of the rights to the direct products of the direct investments done by the corporation. You think it is not. Difference of opinions. As in the example i mentioned before - when contractor/builder gets full rights for the toll on the road he merely built and when he was paid in full for the work - i'd see this as mismanagement, and you'd see this as the capitalist angle and the greater good. Difference of opinions.


i'd see this as mismanagement, and you'd see this as the capitalist angle and the greater good. Difference of opinions.

Fair enough... maybe it's obvious from the name: VladRussian :-)


yep. I've got a chance to study Marxism theory which clearly state main principle of capitalism as "whoever owns the investment owns the profits"


From Figure 5 of Like.com's patent:

510. User selects an image of a merchandise > 520. Perform image analysis > 530. Perform similarity operation to identify merchandise that is similar > 540. Return result with images of merchandise deemed similar

From Fortune column, a description of what Modista could do:

It could take a particular handbag and match it to other similarly shaped and colored items.

According to Kumar, "Modista’s technology was better". Then, Kumar "contends that the patent should not have been given in the first place" and "Modista should’ve been given the chance to compete on its own merits."

I don't know about the validity of the Like.com patent or how different Modista's processes are from those described in the claims of the Like.com patent. But, looking at the patent issue from another perspective: How would Kumar feel if one of K9's other portfolio companies -- such as twilio, whose cofounders have filed for several patents -- learned about an upstart with a very similar technology? Would he be willing to let the company compete with twilio on its own merits? Or would he want twilio to play hardball, too?


Image similarity search has been around since eternity (read early 90's), and a patent claim that Perform similarity operation to identify merchandise that is similar (510 -- 540) should not have been awarded in the first place.


It's patenting a math theorem by replacing the generic variables with more concrete objects. This too could be patented: "a technique for calculating the number of facade bricks of a stone pyramidal structure"


learned about an upstart with a very similar technology?

But the patent wasn't on technology at all. If it was, it would have covered, for example, a method of image analysis. A competitor could then use their own, different method of image analysis.


The lawsuit's a bit aggressive, but taking a meeting with a VC when you're not fundraising in order to mess with someone else's raise is fairly common.


Really? I had never come across that myself. Are there any stories you can relay yourself or links? Very interesting.


Probably not in public, no.

The next time you get called by a junior analyst at a time when you're not raising and haven't gotten any tech press, it's likely because someone else is raising -- ringing up competitors is a pretty standard part of due diligence.

You can sometimes figure out who's out raising just by asking your own investors -- they tend to know what's going on at other firms.

Not everyone is sweetness and light in this situation. Really, blocking a competitor from raising is the only reason why you should take the call.


Citation needed, quantification needed, downvotes needed. This comment needs a lot.


If all we did around here was swap back and forth things you could just Google up a cite for, this board wouldn't be good for much, would it?

You've been given a piece of information you've never heard before, about behavior that people aren't exactly going to jump out of the woodwork to self-report. You might need to do a little digging of your own to determine whether it's true or not - but then you'll have information you can't just Google for, and that's often the most valuable kind.

Or you could just be snarky.


Seriously.

This isn't wikipedia.

And yes, this happens.


The author makes a big mistake saying 'If [Modista] was better than Like.com it would win'.

The best technology doesn't always win, and I'd argue often doesn't due to a myriad of issues.


'Shutting down a competitor' is known colloquially outside Silicon Valley as winning.


Which is why the US is spiraling down the toilet while Silicon Valley is doing better than ever.


And people still believe that there is such a thing as a "free market".


Hmm, underwhelmed to be sure.

The blog article would be better if it wasn't 100% whine. Linking the original [1] would have been better HN material.

Nobody told the founders at Modista to file a patent on their work? There's $2500 bucks that would have been well spent. They could have simultaneously achieved patent protection and a re-examination of Like.com's patent. Since Like's would come up as an office action, they would turn around and send all the prior art that shows why their idea isn't like that patent but that even that Like's patent didn't have any original ideas, which internally could have started other wheels rolling.

Like it or not, you have to file patents for things that you are doing that are cool. This does two things for you, one it brings out people who might have applied for or done something similar, and two it gives you a way to prevent other people from jumping into your space. The Like.com guys did that, the Modista folks did not.

[1] http://k9ventures.com/blog/2011/04/27/modista/


The article pointed out that Modista didn't have the funds for a legal battle. That suggests that they also wouldn't have had the funds to file a patent that would almost certainly go to waste, and even if they did, they wouldn't get any value out of it without fighting more of a legal battle (either by suing someone or defending themselves from a lawsuit).

And more to the point, why should you have to file a patent just to get a different patent re-examined? And with as little effort as the patent office puts into examining patents (as evidenced by the patents they grant), why should anyone expect them to find other relevant patents for you?

Patents just provide a tool for stopping a competitor rather than actually competing with them. "Genius is 5% inspiration and 95% perspiration. A patent is a tool for doing 5% of the work and then sitting around waiting for someone else to do the other 95% so you can sue them."


Good luck getting a decent patent attorney for $2500... order of magnitude or 2 off.


Actually I've filed several patents, have a few that have issued over the years too. I hope you weren't mislead by some attorney that you have to be a lawyer to file a patent.

I don't know if you read stuff from Nolo press [1] but we are fortunate that you can get the help you need for filing a patent in Sunnyvale [2] The last time I checked the filing fee was like $1500. You do need to put it in the form required for the application, and you need to respond when the patent examiner sends your patent back with questions.

Attorneys are nice, they provide a service which takes a lot of worry out of things and some of them are pretty aggressive at getting the patent issued even if the claims whittle down to something fairly obscure but they do charge by the hour. The last time I saw the bill for doing the patent paperwork it was about $12,000 total.

[1] http://www.nolo.com/products/patent-it-yourself-PAT.html

[2] http://sunnyvale.ca.gov/Departments/SunnyvalePublicLibrary/R...


useful info, thanks, though even a $2500 could be a problem for a startup, not even $12000




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: