Nowmov (YC W10): Sit Back, Relax, And Watch An Endless Stream of Videos | Hacker News Reader

Nowmov (YC W10): Sit Back, Relax, And Watch An Endless Stream of Videos

techcrunch.com

72 pointsthomaspun65 comments

I found one funny video and a bunch of random spanish stuff and a lot of pop music. Not sure I see what this has besides some novelty, sorry guys.

I watched it for 10 minutes, which consisted of about 30 clips. 70% of them were bad quality music videos, the rest uninteresting other content, partly in spanish. The "best" one that came across was the recent Chrome Ad. Even voting down every music video just caused more of them to come up. Sometimes the videos had no sound for some reason.

Yeah this is nothing more than a simple technological stunt. What's the value in this, and to whom? Just because a bunch of random clicks on Twitter or Youtube think something is popular doesn't equate to meaningfulness to me.

We agree, and we're working on making the streams personal and more meaningful. But Twitter popularity is an interesting proxy for overall popularity, and we've seen some stuff bubble up much faster than other video sites. It's a work in progress.

It's Portuguese, not Spanish.

I don't mean to be too discouraging, but I really don't understand the point of this as a business. It seems like something that anybody could code in about five minutes, and YouTube itself has started to add playlist functionality.

Recommendations is an 11-figure business. Seriously. This is a pretty reasonable angle of approach, so I can understand that part.

What I have trouble with is what chaosmachine points out: the recommendations are not that good. At least from a cold start.

How does this have 18 votes in 16 minutes? At a time when the site is pretty quiet? That is very suspicious.

From other YCers one would guess ? - maybe someone did a shout out at the scribd party or something to get upvotes.

I get that recommendations are a big deal. But YouTube itself already DOES have recommendations, on the home page if you're logged in!

If a service like this catches on, YouTube could replace it overnight, especially as they have data on what you (and the people you correspond with, via Google) enjoy watching.

The Pandora analogy has been thrown around a fair deal, and it is flawed. Pandora does not play and recommend content that is owned and managed by exactly one company.

>How does this have 18 votes in 16 minutes? At a time when the site is pretty quiet? That is very suspicious.

I'm sure it's nothing nefarious. But a good number of people who use this site have strong affinity for YC startups either because they're YC alums or just they're people who really like YC startups and feel like part of the YC community due to being members on this site and having watched YC grow.

I would guess that a subset of said people will basically upvote anything about a YC startup. And launches are especially prone to this treatment because people think that users of this site will find any YC launch interesting (which many users will).

I upvoted it because I worked on a similar idea with abossy and randomwalker. I agree with abossy's comment above: this is a great idea, I wish them all the best and I hope they kick ass, but the business side is difficult. So long as you rely on YouTube for videos, you are at the mercy of changes to the YouTube terms of service. You can look for other sources of videos besides YouTube, but this is challenging.

So I'm hoping they succeed where I failed. Hence the upvote.

Can you name one recommendation service that is an 11-figure business? The only company I can think of whose core business is mostly recommendation is Pandora, and they're not even that large; nine figures at best.

One could make a pretty strong argument that Amazon is at its core a recommendation service (with really good logistics that take care of the rest of the whole "oh, I just bought this thing I didn't know I wanted" process).

I also can't figure out why in the world would anybody use this - neither me, nor anybody I know, regardless of age/sex/other characteristics.

But then again, I said the same thing about Twitter when it came out and I was wrong.

YouTube could also easily replicate this idea and probably do much better recommendations since they have more resources, more data on the users and videos and more people that are experts in ranking algorithms.

Their business model might be being bought by Google :)

What business model?

I pursued this idea with randomwalker for about 6 months before realizing there's absolutely no business value here. My advice to Nowmov: choose another market before it's too late. We did, and it's been going much better for us. Fortunately, we didn't come across a strong celebrity endorsement like Ashton Kutcher's, so there was little cost to switching gears.

The problem with video recommendations is content ownership. YouTube has a monopoly on user-generated content so they immediately became our number one risk. Our business was henceforth bound to their ToS. Shortly after we switched gears, the Totlol debacle surfaced on Hacker News, confirming our suspicions that the risk was too great for a business to tolerate.

The main obstacle with the YouTube ToS is that they don't allow advertising on pages where the video content is the focal point of the page. Nowmov clearly focuses on videos, so the 'get-big-and-strike-it-rich' model is not an option for them. Moving beyond YouTube -- say, using licensed content -- cripples their product, because there simply isn't enough content to create a compelling recommendation engine. Even Hulu, being backed by industry giants, is bleeding content providers.

I can't emphasize this enough: this is a cool product, but not a good business. We had cool algorithms that provided fantastic recommendations and kept users on our alpha site for an average of 13 minutes before adding any type of social features. "Pandora for videos" was dead simple to pitch, too. I wish NowMov the best of luck and I hope they can succeed where we failed. The world needs a product like this.

Look at it from the other direction. In 5-10 years, there will be a lot of really successful businesses made as the line between TV and internet blurs. Who knows what those businesses will look like, but I'd bet that a few smart hackers who are able to build and iterate quickly will figure it out before Comcast - and maybe even before Google.

Sometimes I'm more interested in knowing the thought process of PG and YC about why they decided to invest in such idea. What they see this as a business opportunity? Or technology opportunity? Or just big market opportunity? What is it that made PG and YC to invest in this?

And this is just out of curiosity - to learn what they have seen that I'm missing and cannot see.

Keep a few things in mind.

First, a lot of companies release their long-term vision in stages. Think of Bump: their first release was a way to share contacts on mobile devices. Which is kind of cool, but maybe not IPO-worthy. Now they're doing mobile payments with PayPal, and I can only imagine they have bigger things coming.

Second, YC invests in founders, and the Nowmov founders are amazing.

It seemed to me a pretty good hypothesis about what the future of TV would look like. Roughly Nowmov has the same relation to watching TV that Reddit does to a newspaper or magazine. The format's similar, but instead of getting what one media co chooses to pipe down the channel at you, you see a collection of smaller things from all sorts of different sources (both big media cos and individual people), and the best stuff floats to the top.

Nowmov is not very good yet at finding the best stuff, obviously, but they only just launched. You have to imagine what it looks like when it can customize per user.

Fair enough. I totally believe that don't judge startups from where they are now, but judge them based on where they will be after few years.

I just had a different experience when I met you the other day at Chirp and demoed our app. It was a 4 month app developed in the part-time, has a decent traction, and growing at decent rate, but with great potential in future. I was more interested in explaining you my vision and getting your feedback on it, but it seemed like you gave lot of emphasis on its current usefulness and didn't seem to imagine how great and useful the app would be in few months.

So I was just curious, how your thought process works. But I'm glad that you seem to believe in the approach of what we potentially can achieve based on what we have achieved so far.

Thanks for your reply!

Here in the uk there is a semi popular tv show called rudetube; which shows Internet video clips based on a rough user ranking metric (YouTube views). So the idea could have legs.

Sure, I could write it off as the: omg where is the fucking business model this is such a toy on what is only iteration1! I'm not going to. There are smart people involved and im sure this is a small piece to a bigger puzzle.

Here's what interests me: it makes video online more like tv. ive been reading mark cuban's articles and some others. i want a more tv like experience ie- things start playing instantly as i navigate.

good job guys and keep up the good work.

We explicitly want to mimic the TV experience; we think that the page 'o videos + ad links is a sub-optimal way to watch user generated content. Think about how much video is uploaded to the Tube every day -- something like 24 hrs per minute -- and think about the discovery problem that presents. TV + discovery is what we're aiming for.

It's true that discovery can be a problem. It's interesting though to tackle it by trying to emulate TV, the thing that that online video has been trying to move away from as a selling point.

It gave me 4 spanish language videos in a row, despite thumbs-downing them.

Me too. A filter based on language selection would be nice.

It's work in progress. We had one in there, but it wasn't reliable enough.

Because we rank videos based on twitter mentions, we're pretty time-zone dependent. Currently it's brazil o'clock. Smoothing that out is on the to-do list.

The rec engine doesn't tailor the playlist yet - currently we're just using it to collect data.

I actually liked it. I watched for an hour straight. I think the comments are a tad biased because the folks who liked it probably didn't come back to comment immediately. I am pretty happy with the variety of stuff nowmov showed me. The recommendation algo could use a bit of work. Keep at it.

Thx. Glad u like it. We are working hard on the recommendation engine.

I'm sorry, but your recommendation algorithm just plain sucks.

It's so basic and ridiculous of an idea to just take the most popular twitter links and stream them one after the other. Recommendation Engines isn't a joke and isn't anything even remotely easy (see the netflix challenge). There are a million factors including age, location, culture, topic, timing, etc. and your engine (as far as I can tell) fails to take any of these into consideration even when being told through the thumbs-up/down buttons.

Sorry if I came across as being very harsh, it's just that this is a good idea very, very poorly executed. I hate to see this idea bombing because you're underestimating its complexities.

The data this company will be able to collect at scale could be valuable in and of itself.

That's true of any company; it's a matter of reaching scale in the first place.

It's a step by step process :)

I was just thinking of buying cable TV today because I never feel like internet media allows me to just turn my brain off and relax. This product seems like a great solution. I don't think these guys need to emphasize personalized/customized content, but rather they need to find content with universal appeal. In other words stuff that is tolerable enough to watch but not too interesting for any particular user, because interesting TV requires too much brain power.

Have you tried netflix?

welp, i'm back after using nowmov. thanks to nowmov i found one of the coolest youtube videos i've seen in a while http://www.youtube.com/watch?v=nCgQDjiotG0 Google Chrome speed tests

"Interesting sidenote: Ashton Kutcher is actually directly responsible for this site existing; the Nowmov guys were considering working on another idea until Kutcher told Y Combinator founders Paul Graham and Jessica Livingston that he wanted something like this. Kutcher decided to invest in and advise the startup, and Nowmov became a reality."

It's the kiss of death.

Does Kutcher have enough of a track record to compare him to the cover of Sports Illustrated at this point?

I enjoy how it feels similar to chatroulette. The 'next' feeling allows you to quickly skip through unwanted videos. What felt intuitive and I was disappointed was not implemented was up and down arrows to give a thumbs up or thumbs down. Another feature I felt lacking was the ability to full screen. A recommendation system could dramatically improve the quality. Perhaps based on the amount of time before the 'next' button is clicked. I see this platform as being addictive, easy to monetize and should quickly build a fanbase.

After a bit of digging I found some good videos. An informative video of liquid mountaineering, the art of walking on water, had me chuckling.

Thanks for the kind words! We're working feverishly on features, some of which are mentioned here and otherwhere on the page. Please keep checking in (or follow @nowmov on Twitter).

I can't help but notice the lack of YouTube ads. Are they being stripped out or something?

We drop everything that isn't embeddable without ads.

As far as I can tell you are using the chromeless player. I've been using this for two years now at Totlol.

Can you elaborate at a technical level what is it exactly that you "drop"?

Yep, it's the chromeless player. We grab videos links from twitter, then we use the youtube data api to drop videos where the embed access control is disabled.

Ah. That has nothing to do with ads.

I love this. I am coding on my Mac with Nowmov playing on my other laptop. Love how I can use keyboard to control.

I don't know about your algo, so it would be a nice addition to let people have accounts and then guess videos based on what they liked before, it will eliminate the need to touch keyboard to skip videos, overtime. :)

Unless the team is really smart and very good at various recommendation algorithms and mathematical tricks(which is not doing so well now),I don't see them going anywhere. Everyone can pull out the videos overnight using the Youtube API but only everyone can write a good filtering algorithm to decide what to show.

Recommendation algorithms are not that tricky, anyone can get 90% of the best algorithm in a couple of weeks. (OK, maybe not anyone, but my friend at Google did it for fun to compete in Netflix competition.)

The thing recommendation engines need is data and lots of it, which Youtube has in spades and these guys have to start from scratch.

Meta-ness is a disease plaguing the web. That aside I wonder how they are going to monetize it?

I could see this for parties, as in, you hit play and all it spits back is Pop and Techno and Raggaeton for hours on end. That would be really nice. I would pay $10 for four-seven hours of it if it did that well.

Very cool. I thought about pursuing an idea similar to this after spending nearly 2 hours (it felt like 10 minutes...) on http://wimp.com the other day.

without the ability to like or dislike a video and to personalize the stream, this is essentially useless for most people.

Pandora for video, but without personalized settings.

It's coming.

They also have trending channels (click on the tv icon near the play button) - which don't seem to be working for me. When I click nothing happens. Prob an MVP thing.

Channels should be working. You should get a box to the left of the player with a list. They're grabbed from the trending topics on twitter. I'll look into it.

There is a glitch though. I paused a video and when I resumed it started playing a totally different video, the video I paused is not even in the clip history.

This is possibly one of the best sites ever made.

have you seen www.yawtv.com , its similar but there you can fire a search and play the videos i.e to say if you search for 'lady gaga' or 'michael jackson' you can play all the songs present on youtube non stop and instantly.

I think people are missing the thumbs up/down. This is more like pandora for videos.

I just lost another 15 minutes of my life to NowMov. So addicting.

Time to announce the strategic switch to HTML5. =)

http://www.upl8.tv/

circa 2008.

Reply on news.ycombinator.com