Hacker News Reader

Comment by AndrewKemendo | Hacker News Reader

The engineering community is sleeping on Snapchat, but shouldn't be.

My life is AR. I'm the CEO for an AR company, I speak write and live and breathe all things AR.

Snapchat is poised to be the most important AR company around for several reasons. 1. They have a MASSIVE userbase that is comfortable with sharing their lives 2. That userbase has already shown an interest in AR - and yes whether you like it or not their filters are on the AR spectrum - the filters system maps, tracks (SLAM) and modifies in real time.

Their Spectacles release has the potential to build the largest real world image dataset on the planet - something needed (for reasons I won't go into) for AR everywhere. Think on the order of 1.8 Billion images per month with depth data from the stereo cameras (30FPS * 10 sec snap * estimated 20 snaps per day * estimate 100k spectacles users in first month).

Most people have no clue what is going on in the AR world, and I can tell you Snapchat is going to be formidable. If they can invest in a VRD, then they could win the whole AR market, but they need to raise more than 4BN to do that probably.

View on HN

I'm inclined to agree. I was very bearish on Snapchat around the time they rejected the Facebook offer. But more than any other app right now, Snapchat is killing it with users. This is a what a product users love looks like.

On top of that, I've been consistently impressed with their technological prowess, even if the AR component seems somewhat inconsequential to me right now.

You better believe Snapchat is gunning for that top spot in social - that's what the IPO is about. If they're at or surpassed Facebook's level by 2020, I seriously would not be surprised.

I'm bullish too, but any comparisons to Facebook are lazy. Snapchat and Facebook are completely orthogonal (the closer competition is Instagram, owned by FB, and Snapchat, but even then I would argue it is lazy to say they are direct competitors, ignoring the Instagram Stories product).

For better or worse Facebook is a piece of social infrastructure and utility. Everyone has Facebook. All the groups I know of are organized through Facebook. All of the events. And so on.

Yes but money is made through eyeballs and not infrastructure. The most money will go to the app that gets the most viewing.

> The most money will go to the app that gets the most viewing.

...and figures out how to draw consistent revenue from that. Snapchat is making $365m this year in revenue, projecting $1b next year, and has a newly launched ads API.

Completely anecdotal, but my siblings and cousins that are teens don't have Facebook, and most of their classmates don't either. It's weird to imagine a world where not everyone has Facebook, but I don't think it's as ubiquitous for those who grew up where alternatives existed, such as Snapchat and Instagram (even Twitter as well).

As teenagers with close friend circles, that's becoming more common. I wouldn't be surprised to see them pick up and get more involved with Facebook if/when they enter a larger community at college, or move to a new city, etc.

I have been facebook clean for almost 3 years. Never felt better!

>I've been consistently impressed with their technological prowess

Like what, the glasses or the apps? Would be interested to learn what they've done so well with on the software side.

I attribute their success to product design and knowing their market, but hadn't really thought of it as technical acumen.

You're trying to convince yourself very hard, as you're all about AR. But nobody will wear those dumb glasses.

Famous last words. Reminds me of the old adage: "640k of memory is all that anybody with a computer would ever need."

I already know a lot of younger folks that are pumped about the Snapchat glasses...

(I'm not saying you're wrong... I just wouldn't make such steadfast predictions about the future either. But then again, I don't understand SnapChat and am not their target audience.)

You can trot out that 640k comment to sell any piece of technology, it doesn't make it true. I could have said the same about Google Glass.

Indeed you could. And sometimes you'd be right; other times you'd be dead wrong. If you could be right even half the time, you'd be rich as all get out. Which is why I would suggest: Don't make overreaching bold predictions (or become an investor).

I'm right way more than half the time. How do I get rich predicting $newTech will fail?

Short the stock :)

Go for it flukus! Easy peasy! If what you say is true, you can probably become a millionaire before the decade is out!

and be ready to pay for as long as you're wrong.

Being right about half of failures isn't worth anything. You can't short SnapChat Spectacles.

You need to be right about half of successes, which is much harder for a raging pessimist like OP.

The problem with the Snapchat glasses is their re usability and if it catches on.

If your friends aren't using them and you're the only one with a pair you'd be less inclined to wear them. Everyone thought Google Glass was going to be the future but those glasses made you look like a total douche. Hopefully the Spectacles wont be branded as such.

I would posit: your perception of "total douche-ness" (your words, not mine) are largely driven by marketing and/or culture.

Glass definitely failed in the marketing, branding, and roll-out departments (among others) -- but they're no more or less doofy (to me) than many other perfectly-accepted fashions: baggy pants, beats headphones, crocs, etc.

>(to me)

That is a fairly small focus group.

Not sure network effects are necessary there. It's fine if one person makes videos with glasses as long as friends enjoy watching them.

If the glasses are promoted as something you only wear when you want to record video and that you otherwise put away it might take off.

Nobody minds when you pull out a cell phone and record video because it's obvious you're filming. I think it was the omnipresent threat of being secretly recorded that made Glass creepy.

Those glasses are designed to be the party Polaroid of a generation. The person holding the camera gets the attention of everybody keen on presenting themselves, the bet is that putting the camera on your face intensifies this experience. Expect these things to get passed around a bit when in use, so there will quite some local virality within groups. This is good for immediate sales, but long term usage could suffer hard if market saturation coincides with the novelty wearing off. If there is one Polaroid at a given party, it is the center of attention, if there are three of them, they are all just annoying. The me-too buyers well get quite the opposite result compared to the early adopters.

If Snapchat sets up the glasses business as an additional revenue stream, it could well become a reasonable but limited success, if they set it up "strategically", maybe even selling at loss, I predict terrible failure.

No, young people will not stand in line to buy or wear those glasses. Snap chat is very popular. However, the current user base will outgrow it and those younger will have something else. Better get that 4B while they can. It doesn't have the roots or staying power of a Facebook or Twitter (I know Twitter isn't making money).

I agree with both of you. The glasses are silly, don't seem that technically complex, and I think they are poised to be a huge hit with Snapchat's base audience. Google Glass was a technically well made device that forgot product, I think those stupid glasses hit their product-market fit quite well. The recording indicator and their 'retro' style are both what I consider to be key components in this assessment.

Lots of people are going to wear those glasses. Is there some site where you can bet small amounts of $ with strangers on the internet? :)

You mean like a prediction market? Yes, hundreds.

Nah, I mean like I want to wager with simooooo. I'm 25% serious.

http://longbets.org/

Depending on the strike price, can't you just buy or short a few shares?

I mean that's a bet with a payoff & loss.

And nobody would carry Palm Pilot PDAs. Well they didn't wind up carrying them from Palm, but Apple and Samsung sure have something awfully close. Snap may or may not be the AR disruptor, but AR will be the new smartphone. Whoever owns it will be the new Apple.

But before Palm Pilots, there was the Newton so loved by techies, just like Glass. So snap may end up the palm pilot or one of the dozen PDA knockoffs that were around (I seem to remember some Microsoft ones too). It seems like concepts float around in several iterations, or as many as it takes, until someone hits the sweet spot.

I'm not really knowledgeable about the "wearable" market, but to me it doesn't seem as ubiquitous as the smart phone.

If they license the technology to premium brands, say Prada, Calvin Klein etc. everyone will buy them. Especially their target group which is primarily youngsters.

Those "dumb glasses" are all I hear high school kids talking about when I volunteer. If they hit a cheap enough price point with them, kids are going to wear them everywhere. There'll be plenty of backlash as schools/etc ban them, but they're targeting a market that's already crazy about them.

Sounds like twitter: interesting, perhaps the future, but not a business worth investment. How can they succeed where twitter failed? The ads have ruined the app. Everyone I know who uses it agrees it performs terribly and is extremely difficult to avoid unwanted, irrelevant ads.

Again: it's twitter.

It's a better case than twitter though. There is no obvious way of monetizing twitter, while there are many ways that I can think of that snapchat can monetized. For example, Line and Kakao have massively succeeded with their sticker/character business, and snapchat could bring something similar. There are many other ways I can list. So, the issue is not "What are possible methods of monetization?" but "How do I optimize profit?" This is definitely a more favorable setting than twitter.

So 4B for a selling stickers business model?

Kakao's business model history is actually rather fascinating, and they keep grabbing sizeable market shares in unlikely or previously inexpressive markets to great success.

> In 2012, KakaoTalk's $42 million revenue is broken down to 67.5% ($31.1M) gaming, 26.2% ($12.1M) advertising, and 6.3% ($2.8M) emoticon sales. [1]

[1] https://en.wikipedia.org/wiki/KakaoTalk#Company_Business_Mod...

$200MM rev is crumbs when you're taking billions. I don't think this provides a good example of how it can work.

"Their Spectacles release has the potential to build the largest real world image dataset on the planet" With Google (and now Uber) cars driving around numerous places to capture data specifically for map imagery, satellite imagery and Android/iOS phones taking tons of photos + videos I highly doubt Snap will be the one to build the world's largest image dataset. The rest of your argument makes sense though.

With only 100k users they would generate approximately 9BN images per month (30FPS min, 10sec Snaps, average 10 snaps per day).

With 1M users of spectacles (.6% of their current MAU), that's potentially 90BN images per month and over a trillion per year. In the same dataset. Machine vision wet dream.

There is no way the average user is shooting full 10s snaps 10x day. I use Snapchat regularly and that describes the top eschelon of users, which might be Spectacles users, but probably not all of them. I believe that is a vast overestimate.

As with most statistics it's likely normally distributed. I know teens send/receive in the hundreds per day - mind boggling really - and the celeb accounts post dozens daily. Even back in 2013 the numbers were staggering.

[1] http://www.businessinsider.com/how-many-snaps-snapchat-users...

Nitpick: I would strongly expect that the number of snaps sent/received per day is exponentially distributed over the space of users, not normally.

And how will they make money? Selling ads in the AR world?

How do facebook and Google make money? Selling ads in social network and search land.

Having an enormous, religiously active user base has proven time and time again to be a huge cash cow.

That was Twitter's playbook and I thought Twitter wasn't doing so well recently...I'm not saying massive amounts of users isn't valuable, but you still need to play it right.

Twitter is massively profitable. It makes billions in advertising.

(for every $1 twitter make, it spends $2, so their accounting is deep in the negative. But if they fire half the people (who are useless anyway) and stop spending money like madmen, they'll be positive, VERY POSITIVE. thanks to advertising revenues.)

Well I'm sure those are the reasons why Google pass on :)

They can't buy first and fire people afterwards. The cleansing would have dire effect on the alphabet parent company.

no need. Twitter announced it was firing ~20% of it's userbase.

Guess Google can swoop in and buy it now :)

Or they can repeat the process till they have 20% left ;)

Twitter is massively unprofitable. 12 months ending 2015-12-31: loss of over USD500MM. 3 months ending 2016-06-30: loss of USD107MM.

twitter is worth 11B and makes 2.6B/yr. The playbook works

Twitter loses over USD600MM / yr. The playbook is yet to be proven. If they were earning 2.6B / yr, news articles and market perception would be very, very different.

Google and Facebook capture user data across many dimensions, and so are able to present advertisers with highly targeted groups of prospects. I don't know if that's the case with Twitter or Snapchat. I would say that the more a site presents you with ads you have no interest in, the less likely it is that they have fine-grained profiles of their users (which is where the money is).

The way it stands, Snapchat already has more data points to tap from and richer content and interaction opportunities than traditional media (television, radio, print, OOH...), and this could be further improved in future versions. You don't need to set out to compete against Google and Facebook when the actual big fish in the market are in fact easier to position yourself against.

I'm not sure I understand what you mean by the actual big fish in the market. Google's Q2 16 ad revenue was $19.1B, NBC's was $2.1B, CBS's was $1.6B, and ABC's was also $1.6B.

I was thinking verticals as a whole, not individual companies. The budget for TV, print, radio and so forth tells the size of the market, the fact that there are more players in these verticals than in online search is maybe yet another reason to attack them rather than Google.

The fairy ad business model.

Don't know about Facebook (never been) but I keep hearing about this fabled Google Ad business and honestly, not only I can't recall the last time I even noticed one, needless to say have never even clicked on one. Some interested party wants to keep these surveillance cum social networks afloat and the fairy ad business model and overblown evaluations and VC cash is the front for funding them.

> I keep hearing about this fabled Google Ad business

The one that makes 19 billion dollars per quarter?

That's the cash flow, the hypothesized white washed flow of funds to select platforms under control of the said interested parties.

You can't use google much if you haven't noticed the top three results on your searches sometimes being tagged with the word "sponsored". Unless the reason it's showing here, is an European law thing?

They say it here in America. But uBlock blocks it :)

I've been using it since pretty much since it came out. In the early days I definitely noticed the ads on the side, but very quickly it pretty much got filtered away in my eyes along with window decorations, etc.

Google ads are something we only see when searching on google. They're easy to tolerate.

No one will put up with virtual ads in AR.

I'm not so sure. Even with their rudimentary pre-AR camera filters, some were pretty popular and successful [1]. The Gatorade dunk filter generated 160 million impressions and the company was very pleased [2] -- and people seemed to like it.

The filters are also interesting because they involve user choice: the ads are not unsolicited; rather, they are picked by the users. That's pretty interesting.

We've had in-game advertising in video games for a while; certain games that mimic realistic urban environments lend themselves well to this. In some settings, AR could include a density of advertising -- passive or interactive -- similar to the real world without much trouble. In fact, an AR setting devoid of any commercial presence may be unsettling on its own.

[1] https://medium.com/comms-planning/11-branded-snapchat-filter...

[2] http://digiday.com/brands/inside-gatorades-digital-ad-playbo...

> In some settings, AR could include a density of advertising -- passive or interactive -- similar to the real world without much trouble. In fact, an AR setting devoid of any commercial presence may be unsettling on its own.

I'm going to use it to block IRL advertising, not replace it.

Adblock made an April Fool's video about this!

>In some settings, AR could include a density of advertising -- passive or interactive -- similar to the real world without much trouble.

Occasionally I'll watch an NHL hockey game and on certain broadcasts I started to notice large banners on the glass behind the goals and thought to myself, "Boy, you pay for seats that close to the ice and you're looking at the back of a banner ad?"

When I noticed the ad changed during the broadcast I realized that it was a virtual projection for the TV audience[0] -- a sort of AR in its own right, since it appears so seamless that I originally thought it was part of the rink.

[0] http://www.sportvision.com/hockey/virtual-advertisements

Snap chat inserts sponsored content into your feed. The trick is the content is engaging and heavily tailored to their demographics. Lots of celeb gossip, sports news, random stuff you'd normally see in magazine articles. There's certainly a distinct, high potential business opportunity there.

https://www.spectacles.com/

Am I the only one bothered by the fact that their models are topless?

I can't help but feel they are signalling their intent to be an eye into my private life, not unlike the monitoring predicted in 1984.

Just you

They could be wearing tops with no shoulders...

My intuition is that these will be a hit. Simple, fun, youthful & on trend, not fussy to charge, huge captive userbase to sell to. Can you turn them around and use the mirrored lenses to take a selfie?

Snap probably has a winner here if their users will further embrace the idea of turning cameras away from their faces and onto their friends and experiences. How big an "if" that is, I can't yet say.

I think they're too trendy looking. They'll be fine for a season/uni semester, but after that will drop out of fashion. And they'll be too pricey for their core demographic to keep buying new ones.

There are too many options for monetization in AR to spell out, or to narrow down exactly which ones Snap will utilize, but just assume if you can make money on the web now, you'll be able to make money in AR in a similar way.

http://www.businessinsider.com/snapchats-sponsored-lenses-ar...

Man seeing those peoples faces plastered in ads is positively dystopic.

My initial reaction was the same but we already have people walking around wearing or carrying items with corporate branding (sports apparel, well known clothing brands, iPhones etc)

True, I'm not opposed to branding or designer wear, it's just something about seeing a face just covered in them and then sharing it...

> It’s like wearing a sweatshirt with nothing but the word “Nike” on it

Which people currently PAY for the privilege of doing. I agree that the face feels a little more identity-usurping though.

The ad makers are people too, for now. One would assume they would get wise of the stigma of having ads plastered on your face and would try to create commercials in different styles and formats that are more subtle/interesting than the banner ads we're all accustomed to.

That's horrific, you'd hope the novelty wears off quickly.

Snapchat already makes like 700 million in revenue. (edit: jk they hope to make 300 million and then hope to do much more than that by 2017)

The real question is "do all these companies buying ads actually make a return on investment" and the answer is DOESN'T MATTER! Lots of ad dollars still to move away from older methods into digital media lol

Maybe there will be a Netflix documentary in the year 2027 about what caused the burst of the collateralized ad market and financial crises of the roaring 20s, with the undertone of "why didn't anybody see this coming"

but in 2017, that absolutely does not matter.

I mostly agree with you, but you're off by about an order of magnitude (±0.5) on Snap's revenue numbers. They did about $60 million in 2015, and were projecting 4x that for this year.

Source: http://www.bloomberg.com/news/articles/2016-05-26/snapchat-r...

whoops

yeah they're extrapolating to do 4x of this year's theoretical revenue numbers in 2017

I'm not even mad, this is amazing

I sure as hell am tired of all these damned ads. I will pay for a service to avoid myself being the item that is sold. How large is the population of users like me? Probably not very large. Also, privacy minded people seem to have a disdain for social media.

Is the reason AR benefits from billions of photos of the real world a secret, or is there something you can point to to read about that? I'm intrigued :)

Also, you probably know better than me, but to me it looks like the beauty filters are the ones that have really stuck for Snapchat. If I'm right and their users only really care about filters that make them look better, does that affect your outlook at all regarding their users showing an interest in AR?

The simple answer is that you can make very large 3D maps with those images.

I disagree with your second assumption, though I can see where it comes from. It's not so much the specific implementation, beauty or whimsy, but rather that people will take the action of modifying their real time view consistently.

Andrew I think you're too optimistic about spectacles given no one has even proven it's a viable market, let alone that Snapchat will have a successful implementation.

What's your best guess of how many pair will sell through to actual customers?

At a minimum 100k, which is enough to start proving the market out and getting some virtuous cycles building.

They are not the only player, but they probably have the most accepting user base. They don't have the largest user base amongst the AR players though.

*they don't have the larger user base amongst AR players though"

Yes they do, by far. The next biggest pure AR player (by mau) is arguably blippar with about 8M. Metaio had a larger install base across all products, but they are gone.

I guess if you count Google translate they would eclipse test in their but I don't know how many people are using word view translate implementation.

I don't consider Pokemon go a pure AR play and even then their at numbers are hard to come by as a significant population used it without AR

Don't leave us hanging like that - who in the AR world has the largest user base?

There is only one camera on the Spectacles. No stereo.

You appear to be right. For some reason I thought I saw a stereo version. Missed opportunity there in that case. I still stand by my assertion though - as you don't need stereo for depth but it does make it faster and more robustly recovered.

I used to love Snapchat, it felt like the future of communication in many ways. Unfortunately it seems that their Android app became completely unusable on mid-tier devices in the past year. Painfully slow UX and unreliable in most situations. Wish they'd do something about this as it forced me to abandon the platform entirely.

Their spectacles is going to be a game changer. I hope it does well (buy also part of me hopes it does not for the sake of everyone's privacy in public).

https://www.spectacles.com/

No, they are not. It's gonna be another gimmick lifestyle tech product that gets trumped up in the press and then no one cares about. They are glasses that only take ten seconds video and can only post it to snapchat. And they cost $130. Come on.

No WiFi. Less space than a nomad. Lame.

Lol people post this everytime someone critiques a hyped up new bullshit tech product. Guess what, the vast majority of them fail. Spectacles is not the iPod. It's going to flop, I stand by it.

Your analysis that it is a "gimmick", "bullshit", and "hyped up" is over-simplified and, I think, wrong. There's no need to swear or be mean about this. Some people, myself included (I hate hate hate hate the snapchat app, I'm not able to use it and I can't stand the idea of it) think that Snap does understand product and its market, and has incredible visions for the future with an ability to execute. I think that Snap will be a huge success for a long time to come.

Naysaying Snap is fine, but not if your argument is "bullshit hype".

His argument was cost and lack of utility...

Bullshit hype IS an argument in and of itself and if we can't speak plainly how can we speak honestly?

$129 is pretty cheap nowadays, you can get a Google Home for that price.

Snapchat is targeting teens and young adults with this - not exactly people who have $130 lying around

Did not stop Gopro. If you have been to a ski resort in recent years you have probably noticed that at some point, Gopro saturation started to become inversely proportional to riding skill. Where Gopros are sold with the promise of being like those who get really fast and have spectacularly high jumps, Snapchat specs will be sold with the promise of being like those who get really drunk and have spectacularly great parties. That is a very promising market to tap into, but no way that it could ever be a four billion market.

There are important differences between GoPros and Spectacles, though.

GoPros are a convenient, high-quality, expandable memory camera that creates a regular video file as big as the storage space available that you can use however you want.

Spectacles are a convenient, unknown quality camera creating a ten-second video that requires smartphone interaction, is designed to be on your face, and is tied to one specific (albeit popular) platform.

Aside from both being able to take video and being easy enough to carry around, the products and the target markets don't intersect perfectly, so it's hard to predict one's success from the other.

I base my comparison on the observation that both are purpose built for capturing a very specific kind of experience that few have, but many think they should have (but never will).

This is certainly no predictor of absolute success, just an element of understanding the potential market.

But they do all seem to have $500+ phones already. I don't think it's a stretch to assume that they could afford the glasses if they deem it worth it.

I hope not.

I really dislike the asymmetry between the ability to take pictures of anyone in public easily (assuming spectacles or an equivalent comes out), with low chance of detection (unlike cameras/phones, which are harder to disguise) while being unable to as easily wear a mask in public to preserve some semblance of anonymity. Anti-mask laws exist for a variety of reasons, some of which I am fine with (e.g near banks) and some of which I am not (e.g abuse to shut down things like the Occupy movement). I would like to see a similar treatment of devices like this (spectacles, google glass, or whatever other product people work on in this space).

They should be treated on par with hidden cameras and the like. Until legislation takes care of this, I am not happy with the situation.

They are at least trying to avoid the Google Glass fiasco by adding recording indicator lights to the front.

And I bet we'll be seeing our first Spectacles-based-sex-tape scandal within a year, once someone figures out how to short-circuit the indicator light. ;)

Google Glass was meant to be worn all the time, people were wearing them to the bar and stuff. Spectacles are pretty easy to put in your pocket. For the most part kids will be wearing Spectacles in situations where they would otherwise be walking around with a camera in their hand, ready to shoot stuff.

Snapchat spectacles are specifically designed to transfer the "everybody smile for the camera" effect to the wearer. The absolute opposite of a disguised camera.

Hidden cameras (not terribly well hidden cameras, but hidden nonetheless) are nothing new, the Snapchat innovation is to reverse the camouflage to create completely different usage patterns.

If the Snapchat specs become popular there will actually be a knockoff market for fakes that don't even have a camera, because imitating just the recording status lights will be enough to get the "say cheese" effect.

Yet another example of a massive amount of user collected and provided data that is only available to one company.

I wish that there were more open data laws. A lot of startups would benefit from having the data volumes that more massive companies have.

Can you elaborate a bit on how they use SLAM? I was my (mis?)understanding that SLAM required a pre-existing map. If this is true what are the source of the maps? Are the maps built from previous users's snaps?

But don't you think that because AR is just starting to have a place in our mobile phones it is open to a lot of newcomers? This is not a mature market sector like social networks, ads, or search.

What is AR?

augmented reality. https://en.m.wikipedia.org/wiki/Augmented_reality

Their filters do SLAM?

Not in the robotics sense where you map out a hallway of a building or something. But it does map facial landmarks very well (and I suppose that will also give you localization relative to the face).

No, it does face recognition and orientation, it's far simpler than slam and have nothing to do with it.

Yes, but a narrower interpretation than used in more robust systems like those that we built. They don't need an entire SLAM system with loop closure and relocalization because you can mimic those with the simple facial recognition portions. However they added orientation functionality but without visual odometry. Functionally it works the same, and is the easiest way to describe it.

what does AR, SLAM, and VRD stand for?

AR: Augmented Reality SLAM: Simultaneous Localisation and Mapping VRD: Virtual Retinal Display

Sorry... I tried googling this but what is AR?

Augmented Reality.

Thanks - why the down votes on this?

I don't know, I didn't downvote anyone.

Reply on news.ycombinator.com