Ep. #11, Hybrid Broadband TV with Jeremy Brown

Matt McClure: Hey, everybody. This is the first time we've actually been back in the studio within two weeks from the last episode that we recorded. We're on a roll. 2019 is off to a roaring start.

Today we have Jeremy all the way from Australia with us, and Phil. We have two international people in the studio today, which is awesome.

Jeremy Brown: And three video meetup organizers.

Steve Heffernan: Yes. SF Video, London Video Technology, and Jeremy runs Sydney Video Tech.

Jeremy: That's correct. Thank you. Thank you for having me.

Matt: Nice. Some normal things that we'll run through, I will continue to mention this in perpetuity forever until October. Demuxed is officially announced. It's the-- Well, I don't know if we've even technically announced it.

But you've probably heard rumors. October 23rd and 24th, it will be announced by the time that this is released. We've opened up sponsorships, so if you're interested sponsoring shoot us an email Sponsor@Demuxed.com.

Otherwise, we'll see you there. OK. Now that stuff's out of the way, Jeremy, why don't you introduce yourself?

Jeremy: Yeah. I'm Jeremy Brown and I'm head of video architecture in a company called Seven West Media. We are a broadcaster in Australia. We own a bunch of magazines, newspapers and media companies.

But the meat and potatoes of the company would be a product called Seven Plus which is a catch up and live streaming service. During major sporting events like Olympics we might run an extra 50 channels on top of our BAU live channels.

Lots of ad replacement, lots of devices we support. Which we'll get into today.

Phil Cluff: Cool. You're from Seven. One of my favorite things about the Australian market is every television channel is the number.

Jeremy: There's two government subsidized channels, ABC and SBS, but everyone else is pretty much a number until you go to regional areas, and then it gets all messy and you'll find channels with a bit of content from every channel mixed in. It is a very different market. I get that.

Phil: What's interesting to us is, in talking about what the differences are, I've worked with you for a long time.

Jeremy: Right.

Phil: We've been deep into that market of things that happen and technologies that exist in Australia, and situations that exist in Australia, and also New Zealand that don't exist out in America or Europe and those sorts of things.

Let's talk a bit about, connected TVs and smart devices, because I think that space is really different out there.

Jeremy: What we're finding as a broadcaster catch up service that the connected TV is actually our biggest audience. We're over 60%, we're aiming for 70% on connected TV.

We've got a flavor of Roku called Telstra TV in the market that's very popular. We've got Samsungs and PlayStations and some have Xboxes, but probably the interesting one that you want to get on to is what we call Free View Plus which is HB TV.

Phil: HB TV is hybrid broadband baseband? Is it the other way around?

Jeremy: Yes. The version we're running right now is 1.5, which is the C HTML spec. That was originally written for refrigerators.

Matt: Is that a joke, or is that serious?

Jeremy: No, I think it is. So for all its quirks, which I'm sure Phil's worked on in previous life, it's actually a pretty interesting project to work on. One that all the broadcasters collaborate on together.

We created this consortium in Free View Plus, which is all the broadcasters are allowed to participate in.

We created an experience across all the channels where you have a catch up service for each channel that's activated by a broadcast signal.

Phil: Yes. So, coming from a technical perspective, if you think about how this is working we've got metadata going out in the transport streams that the content goes out through that can trigger and build applications and services to appear on the TV. Right?

Jeremy: Yeah.

I usually dumb it down and just go, "You've got your broadcast and we're running a web browser on top of it."

It's transparent, and you can decide how much of that you want to take over. From that experience you can see how you can have ad experiences, or catch up experiences, or enriched experiences.

Like the cable providers have been doing in their boxes, but every channel can build their own different experience. That's what's been happening.

Steve: How reliable is the HTML interface on top of this? Like, are they using some standard browser engine or is it a custom thing that you got to be careful about which browser features you try to use?

Phil: Take a wild guess, Steve.

Matt: "Does the marquee tag work?" is the real question.

Jeremy: The blink tag works, yes. It's only fair to say that it's been a lot of work and to try and keep it simple would be an understatement. Products like server side have been absolutely critical in getting this to work.

I think the original intent of it was in Germany where there's a lot of government owned broadcasters, and so commercialization of product wasn't as important. Whereas for us, being a commercial broadcaster that was the driving requirement.

But when you get through those, the video delivery has never really been the complicated bit. It's always been the refrigerator operating system that's running the browser.

What we are finding though is some of the modern TVs just have an HTML 5 browser right there, and you can sniff for it.

If you find it, then you use it.

Phil: What's super interesting from a delivery perspective is that HB TV, when you start thinking about delivering actual video content, it is a derivative dash.

It's an old version of the dash spec with quite a set of limitations on it, really one of the big ones is a single period. One of the ways of doing a dash of insertion is obviously multiple periods, but that doesn't work.

Jeremy: Again, to my point, there is enough devices out there that are smart and you can do a user agent's name for something like that. In fact even HLS, we found most of them, it's not written anywhere in the spec but it can do it.

But the big challenge is you're finding these browsers are very limited. They're playback very different. I liken it to the good old days of web development where each browser was different.

You just have to think about it that way, and you're running in iOS 2 or 3. So it's a very old JavaScript.

Matt: Does Babble even take you back to there, if you--?

Jeremy: Yeah, we've-- I think we're using web pack and Babble. So, that's one experience that we build. The other thing that we're doing collaborative is Austin which is a Nielsen byproduct. Its analytics, but its analytics across all the broadcasters and it's a publicly available.

Every day there's reports out on what show is the most popular in catch up, which channels are the biggest. That's a real game changer in our market because suddenly all numbers are considered equal.

We found that across the industry, someone will talk about minutes, or sessions, concurrence. That conversation stops. It's minutes, and that's the currency.

It's been playing against all the channels against each other, and one of the cable providers. I've learnt today, even an s-spot service from America is running it too.

Matt: Is this a-- Is it a requirement that you--?

Jeremy: The driver for it was minutes appended to broadcast minutes. So, the way that you see broadcast minutes is you see live minutes and you see the catch up minutes with people whose DVR may have recorded it, and that's being tracked by the Census previously.

This is more minutes you can add onto that, because we've all agreed that these are all equal.

Steve: What is the main value behind that? Is it understanding where advertising dollars should go? Is it like, knowing what's the most popular show?

Jeremy: Yeah. We've all, everyone in the industry, has accepted that the digital minutes are climbing. I mean, they are doubling if not every year. How do you turn those minutes into the traditional currency of minutes that you would see on a broadcast, like on a Nielsen census?

If you can add onto those minutes and create a total perspective across all of catch up and live streaming, as well as live broadcasts, you can create a currency that doesn't look so bleak for the broadcaster.

Matt: How are you guys doing those analytics?

Jeremy: It's a lot of work. Each player needs to be accredited, certified, and so there is people in West Media who are working very hard keeping all the players across all the platforms across all the broadcasters accredited and tested.

Matt: Are the players beaconing? This gets back to what I was asking earlier.

Jeremy: Beaconing every minute, and then ad events, starred events, complete events. There is quite a lot in there.

Matt: Are you required to implement that legally? Or is it just something that, "Why wouldn't you do it?"

Jeremy: It's probably more the latter. You need those minutes to feed in to a bigger thing.

If you've got devices without supporting it, you are just missing out on the greater picture of your audience, and that's publicly available. You just don't look as popular as you probably are.

Phil: Are those beacons caught by ad blockers in general, or not?

Jeremy: I don't know. I know that we are working very hard to keep PII and GDPR compliant, and I know they're working very hard on not being blacklisted. So, I don't think they are.

Phil: That's cool.

Steve: I imagine they can throw some money to ad blockers way too, maybe circumvent.

Phil: Let's talk a bit more about ads in Australia. It's different, right? It's not the same as it is everywhere else.

Jeremy: I have great conversations here with people in ad markets here, they talk about brands that just aren't that popular in Australia. But then they think I'm crazy, because we deal with some variance.

I'm not going to name them by name, but you get what I'm saying. There's definitely a different ecosystem in Australia.

It's interesting, I've been talking to another company who is in Africa, and their ecosystem is completely different as well. Brands and agencies that are popular there I wouldn't even thought they were popular in Australia or in America.

So, the ad tech experience is being that there is a small number of broadcasters, they are all catch up.

When it comes to premium advertising as a product, it's pretty easy to narrow down to who you need to be buying from.

That's probably changing the conversation away from the cat videos and the pretty girls.

Phil: From what I understand the CPM is a great out in Australia in comparison to here.

Jeremy: Yeah, and we noticed that when we use some of the American DSPs and SSPs, the rates we're getting are much higher. That comes with work, and relationships. There's a strong relationship between the DSPs we do work with and the ones with--.

The one's with local presence, especially technical, seem to be the ones that perform the best. Because they're making that investment into getting it to work. As I said, connected TVs are so high that you have to do server side really if you're going to build a platform with a small team.

Those teams that understand that and are taking time to work with those other quirks of server side advertising, it's paying off for them. It's paying off for us, too.

Steve: So, we're saying CPMs are higher in Australia. Is it just like that there's maybe better technical integration between the platforms and the ad vendors, or is there other unique details to the Australian market to make that possible?

Jeremy: It's a smaller marketplace.

Steve: Smaller?

Jeremy: You kind of know everyone, so you can work through these problems a little bit more, and better relationships I guess. America is a massive place by comparison. We're probably the size of Connecticut, that's the whole country by comparison. I think that helps.

We're based out of two major cities as a country, so that lends itself to better relationships and better targeting and better outcomes for each partner. I don't think it's a premium that we charge because we're Australian.

Steve: No, yeah. I would imagine it's something--

Jeremy: It's more a longer term relationship that works for both parties.

Steve: Yeah, and maybe targeting like you mentioned, you can target a little bit easier than you can probably target in the United States.

Jeremy: Yeah. I mean, we don't have GDPR yet. I'm sure it's coming.

Matt: Is it coming?

Jeremy: Well, I think it's coming everywhere, isn't it? I bet California laws are pretty strict. Saw them coming in out of nowhere, really. That was fast.

Steve: Might even just be knowing the audience better. Like, knowing what audience each channel has and be able to target your advertising dollars a little bit better.

Jeremy: That goes hand-in-hand with those relationships. They've been very direct in telling us what they want, what they require. Sign-in is something that we've had to roll out everywhere, and I think just about everyone's doing it now.

Steve: Sign-in?

Jeremy: Yeah. Getting basic data about demographic data and being able to target against that creates value.

Steve: Is that the viewers logging into your service?

Jeremy: Absolutely. Each service has their own sign-in platform, and that creates like an A/S/L, isn't it? Back in the--?

Phil: Those who remember chat rooms.

Matt: Right. Got it.

Steve: Yeah. Gotcha.

Jeremy: Then you can do demographic targeting, and with that comes a premium. Even location targeting, to a lesser extent, is becoming a priority.

Matt: Most of these platforms are free, though, right?

Jeremy: Yes, they're all free. We're all public broadcasters, so it's a catch up service typically.

Matt: Got it.

Phil: That's suddenly becoming more popular across England as well. Even BBC putting everything behind a sign-in wall now. It's already been on the majority of those catch up services, but the rest of them are now pretty much in alignment as well.

Jeremy: The other issue I see coming through a lot is for about five years now everyone's decided that JavaScript can solve everything.

Matt: It can't?

Jeremy: It's been every year, a new issue, whether it be fraud or ability. Every year that's a new big issue. Out of that they've gone, "Here's the JavaScript to solve that problem." Which is fine.

But as we've moved to the connected TV audience, well there is no JavaScript, and there are no plugins . Or, some of the attributes that you can get from JavaScript you can't extract out of a Roku TV. It just won't let you ask, "How big is the TV?"

So, there's an education there, and that's happening. I'm really excited to see where that goes. Open measurements, a funny thing that is getting stronger in growth and people are committing to it, but I'm yet to see how that's going to work on connected TV.

Matt: Taking you back to what I know about Australian connectivity is that most connections are metered, and bandwidth is really expensive. Those are the two things that I hear often. So my question is, do you guys fall prey to that with your spot services?

Jeremy: Yeah. It's reflected in the rendition letters we have. They top out pretty low in comparison to somewhere like America. Somewhere like India, we're considered HD, but what you're alluding to is connectivity is poor by comparison.

That's come from a long drawn out project with fiber connectivity that's become political, and it hasn't actually ended up in a good position. Basically, "Why would you invest in copper if you know fiber is coming?

And if you're the fiber company and you're in inheriting a copper system, why would you invest in that?" So, we're probably behind in that regard. Having said that, the mobile telcos have taken over.

Matt: Really? So high connectivity in terms of 4G? Are most of your customers using 4G in the house? Or, not most, but are you seeing that as a thing?

Jeremy: A common experience for a user at home would be on the phone scrolling through something like Facebook, getting to a video, and maybe even turning the Wi-Fi off just to watch that video.

Steve: Really?

Jeremy: Yeah. If something is taking too long to load, you turn off the DSL and you go into the 4G. Because the 4G connectivity is so good. But then you're paying the metering on a telecom plan, which would be a lot more than a metering on a DSL plan.

You're constantly in a seesaw of "How am I connecting? For what purpose? How quickly do I need this?"

Matt: How do you think this affects your viewership?

Jeremy: I think it does. There's an educated user out there who has very technical demands. "HD. Where's 4K? Where's 22K sound audio channels?" Just like, "Guy. We can't." We know.

We've run HD streams up to 7 meg at 1080, which some users got really well, and other users didn't. A lot of users couldn't even stream in that capacity. So it was like, "We can do this. It's just not worth the investment in time and money."

Then there's the licensing on top of that. With our live streams, because we're a catch up service, we don't necessarily have to DRM as long as we keep the quality low.

Matt: As long as keep the quality low, really?

Steve: Interesting.

Phil: Yeah. If you're not a 720p or higher. I've seen that in studio contracts before.

Steve: So, that's a content owner requirement? Anything over a certain resolution, have DRM. But if your streamers can't even support that then don't even bother with it. Basically?

Jeremy: Yeah. So it means that for us, we can get away with just clear HLS, encrypted HLS, which makes streaming 50 channels financially viable and building the players to support across all these devices.

Matt: I assume that the refrigerator operating system and DRM play nicely together.

Jeremy: I don't even think there was a spec for Dash Live back then.

Phil: No. HB TV with DRM is totally a thing HDTV 1.5 with DRM is totally a thing. It's pretty much PlayReady. It's an old version of PlayReady, though. An old version of the key exchange and everything. So, that's fun. I have done it. It sucked.

Jeremy: We do thank you for your effort. Because we were all looking at each other wondering, "How is this going to work?" But it paid out in the end because there was a universal platform, and something Australia can showcase to the world. You can work with other competition, you still get to earn your own video.

You're still paying your own CDN bills and creating your own experiences, but as a collective broadcast it's you creating that next generation of experience.

That hybrid broadcast and internet together, which I think is cool. We just need to keep investing in it.

Phil: There's the HB TVs too. For example, that does pull in later versions of the dash spec, gives you a lot more options. But unfortunately, TVs are already in people's houses.

Jeremy: Being on the bleeding edge, iOS is so easy because it comes out with a new release and everyone gets the new operating system that weekend. With a television they don't necessarily update, ever. You're stuck with a seven year window of keeping this platform alive.

Steve: Seven years?

Jeremy: Yeah, potentially.

Phil: I once worked on a project where supporting some of the older TVs was so hard that the customer shipped people Chromecasts. Said, "Sorry your TV is going to stop working. Here's a Chromecast in the post."

Jeremy: I've often joked it would be cheaper to send them a DVD.

Phil: Yeah.

Matt: So, back to the connectivity bit. Your viewers have an expensive experience, what is that like from your perspective as like what a CDN build looks like? I've heard it's higher, like your Fastly rate or whatever is going to be higher than it is in the US?

Jeremy: A lot of that comes from the connectivity at the pairing level. There's a total cost of CDN that I think everyone gets stuck on the sensible gig conversation going out, without actually factoring in "If you were to have an origin on a different CDN how are you going to transport that content across, and what's the rate on that?"

That's something we probably talk about 10 times a week. It's--

Matt: It comes up occasionally.

Jeremy: Yeah, it's definitely trying to work out the best system for where should our origin and where should our CDN live, and how do we get the video to transport across? It's always a big decision in how we architect now.

I think everyone's got a plan for multi-CDN, and I don't think anyone's looking properly at the total cost of that.

Steve: Are all your points of presence in Australia?

Jeremy: We are geoblocked to Australia, which makes it a little bit easier in that origin egress charge. But when your origins are overseas it makes a egress-- Even just trying to do an HD stream.

How are you really going to pull down those renditions across the Pacific Ocean? It took me 14 hours in an airplane to get here. What makes you think you're going to be able to pull it down in a fraction of a second?"

Phil: That changes your OVP marketplace out there, because really the people who are relevant, the architectures you build are all about the people with local deployment.

Jeremy: Absolutely. You've pointed out Fastly, I think they're a great story. That experience that everyone has on Spotify, you press a button and the audio plays immediately. I think if you go to use some of the other audio programs, you notice very immediately that you're hitting buttons and waiting even half a second or second.

That experience that they've built is amazing, and I think everything we should all be achieving towards as an industry. It's pretty impressive tech.

Matt: That's Spotify in Australia, specifically?

Jeremy: Spotify globally. It's an experience. You should get the guys to talk about it one day, I won't explain it well enough.

But I think there's a library of the first few seconds of every video sitting really red hot on the edge, so that way you've got time to go and get the second, third, fourth, or fifth part of the video. But the actual experience is almost instantaneous.

Matt: I've always just assumed that it was because I have fast peering and I'm listening to content that other people have cached already.

Jeremy: No, they go to the red hot creation on populates. Obviously as you go between countries the music that's popular in that country is different. So they've had to factor that in.

Anyway, get those guys to explain it. It's definitely an experience, that instant click to play, that startup time seems to be a really big thing. Probably one where ad tech lets us down so badly.

Matt: Yes. It's bizarre, whenever we look at data with customers and they're like, "Our viewer experience score is low,"and once you actually start digging into the data it's three seconds added to startup time because it's waiting for an ad.

Phil: Yeah, got to go find an ad and make sure it's stitched in, got to get some impression beacons on the way.

Matt: To be fair, it is an incredibly complex-- The quote that I've seen repeatedly lately is that comment about "The most brilliant minds of our generation are trying to figure out how to show more ads efficiently."

Especially when you look at, when you think about these exchanges and how they work. I mean it is wild that there's this whole auction happening when you load a website. It's also terrifying.

But anyway, I totally agree. Who knows how much better most video experiences would be if monetization by ads wasn't such a hard requirement for so many people?

Jeremy: Absolutely. When you've got auctions sitting behind auctions, behind auctions, behind auctions. I've seen wrappers 10 deep who circumvented the whole globe. Bounced between UK, Australia, England, Singapore.

You can talk of three second startup time, I dream of that.

There can be some real nasty ads in startup times because of that, and that's probably another reason why I like server side so much. You pay a tax at the start of the clip but you get that steady experience the rest of the way through.

Matt: There's nothing quite like watching something, enjoying it, all of a sudden the video stops and the spinner shows up. You're just sitting there waiting and then suddenly a mid roll plays for a little while. So on top of getting to enjoy a mid roll, it hits also this terrible-- like, "Get it over with."

Jeremy: If that experience to get the mid roll takes the same duration as the mid roll. It's--

Phil: Even better if it falls back to a bumper or something. "We couldn't fill this ad, so."

Matt: It's like the advertiser you watch it when you're bingeing a show on Hulu or something?

Phil: I wouldn't know.

Matt: OK, sorry. When I've been binging shows, and all of a sudden you see the same Geico commercial 40 times across six episodes of a show. There's been times where I saw two ads, seeing the same ad twice. Just immediately played again.

Phil: That used to happen to me on a platform I won't name, but the worst part was they had no direct ad sales in the UK. They forwarded that to charity ads.

So, it was charity ad, after charity ad, after charity ad. Three, four, five. Same ad, same break, same part. It was just--

Matt: Like Groundhog Day.

Phil: Didn't make me want to give money to charity. That's what was really upsetting about it.

Jeremy: There's definitely a-- VAST4.0 gave us the creative idea, and if everyone listening could please use it we would solve this problem. But what you're finding is the same ad, but coming from multiple DSPs.

They're just not are not aware of each other, and that's where they're colliding.

And that's where a lot of that happens. The solution is out there, it just needs people to use it.

Matt: We could also sprinkle some block chain on it at some point. All the problems would be solved.

Steve: I want to touch back on something really quick. This idea of the CDN in Australia is interesting to me. Most people think of a CDN as like, "OK I have a central place for my files, and if I have viewers all around the world I need this file cached everywhere so I can reach them quickly."

But what we're talking about here is Australia, and you said earlier like the size of Connecticut. I mean, people-wise. Your geoblocked to the Australia region. So, I'm interested to understand where does the specific value of the CDN come in for your local area?

Jeremy: I mean, it's scale, it's concurrency. When you talk to a back end developer nowadays and they stop talking about the threat of IOT, and suddenly having to manage lots of scale and connections.

It's like "We've been doing that for a long time. We've got this segmented video, it's really small bits, we're going to need lots and lots and lots of concurrence, and we need to scale, and we need to be reliable."

You know It's really interesting to watch where CDNs go over the next five years with LAMBDA on the edge or web assembly, those sort of things. How that drives video technology is going to be really exciting, and probably address a lot of these issues that we're raising today.

But I think at the end of the day there's the compute price that you pay, and there's a price on egress or serving across networks that we have. CDNs obviously help that, and lock down what your actuals are going to be. What your costs are.

You can't really fix price, but you can at least estimate and get a good rate and look into that.

Steve: Get to some level of predictability.

Jeremy: Yeah, predictability is hard.

Phil: While Australia is small in terms of population, it's big.

Jeremy: It's the size of America.

Phil: Matt just looked really surprised.

Matt: So, repeat that one more time?

Phil: Australia geographically is huge. It takes what, five hours to fly across Australia?

Steve: It's a little down on my map, in the corner.

Phil: We had this discussion last night about map projections. This is this again.

Jeremy: There's multiple hour time zones between each. The big joke I have is the English backpacker, Phil you know these guys well. They fly in to Perth and they say "I'm going to go to Perth, I'm going to go to Sydney, I'm going to Melbourne, I'm going to Cairns. Then I'm going to fly home."

I say, "How long are you here?" They say "Two weeks." I said "You better start driving, you're not going to make it." Because it's thousands of kilometers across.

Matt: So it's Connecticut, but just exploded across--?

Jeremy: There's a big wide desert sand in the middle and a red rock. Yeah, it's spread out.

Matt: There's a note on here in our notes about New Zealand, which according to John Oliver apparently isn't on like half the maps in the world, but when you talk about your geoblocking. That would also block New Zealand residents I would imagine?

Jeremy: Yeah. It's the same conversation if you guys went overseas and someone asked if you're Canadian. That's a whole other country.

Matt: Right.

Jeremy: At least in Australia, New Zealand, they're two separate landmasses with two separate governments. We're friendly. This is a country of 4 million, maybe even just? But they're very well connected, very different to us. They've all got their own fiber network, run out to their houses.

Matt: Really?

Jeremy: Yeah. A lot of the deep sea cables go via New Zealand, too. So connectivity is very good in the country.

Phil: One of the things that I found super interesting when working on deploying stuff to Australia and New Zealand was you often build out your origin in Australia, but then be delivering into New Zealand from Australia, because for example--

A great example is AWS doesn't have a region that's in New Zealand, it only has basically Sydney. So you are hamstrung by the links between Australia and New Zealand, of which there are to my standing: two.

Jeremy: That many? There must be redundancy. Yeah. Sometimes one of those doesn't work.

Jeremy: It's a long way away. Again, it's 14 hours flight. You can imagine dropping a sea cable between here and there. It's a huge investment.

The same goes off to New Zealand. It's still a three hour flight, or something. So you can imagine, you've got to run a cable the whole way there and the whole way back, and then try and monetize that to a population of 4 million. I don't know.

It seems like a lot of work. But having said that, internally their networking is amazing. Enviable.

Matt: Nice. This is a slightly embarrassing story, but this fight stuff is reminding me when Steve and I went and did a trip around-- We went and spoke in at a BrightCove event in Tokyo, and we also went to the first Sydney video meetup which was awesome.

So we had this Tokyo event lined up, and I was like "We should just go to the meetup while we're over there."

Steve: Via Australia, because they're so close.

Matt: Yeah. I was like "Since we're on that side of the world, we'll just hop by Japan." I didn't really realize until I got on the flight. I was like, "How long is this flight again?" It's like 12 hours, or whatever it is.

Jeremy: Yeah. You're close in time zone, but a different hemisphere.

Matt: Yeah. Minor difference.

Phil: But there's so much going on in the market out there, it's so exciting. Thank you for coming and joining us.

Jeremy: Thank you, Phil.

Steve: Thanks, Jeremy.

Matt: He's also-- This recording is going to go out a month after this, but there is a meet up tonight. So if you hear this in retrospect you should look up his talk.

Jeremy is speaking tonight at SF Video. If you hear this in retrospect, look his talk up online, we'll have it on the YouTube channel.

Phil: Who else is talking tonight?

Matt: I don't know. I think it's just Jeremy.

Phil: Some guy?

Matt: I think it's just Jeremy.

Phil: OK.

Matt: No, I'm just kidding. Phil's talking too. Look up both of these talks.

Jeremy: If you're ever in Sydney, get in touch with me. Anyone out there, really. We'll see. We run a monthly event and we're happy to invite people, or have them speak if they've got a great topic.

Matt: I can neither confirm nor deny that Sydney Video Tech is technically bigger than SF Video Tech.

Phil: It's the biggest video tech meetup in the world.

Jeremy: Thank you.

Matt: So, everybody. Go to SFVideo.org and make sure that you join the meet up. Thanks again, Jeremy. This was awesome.

Jeremy: Thank you very much.