Discover more from Future Commerce
Episode 54
December 12, 2017

Microsoft Paint, but for Augmented Reality

What *exactly* is Deep Learning? What can HAM radio teach us about decency in the social age? Danny Sepulveda joins us for a new "Future Policy" segment. Also - what exactly is Facebook Messenger Kids all about?

<iframe height="52px" width="100%" frameborder="no" scrolling="no" seamless src=""></iframe>

this episode sponsored by

From the Gutenberg Press to Twitter - how tech innovation gives legitimacy to our words. Phillip gets down and dirty with AI and warns retailers of marketing confusion. Also - what exactly is Deep Learning?

Show Notes

Vibe notice: if the vibe is different, it's because this is the first in-person episode with Phillip and Brian and they're having way too much fun with no imbibing of ethyl alcohol before recording.

For more information about the evolution of media and journalism check out Understanding Media by Marshall McLuhan, a fascinating read covering the evolution of media in the Western World.

Shoutout to Kiri Masters:

  • We're sorry we forgot to mention your great podcast on episode 52. We're sorry; we're the worst.
  • Come on the show to talk about building a brand!

In the meantime, folks, go listen to Kiri Masters' Ecommerce BrainTrust.

Our Inaugural NPS:

  • Very first NPS went really well. It was fun to hear from our listeners.
  • The FutureCommerce's copywriter's job is in jeopary: thanks NPS commenter.

The Big Think Segment

  • What can HAM radio teach us about decency in the social Age?
  • Evolution of the written word: the Gutenberg Press gave authority to printed material due to the medium in which it was distributed.
  • When something is in print, it carries weight and authority.
  • Journalistic practices evolved out the necessity for us to bring ethics to the printed medium.
  • Fast forward and Facebook and Twitter over the last 5 years have become the authoritative choice for disseminating news.
  • Riley Florence tweeted parallels between Twitter and HAM radio's early toxicity of the medium.
  • Early adopters had to come up with a set of guidelines to root out rampant toxicity.

Retail Prophet

Retail Prophet Doug Stevens first podcast says the future of commerce is social.

  • Implications for retail: if some of our greatest thinkers say social is the next frontier for retailers, and social is a toxic place, then we need to know how to behave ourselves as retailers and consumers in the social medium.
  • New mediums create paradigms that require getting used to and understanding.

Facebook Messenger Kids

  • What you think it is vs. what it actually is:
  • What you probably think it is: we don't need another product to help our kids get on chat.
  • There's an Inherent creepiness to marketing chat to kids.
  • A real fear that creepy people can subvert the platform for dangerous purposes.
  • What it actually is: a way to control and keep your kids safe when chatting online.
  • It gives parents the tools to keep your kids safe when using a chat platform.
  • It's like kik or snapchat with parental controls.
  • It's, "I as a parent get to moderate who gets to talk to you." Which is smart and healthy.
  • Important to shepherd and teach our children that these can be mediums for both good and bad.

Future Policy with Former Deputy Assistant Secretary of State, Daniel Sepulveda:

  • Episode 51's conversation about the Digital Divide serves as a good frame for the policy conversations we're going to have.
  • It's exactly how the legislative process works: a conversation between 2 people from either side of the aisle hashing out problems and solutions.
  • Together they come up with a zone of agreement and bring it to their bosses and the bosses talk to each other, and then bring it to their colleagues and they take it to their colleagues, and eventually a solution is proposed.
  • Brian took the course of free market economics and the rise of innovation as the natural course of economics, so net net it will be a positive.
  • Phillip worries that we have responsibilities to each other and and have communal responsibilities for those who are going to lose out that aren't being discussed.
  • They each held two different points of view, each listened to each other and have a natural respect for one another and were able to have a respectful conversation. / What's missing today is the ability to come together with mutual respect and listen and examine the question to eventually come to a solution
  • That's what Danny hopes we would do as we talk about future issues going forward.

Machine Learning and AI and Marketing Confusion:

  • Google Brain's auto machine learning (Auto ML) created its own Artificial Intelligence.
  • The researchers at Google Brain announced the creation of autoML, an AI that is capable of creating its own AI. The babies are having babies.
  • It created something called NASnet that recognizes objects in video at real time and has an 82.9% success rate.
  • Brian wants you to watch Person of Interest.
  • AI term and the Machine Learning terms are being abused because people don't understand the difference.
  • See episode 14 for an overview of this with Jonathan Epstein from Sentient Technologies.
  • Retailers: caveat emptor! Be highly skeptical of any technology provider telling you they're using deep learning or AI.
  • It's only been recently that Google and Amazon have productized deep learning.

Explanation of Machine Learning

  • Machine learning is trying to find the best fit algorithm. Think of a scatter plot in Microsoft Excel
  • You can make a best fit straight line with a particular slope that will try to hit an average or median between all of the points on your scatter plot
  • Imagine what that looks like. You'll see that the straight line is really far off the mark from most of those points because it's an average. That means there are major outliers.
  • The difference of machine learning is the straight line. The deep learning continues to perform refinements to the line to get it closer to all of those data points.
  • That's called gradient descent. It's not just 2D it's a multidimensional scatter plot.
  • It's still just trying to find a better fit line, and n finding that better line, it can begin to make predictions about where a particular data point may fall along that line.
  • What most people are selling you is the straight line. They're selling you a really average product.
  • It's the difference between A/B testing and 1 to 1 personalization.
  • Most are doing A/B testing and calling it 1 to 1 because they don't have the means to do it.
  • Retailers: if you're being sold that you have to bake AI into something, or every single product has AI branded on it, don't be fooled.
  • If they are not using Google, IBM, Amazon, Microsoft or Sentient technology, it's probably not legitimate.
  • If you're listening to this right now in 2017, be skeptical.

Future AI/ML

Wrap it Up! Bitcoin Edition:


  • Our 2018 prediction show is coming up! Subscribe to our podcast anywhere you listen to your favorite podcasts and sign up for FC INSIDERS for exclusive content.

Download MP3 (41.2 MB)

Phillip: [00:01:07] Welcome to Future Commerce, the podcast about cutting edge and next generation commerce, rated as top five podcasts for consumer brands by Bobsled marketing. I'm Phillip.

Brian: [00:01:16] And I am Brian.

Phillip: [00:01:17] And as an aside, we didn't include Kiri Masters of Bobsleds Marketing in our podcast roundup back on Episode 52.

Brian: [00:01:23] That was a mistake.

Phillip: [00:01:24] Yeah. Woops. Sorry, Kiri. And so we are the worst. We're awful. So Brian, let's get Kiri on the show some time to talk about how to build a brand on Amazon.

Brian: [00:01:35] Yeah let's do a full segment.

Phillip: [00:01:36] Yeah, because it's not like we never talk about Amazon on this show. It's really relevant.

Brian: [00:01:41] Yeah. Like maybe 55% of the time.

Phillip: [00:01:44] Can we just like rebrand as like the Amazon... Fifty five percent of the time. I see what you did there.

Brian: [00:01:49] {laughter}

Phillip: [00:01:49] That's really funny. Yeah. Letting you in on the joke. Fifty five percent of all Black Friday and Cyber Monday sales took place on, but anyway.

Brian: [00:01:59] Of course you would have known that if you'd have listened to episode 53, so...

Phillip: [00:02:01] And well you may have known that if you listen to Kiri's podcast, so go check it out on her show at ECommerce Braintrust. And hopefully we'll get her on here in the near future. So, OK. Fc insiders, Brian.

Brian: [00:02:18] Yeah. As always, we want you to sign up for FC Insiders. We think it's a great way to find out about what's next in retail tech. Be the first to know about newest trends in retail tech. You've got to subscribe to FC Insiders newsletter today.

Phillip: [00:02:32] And you can do that if And we want you to never miss another episode of Future Commerce. And that's the best way to stay...

Brian: [00:02:41] We should start doing some pretty interesting... We should start asking people to sign up for Future Commerce Insiders in fun ways.

Phillip: [00:02:47] Yeah.

Brian: [00:02:48] I think in the next episode we we should do that.

Phillip: [00:02:50] Ok. Yeah. Let's let's think about ways that we could entice you to get on to our newsletter. And oh, one of the ways that we're using our newsletter is actually to get your feedback. And we sent out our very first NPS.

Brian: [00:03:04] It went really well. Really well.

Phillip: [00:03:05] Yeah, really well.

Brian: [00:03:07] It did. It was really fun to hear from our listeners.

Phillip: [00:03:09] Yeah. We had some really really great reviews which was sort of surprising. I wasn't expecting such great feedback. But just to kind of like cherry pick a couple here, somebody says the diversity of content and the palpable vibe that Brian and Phillip exude due to their passion for the topic is engrossing. Can we get that person to be like a copywriter for us?

Brian: [00:03:35] {laughter} It's anonymous. Unfortunately, we can't reach out to them.

Phillip: [00:03:37] But we have an NPS of 79 at the moment.

Brian: [00:03:40] So far so good.

Phillip: [00:03:40] So help us. {laughter} Have you ever heard an NPS of 79? It seems like it's a really small sample size. It's actually not. Like we have a really decent number of people that have... People like our podcast, Brian.

Brian: [00:03:54] Yeah I guess so. If you don't like our podcasts, just don't fill out the NPS.

Phillip: [00:03:57] No, no, no, no.

Brian: [00:03:59] No. Come talk to us.

Phillip: [00:04:00] Yeah. And you'll get that if you go to, and we're gonna have... It'd be nice if you took that every six months or so, and we'll be sending that out over email through FC Insiders as well.

Brian: [00:04:12] Oh wait. OK. Got it. Here's one for FC Insiders.

Phillip: [00:04:15] Okay.

Brian: [00:04:15] FC Inciders is here to help you avoid the retail apocalypse. Join now and stay alive. {laughter}

Phillip: [00:04:23] {laughter} I like it. I like it. I like it. Okay. I don't know that this is gonna be a regular segment, but it's sort of been a thing that we've done accidentally in the last few episodes. So I have a big think topic. OK?

Brian: [00:04:38] Here we go.

Phillip: [00:04:39] It's like an half formed idea. This is a new segment. I don't know that we're gonna do it every week.

Brian: [00:04:44] No, I love it. I love half formed ideas.

Phillip: [00:04:45] Okay. Yeah. This is really on brand for us. {laughter}

Brian: [00:04:51] {laughter}

Phillip: [00:04:51] Okay. It's a half formed idea, and maybe it's one of those things that kind of fully formed in an FC Insiders exclusive later on, which like it's a podcast that we send out in the Insiders newsletter that's just for our subscribers to the newsletter. But I had this thought, and it's really kind of tying the other seven or eight different ideas...

Brian: [00:05:14] Ok. Lay it on me.

Phillip: [00:05:16] OK. Okay. So, I was thinking about the Gutenberg press. Ok, which was...

Brian: [00:05:23] Important tech. {laughter}

Phillip: [00:05:25] Yeah, as far as retail tech goes, Gutenberg press is kind of up there right? The Gutenberg press. And I was thinking about how what it must have been like to live in that age where you went from... You had not only just the mass distribution of thoughts and ideas, but printed sort of beautifully and uniformly in a way and put on paper and put in a way that gave authority to whatever words were printed. So think about what the first 20 years of the Gutenberg press must have been like. Of the kind of even disinformation that could have been legitimized because of the medium in which it was distributed. Does that make sense?

Brian: [00:06:05] Yeah, and maybe not even the first 20 years. Maybe like... The first 20 years there wasn't that much printed there.

Phillip: [00:06:12] Well, I don't know.

Brian: [00:06:13] Good question. OK, I get your point, though.

Phillip: [00:06:15] Yeah. You feel where I'm coming from?

Brian: [00:06:17] Yeah.

Phillip: [00:06:17] OK. So but the idea is that, think about the Gutenberg press and think about what's happening with social today. And, you know, I think we all have felt the effects and reverberation of last year, at least in North America and here in the United States, since the 2016 elections. And look at how the information and disinformation both spread so quickly along the same medium.

Brian: [00:06:48] Quick side note, I feel like this has been happening...

Phillip: [00:06:51] Much longer...

Brian: [00:06:52] Much longer than that through other means like email chains and so on.

Phillip: [00:06:58] Yeah.

Brian: [00:06:58] But anyway, I it plays to your point.

Phillip: [00:07:01] Ok. So I have a string of events. What was it, five or seven years ago people were boo-hooing that Google's RSS reader went away? Remember this?

Brian: [00:07:14] Boo-hooing?

Phillip: [00:07:14] Yeah, it's a verb. Boo-hooing. They were kind of getting a whiny that Google's RSS reader was going away.

Brian: [00:07:23] Yeah.

Phillip: [00:07:24] And a lot of people were saying, who even uses that? Right?

Brian: [00:07:27] Right.

Phillip: [00:07:27] If you want to read the news, where do you get it? Facebook.

Brian: [00:07:31] Facebook was the answer. It was one of the big answers.

Phillip: [00:07:34] So what happened is that Facebook became the authoritative source for aggregation of news and people actually liked it.

Brian: [00:07:40] And Twitter. I mean, really Twitter.

Phillip: [00:07:42] Right. So what I'm saying is that journalistic practices evolved out of the necessity for us to bring ethics to the printed medium.

Brian: [00:07:52] Right.

Phillip: [00:07:53] All right. So not that there wasn't journalism before. But when something is in print, it carries weight and authority. And I think we're going through the same age in social now. And so I have this idea that, and this is not an original idea. Riley Florence tweeted this and basically found a lot of really interesting parallels in the ham radio like amateur radio operator guide.

Brian: [00:08:18] OK. Yeah. Yeah, yeah, yeah.

Phillip: [00:08:19] Which basically that said that, you know, ham radio operators had the same problem in toxicity in their communities decades ago. And they came up with a set of rules and guidelines to follow to help them sort of root out the toxicity. So just a few little samples here. And these will be on our show notes courtesy of Matt Vermillion. Thank you so much, Matt, because he helps us because we're not that organized, but OK. So the first rule is always be polite regardless of the circumstances. And if not, avoid transmitting. You think that that might be useful on Twitter?

Brian: [00:08:58] Yeah.

Phillip: [00:08:58] Possibly? Right? Decency is important. Number two, be a good listener. It will help you better organize your thoughts before transmitting. Remember these are ham radio operator guidelines. But I think it has so much parallel with how we could actually make...

Brian: [00:09:15] B....

Phillip: [00:09:15] {laughter} Yeah, OK. Number three, don't acknowledge the presence of deliberate interference. Ok?

Brian: [00:09:26] Oh, avoid the trolls.

Phillip: [00:09:29] Yeah, avoid the trolls. Don't acknowledge the presence of deliberate interference. That's most likely their overall objective of the person doing the interfering. This was written decades ago for amateur radio operators, and it's so relevant to today. Anyway it's blowing my mind.

Brian: [00:09:48] It's actually not surprising that would be relevant. It's not that different.

Phillip: [00:09:52] It's still human communication.

Brian: [00:09:54] Yeah, totally.

Phillip: [00:09:56] Platform change.

Brian: [00:09:56] And the platform's not even that different. I mean, think about it. It's information that being broadcasted across the world.

Phillip: [00:10:03] Right. Anyway, I again so big think idea. I feel like, you know, the more that people... And the reason I actually got to thinking about social in the first place is Retail Prophet...

Brian: [00:10:20] Doug.

Phillip: [00:10:20] Doug. People are shouting at their radios right now.

Brian: [00:10:25] Radios. {laughter}

Phillip: [00:10:27] {laughter} Doug Stephens.

Brian: [00:10:29] Yeah.

Phillip: [00:10:29] Retail Prophet, Doug Stevens actually just put out his podcast, his very first episode of his podcast, in which he said the future of commerce, which, by the way, thank you for stealing that. That's like such a compliment to us.

Brian: [00:10:41] It is a bit of a compliment.

Phillip: [00:10:42] Yeah, yeah, yeah. A bit of a compliment to us. Future of commerce... He says the future of commerce is social. And I'm saying if everything's going social, we need to find a way for us to make social a better place. I have a different take on it. And we'll have our predictions episode...

Brian: [00:10:58] Oh my gosh.

Phillip: [00:10:58] ...later in the year and maybe an episode or two.

Brian: [00:11:01] I'm laughing really hard at an inside joke.

Phillip: [00:11:03] Yeah, I know. I know. So anyway, the idea is that if some of our greatest thinkers and retail minds are talking about social being the next frontier for us, for retailers, then social is a toxic place. And we have to understand how to behave ourselves as brands.

Brian: [00:11:20] And also navigate through it. Yeah, yeah, as consumers as well.

Phillip: [00:11:24] Sure.

Brian: [00:11:24] Like it's both. It's a two way street.

Phillip: [00:11:26] Sure. Anyway.

Brian: [00:11:27] And I think we're getting better at it. I think you're right. I think these rules, this, you know, sort of dare I say, advanced netiquette.

Phillip: [00:11:34] Yeah.

Brian: [00:11:36] Is starting to make its way out. It's interesting to see what segments of our society have picked up on this faster than others. And I don't think it's necessarily an age thing. It's really funny that baby boomers and an older spend a lot of time complaining about how much time millennials and younger were spending on their phone.

Phillip: [00:11:59] Right.

Brian: [00:12:00] When in fact, I think the baby boomers are now the worst offenders of this. And once they did. And once they realized how awesome it was. Again, I'm making big generalizations.

Phillip: [00:12:10] You're also sort of thinking off the cuff, but yeah.

Brian: [00:12:13] Oh, yeah. I mean, this is a half baked idea segment.

Phillip: [00:12:15] Yes. {laughter} I kind of like the the "Big Think Idea", but yeah. Whatever. Yeah, I like that. I like that brand a lot better than "Half Baked," but carry on.

Brian: [00:12:26] In short, I think they finally sort of picked up on the utility and the fun of it and the addictive nature of it.

Phillip: [00:12:32] Right.

Brian: [00:12:32] But they weren't really prepared. I feel like the younger generation learned the things you just talked about a lot faster.

Phillip: [00:12:41] Probably because it affects their social circles more closely.

Brian: [00:12:43] Exactly. And they were early adopters.

Phillip: [00:12:48] Right. The things that they do can have real emotional impacts to them. And real actual societal impact for them.

Brian: [00:12:55] Yeah.

Phillip: [00:12:55] When you're younger, your acceptance into your small society of people is everything that you have. Your social credo, your social...what's the word? You social currency that you have is relationships. And you are with this crowd or... Someone told me, and obviously I'm much too old to be an avid Snapchater. But, you know, part of the stickiness of social and kids using social is to just keep them addicted to using social, to keep competitions among their friends. Snap...

Brian: [00:13:37] Interesting.

Phillip: [00:13:38] Right. Snap...

Brian: [00:13:38] Interesting.

Phillip: [00:13:38] Like the Snap streak is everything.

Brian: [00:13:41] Yeah.

Phillip: [00:13:41] And kids want to engage in social because it's a competitive social sticky experience.

Brian: [00:13:47] Yeah. Interesting. Interesting. And I can totally see that.

Phillip: [00:13:51] And that affects... What I'm saying is it affects the younger generation more. So they have an immediate feedback loop into negative behavior and toxic behavior, whereas the boomer generation actually does not necessarily. There's passive aggression, there is not direct contact with their social circles because their social circles are more widely flung.

Brian: [00:14:10] I love how general we're being right now.

Phillip: [00:14:11] Okay. Anyway. It's terrible.

Brian: [00:14:12] No, it's good. And actually, I want to add on this a little bit further. Think about this. Again, I'm making generalizations. So if you don't...

Phillip: [00:14:20] What even is this episode right now? This is amazing.

Brian: [00:14:23] Hold on. Hold on.

Phillip: [00:14:23] This is good.

Brian: [00:14:23] This is good. Think about something else. OK. So I think when the older generation... We're going to get a little bit...

Phillip: [00:14:29] Oh, gosh.

Brian: [00:14:30] The older generation, when they first saw stuff like... Let's apply this even further to next gen tech. Augmented reality and things like that.

Phillip: [00:14:37] Right.

Brian: [00:14:37] We talked about sort of digital vandalism on an earlier podcast. I don't remember which episode. Maybe Matt can find that.

Phillip: [00:14:45] That's like 48.

Brian: [00:14:46] Yeah, somewhere around there. I think that in the physical world we had graphic artists use spray paint art on buildings. When an older person would look at that, they just saw vandalism.

Phillip: [00:15:00] Right.

Brian: [00:15:00] And our generation could tell when they were being subversive versus not.

Phillip: [00:15:05] Oh, I see.

Brian: [00:15:06] And so there's like this whole element of like layer.

Phillip: [00:15:10] Yeah.

Brian: [00:15:10] And maybe like subtlety to what people are doing. I feel like I'm really offending an older generation here. So if you don't fall into this category, don't take offense.

Phillip: [00:15:19] No. We're just gonna have Chris edit this down to some...

Brian: [00:15:23] {laughter} But you get one where I'm going with this, right? We're a lot more used to open source edits on edits on edits... Meta thought, meta world, meta life.

Phillip: [00:15:33] Right.

Brian: [00:15:34] Right?

Phillip: [00:15:35] That's sort of how we think.

Brian: [00:15:36] That's sort of how we think. So to get back to the original point. We have the ability to tell when someone is just trolling and when someone's actually making a point.

Phillip: [00:15:45] Right. I understand. I understand. And it could be a generational thing. It could also be, you know, again, digital natives like understanding the medium because you grew up with the medium.

Brian: [00:15:55] Right. That comes down to medium. Totally. One hundred percent.

Phillip: [00:15:57] It's not coming from... You're not bringing 40 years of experience prior to that operating the medium. Ok. Let's bring it back to you retail.

Brian: [00:16:04] Well, let me just make one quick statement. I'm not saying older generations can't pick up on subversive things and subtlety.

Phillip: [00:16:11] Right.

Brian: [00:16:12] At all. But new mediums create paradigms that require getting used to an understanding of. Like those mediums, in order to understand those subversive pieces or statements. And so anyway, my point is that new technology brings new ways of sending messages and making statements and creating new things.

Phillip: [00:16:34] Yeah.

Brian: [00:16:34] And so when we talk about netiquette. Netiquette is something that's going to continue. And it's not even the right word for it. What we consider as good and worthwhile on the net and in technology is going to continue to change. And we are going to have to understand the new technology in order to understand what is good and what is right.

Phillip: [00:16:57] Yeah.

Brian: [00:16:57] I think going back to ham radio, to the Internet, is a great, really, really like stark example. I mean, you went back to the printing press.

Phillip: [00:17:05] Right. Exactly. Exactly. OK. Well, that's really good. So that actually segues really nicely and accidentally into our first news story, which is Facebook Messenger Kids Facebook.

Brian: [00:17:20] {laughter} It's great.

Phillip: [00:17:20] Yeah. OK. Actually, it sounds really awful on the face of it. So I'll give you the if you only read the the headline, I'll give you what you think it is. And then we'll tell you what it actually is. So Facebook Messenger Kids. Basically, do we need Facebook to be helping our kids become more engaged with people on chat? Do we need one more? First of all, we don't need one more of these.

Brian: [00:17:49] I'm thinking back like AIM and MSN Messenger days. Do you know how much time I spent on MSN Messenger as a teen? A lot.

Phillip: [00:17:59] I mean you're dating yourself a bit. The point is, is that when you're targeting specific things to children, even if it's messenger like here's an inherent creepiness about that, I think. Right? And even really well-meaning, I don't want to call it a social network. But even well-meaning properties like Club Penguin were not immune. Like Disney could not control that community.

Brian: [00:18:26] Right.

Phillip: [00:18:26] And so they're not immune to, again, being subverted by people who have less than desirable intentions with using those platforms. But so that's what you think it is. What it actually is, is a greater way of controlling and pushing your child toward and keeping your kids safe online. If they're going to use something, if they're going to chat with anybody online, if they're going to video chat, if they're going to use filters and they're going to have a social network with their friends, then why would you not have a platform that specifically built just for kids that give parents control over who they can talk to?

Brian: [00:19:10] Monitorable...

Phillip: [00:19:10] Yeah.

Brian: [00:19:11] It's one hundred percent transparent.

Phillip: [00:19:14] Yeah.

Brian: [00:19:15] It's a controlled network.

Phillip: [00:19:16] Yeah.

Brian: [00:19:17] I think this is a necessity, actually.

Phillip: [00:19:18] Yeah. It's Snapchat. It's Kik or Snapchat with parental controls like. And not just like let me put a lock on this at the time of day, so you can't use it in your bedroom at night. It's I as a parent get to moderate who talks to you. And I think that that's really smart and healthy because as even though we were just kind of you were sort of talking about an older generation not really grokking social, it is important that we sheperd children and teach our children that these can be mediums for good and for bad and that maybe other people don't have the best of intentions when they contact you there. So I think that this is the right type of application of technology focused for children.

Announcer: [00:20:15] Now it's time for our weekly segment called Future Policy, brought to you by Vertex SMB. And as always, we're joined by Deputy Secretary of State Danny Sepulveda, in this episode responding to the discussion on AI and classism from Episode 51.

Danny Sepulveda: [00:20:35] I thought a) I thought it was a great episode, and I think it actually serves as a really good frame for all of the conversations that we're gonna have. And let me tell you why I liked it so much. First of all, I'm sure you guys thought these are just thoughts off the top of my head. I think this. What do you think? Blah, blah, blah... Like two normal human beings talking to each other. And that's actually exactly how the legislative process works. So I would sit down with a friend of mine on the Republican side of the aisle and say, "Look, I think automation's coming down the line and robotics are coming down the line and it's going to cause X, Y and Z problems." And he'll say to me, "Look, you're overreacting. Nothing terrible is gonna happen. This is just like when the horse and buggy came. Then cars came. Horse and buggy people went out of business, but everybody's fine. So, no, we don't need any of your new laws. And then we'd have a cup of coffee, a couple drinks and talk about where there was a potential zone of agreement. And that's where you just kind of work through a conversation. Then we take it to our bosses and our bosses talk to each other and they see if they can work on an issue together. They present it to your colleagues and they take their sides. And it really is that simple. And we do start much the same way. You just started that last conversation. So one of you took the side essentially of a free market economics and the rise of innovation as a natural course in commercial history. And that whatever happens, it's going to be OK, because net net, it's a benefit. More jobs will be created. The jobs that are lost are actually bad jobs. Therefore, you know, just don't worry about it. Kind of feel it out as it happens, but don't be negative about it. And then the other of you said, well, don't we have responsibilities to each other as a community? And clearly, there are people who are going to lose. Don't we have communal responsibilities to those people? And that is the perfect framing for how a conversation in Congress begins. We have a new thing that's happening. Then we have to determine whether or not it's a problem. And then we have to see whether or not we agree on how to go about solving that problem. And what I really liked about your conversation is that you held two different points of view, but you clearly either have affection or like each other or respect each other's point of view to listen to what the other person was saying and actually say "Hmm, maybe. But..." And then keep going. And the part that's broken about our policy conversations today, and I think in our legislative process is that that's missing. The ability for people to come together as friends or as people who at least have mutual respect for each other and listen and try to identify a problem. And then maybe there's a solution, maybe there isn't a solution, but work together to examine the question. Right. And that's what I would hope we would do as we talk about different issues going forward, whatever that issue might be. I will, to the best of my ability, say, look, people on one hand think there's people, on the other hand think that, and this is what I think, and this is what the process is going to look like going forward.

Phillip: [00:25:41] Of course it doesn't really have a lot to do with retail, but it's not long before the Matels of the world are marketing in those channels because they're already doing it everywhere else our kids have their eyes glued.

Brian: [00:25:54] Kind of. Actually a side note.

Phillip: [00:25:56] Yeah.

Brian: [00:25:57] My wife and I were just talking about this the other day. Our kids at least are not watching nearly as many commercials as kids when we were kids.

Phillip: [00:26:11] I don't know if I agree with that at all.

Brian: [00:26:13] Yeah. No.

Phillip: [00:26:13] Is it because of Netflix?

Brian: [00:26:14] Yeah. Netflix and Amazon Prime.

Phillip: [00:26:16] Right.

Brian: [00:26:17] My kids don't watch commercials.

Phillip: [00:26:18] Right.

Brian: [00:26:19] We don't watch cable.

Phillip: [00:26:20] Right.

Brian: [00:26:21] We already cut that cord...

Phillip: [00:26:23] A decade ago.

Brian: [00:26:23] Yeah. Roughly.

Phillip: [00:26:24] Right.

Brian: [00:26:26] And so we were talking about... This gets back to other things we talked about maybe even back to when we were talking to Saku about shared experiences versus created experiences and and thinking that we're like now today, kids do have there are like sort of the famous toys that get sold every year. But it's not like when we were kids, it's not like when, you know, when there were stampedes.

Phillip: [00:26:53] Well I don't know about that.

Brian: [00:26:53] This is hilarious and ridiculous. But we went back and watched Jingle All the Way with Arnold Schwarzenegger.

Phillip: [00:27:00] Yes. Yeah.

Brian: [00:27:00] Right. And yes, it was a terrible movie. Terrible movie. But the whole point of it was sort of to make fun of the craze that kids have for specific toys and how essentially commercials sort of played into that. And our kids, well at least my kids, just don't have that.

Phillip: [00:27:18] Well, it's interesting because your kids may not have it, but some kids have to have it because my kids won't Hatchimals. They also have the same exact experience that yours do in that my kids only watch Amazon Prime video or they only watch Netflix. They only watch the shows that they want to choose to watch. And they don't see commercials at all.

Brian: [00:27:42] Yep.

Phillip: [00:27:42] Really?

Brian: [00:27:43] Yep.

Phillip: [00:27:43] But they somehow still want the heck out of a Hatchimal. And so someone at school has to be talking up the Hatchimals.

Brian: [00:27:50] Well yeah, social networks are creating these things...

Phillip: [00:27:51] Gosh. Why are we doing an episode about social? I can't believe this is happening.

Brian: [00:27:56] Yeah. You came up with the big idea. {laughter}

Phillip: [00:27:57] I know this was my fault. This is entirely my fault. All right. Well, OK. So, kids... Kids, hide your kids, hide your wife because Facebook is chattin, everybody out here.

Brian: [00:28:11] Oh, that's the best.

Phillip: [00:28:12] I did not come up with that before just now. That happened on the fly. OK. And by the way, if the vibe is different, this is our first in person episode I think.

Brian: [00:28:22] Not our first.

Phillip: [00:28:23] I think it is our first.

Brian: [00:28:24] First of just the two of us.

Phillip: [00:28:25] Yeah. OK.

Brian: [00:28:26] We've done in-person interviews.

Phillip: [00:28:27] Yeah.

Brian: [00:28:27] This is a way too much fun.

Phillip: [00:28:29] This is way too much fun. OK. So and there was no imbibing of ethyl alcohols before. OK. So anyway. Oh, speaking of kids.

Brian: [00:28:44] No, wait.

Phillip: [00:28:44] No, I have such a good segue. Let me take the segue. Speaking of kids, what happens when kids start having kids? Babies are having babies now, Brian. Did you know this?

Brian: [00:28:55] What?

Phillip: [00:28:55] Yeah. Google's Google Brains, auto machine learning, AutoML, their artificial intelligence. Their child, if you will, created its own artificial intelligence. An article out on Futurism detailed the story that back in May 2017, the researchers at Google Brain announced the creation of AutoML, which was an AI that was capable of generating its own AI.

Brian: [00:29:20] Oh, my gosh. We're talking about self replicating software.

Phillip: [00:29:23] And so the babies are having babies. More recently than they presented AutoML with its biggest challenge for it to create a child AI. And it did. It outperformed all of its human made counterparts. Google researchers automated the design of machine learning models using an approach called Reinforcement Learning, AutoML acts as a controller neural network that develops a child AI network for a specific task. And in this one, basically, it created something called NASNet. And it was to recognize objects and video in real time and it had an 82.9% Success rate in doing so. Yeah.

Brian: [00:30:06] I mean, this is like. Have you watched Person of Interest?

Phillip: [00:30:11] What is that?

Brian: [00:30:12] Oh, it's... I don't want to give it away.

Phillip: [00:30:14] OK.

Brian: [00:30:14] Never mind.

Phillip: [00:30:15] No, I've not watched that.

Brian: [00:30:16] It's about this sort of.

Phillip: [00:30:17] OK.

Brian: [00:30:18] But, you know, this sort of gets into evolutionary algorithms and all of that.

Phillip: [00:30:22] Oh that was a show. There was a network television show on Primetime.

Brian: [00:30:25] Yes.

Phillip: [00:30:25] Not so long ago.

Brian: [00:30:26] Yeah.

Phillip: [00:30:27] Like cameras are everywhere and... Yeah, yeah, yeah.

Brian: [00:30:30] It was pretty good.

Phillip: [00:30:31] Interesting. So I find that incredibly like endlessly fascinating. I've been doing a lot of research lately into deep learning, how deep learning and recurrent neural networks and all those sorts of things, how they differ from standard machine learning. Very, very, very fascinating stuff. So this came up in my reading. And I find it interesting how these technologies are so deep, and they're being... The AI term and deep learning are being sort of abused because people don't understand the difference between, you know, machine learning and big data and deep learning and how all these things are very different to each other.

Brian: [00:31:20] Didn't Jonathan Epstein get into that?

Phillip: [00:31:21] He did.

Brian: [00:31:22] Episode 13?

Phillip: [00:31:24] Yeah something like that.

Brian: [00:31:25] Yeah. Early on. And if you haven't listened to that episode, go back and listen to it. Jonathan Epstein from And they have a pretty cool tool at Sentient and for commerce. But it's a multi varient testing using machine learning.

Phillip: [00:31:42] You were pretty close. Episode 14.

Brian: [00:31:44] Yeah. Yeah.

Phillip: [00:31:45] Yeah, and how basically machine learning is just the beginning.

Brian: [00:31:49] And he kind of gave some some high level overview of what you're talking about, so it's worth listening to.

Phillip: [00:31:55] Yeah. Yeah. Anyway. Beware. I guess I would say buyer beware.

Brian: [00:32:01] Definitely.

Phillip: [00:32:02] If you are a retailer. Yeah. If you're a retailer and you have some technology provider that's telling you that they're using deep learning or they're using artificial intelligence, I would be highly, highly skeptical and very cautious because it's only very recently that Google and Amazon have actually productized deep learning products and they'll go to market on those products to actually build them into real tools. What they've actually done is use machine learning, which is a much more juvenile approach to... Not that it can't be used well and applied well. I wouldn't call it AI.

Brian: [00:32:43] Well and another thing is, I think a lot of the algorithms... They got relabeled machine learning.

Phillip: [00:32:50] Right.

Brian: [00:32:50] So a lot of stuff you are already using has been relabeled and re-marketed as, you know, if any little element of it is related to machine learning, marketers are going to say, "Oh, we do machine learning."

Phillip: [00:33:02] So just to explain, and I feel like I'd like to try to explain in very, very simple terms what the difference with machine learning is. It's basically like trying to find a best fit algorithm. So if you've ever worked it with a Scatter Plot in Excel, OK, and you have data points all over a map, you can click a button in Excel that says give me a best fit line. OK, and that best fit line will have a straight line with a particular slope that tries to hit an average or a median between all of the points on that Scatter Plot. Ok, so if you can imagine what that might look like, that straight line through all of those points is really far off the mark for most of those points. It is an average. That means there are major outliers. The difference of machine learning is the straight line. The deep learning actually continues to perform refinements to that line to get it closer to all of the data points. And that is called gradient descent in this particular, in deep learning, it's called gradient descent. And it's incredibly, incredibly intense because not everything can be... It's not a 2D Scatter Plot. It is a twenty thousand dimension Scatter Plot.

Brian: [00:34:28] Right.

Phillip: [00:34:28] But it is the same concept in that it's trying to find a better fit line. So in finding that line then you can make predictions about where a particular data point might fall along that line. That's all this stuff is. But what most people are selling you is the straight line. They're selling you a really average... It's the difference between, you know, a/b testing and one to one personalization.

Brian: [00:34:56] Right.

Phillip: [00:34:58] And most people can't do the latter anyway. Most people are doing a/b testing and calling it personalization because they just don't have the means and the operations in their business to actually do one to one personalization.

Brian: [00:35:11] People are calling everything personalization.

Phillip: [00:35:12] Oh, my gosh. Yeah, don't even get me started. But that's the point.

Brian: [00:35:16] Point taken. Yeah.

Phillip: [00:35:16] If you're a retailer and you're being sold AI right now, or being told you have to bake AI into something or every single product that you're looking at and incorporating into your digital commerce experiences has AI branded on it, don't be fooled.

Brian: [00:35:29] Yeah. Let's make an even bolder statement. If they're not using Google or Amazon...

Phillip: [00:35:37] Yeah if it's not a Google or an Amazon technology or Sentient, maybe.

Brian: [00:35:42] Well I was going to say Sentient because the person that created something actually created the technology that theory is built on.

Phillip: [00:35:51] Right.

Brian: [00:35:52] And so if that technology isn't built on some derivative of those Apple, Google or Amazon technology or Microsoft... Microsoft's got some logit technology, too.

Phillip: [00:36:07] Sure.

Brian: [00:36:07] But of those main players, it's probably not legit. That's a bit of a generalization.

Phillip: [00:36:14] Well, so what has happened is so there's actually again, going super deep on this. I've been really interested in this for some time, but... Google actually open sourced what they created as an extension on or as sort of an evolution for deep learning is like a tool kit called TensorFlow.

Brian: [00:36:38] Right.

Phillip: [00:36:38] And TensorFlow is like a very sort of it's a powerful tool kit, but you don't have to be a data scientist to understand how to use it. Whereas the variance prior like these Python packages like NumPy, you could do similar things and create these deep learning sort of machine learning algorithms with them. But it required a high level of understanding of what you were doing. And I really do think that eventually, if you're listening to this in 2017 or 2018 Right now, be skeptical. But it could be that Google and the rest and Amazon make these products so simple for any founder, any startup to use that it becomes... And they have done that already. You look at Amazon recognition, which is a machine vision sort of service, you look at Polly. Amazon Polly, which is NLP, natural language processing. You look at the services that they've created and any developer can feed data in and get something out the other side.

Brian: [00:37:47] So broad point here. That's really interesting to me how really advanced technologies, things that took insane amounts of money to understand and to do something with, the average professional can now take advantage of them.

Phillip: [00:38:07] Well, that's exactly where we started this whole conversation. The Gutenberg press required like a literal foundry and people to pour liquid like metal into a form and then typeset it.

Brian: [00:38:20] Exactly. Yep.

Phillip: [00:38:20] And then ink it and then press it. And bind that into books. And now any numbnuts can, you know, start a Tumblr and type whatever garbage they went to into the Internet. And but that's the chasm here.

Brian: [00:38:38] No, a hundred percent. I mean, it's Sumerian. We talked about Sumerian last episode. And that's taking a very powerful tool...

Phillip: [00:38:45] If you missed it, just what is Amazon Sumerian?

Brian: [00:38:47] Sumerian is a basically AR VR 3D platform where you can create these experiences, those types of experiences without being a developer.

Phillip: [00:38:59] Right. It's like Microsoft Paint for AR.

Brian: [00:39:02] Paint for AR. They need to take that and do something with.

Phillip: [00:39:06] Yeah.

Brian: [00:39:07] You know they .... Paint. Microsoft did.

Phillip: [00:39:09] Did they really?

Brian: [00:39:10] Yeah.

Phillip: [00:39:11] I didn't know that.

Brian: [00:39:11] Yeah. I mean maybe there's tools...

Phillip: [00:39:12] It's like all these things are dying. Aol Instant Messenger.

Brian: [00:39:15] Yeah AIM's gone now.

Phillip: [00:39:17] Now we've got Facebook kids.

Brian: [00:39:19] It's cause we have things like Sumerian coming out. Why do we need Paint when we have to Sumerian?

Phillip: [00:39:25] Ok. All right. So, gosh, this has been a super deep dive episode. OK.

Brian: [00:39:29] That is not what we meant to do.

Phillip: [00:39:30] No, not at all. I hope you enjoyed this. I've really enjoyed this.

Brian: [00:39:35] This has been fun.

Phillip: [00:39:35] OK. Let's wrap it up.

Brian: [00:39:37] No, hold on. Do we want to wrap it up or do we want to talk about our next topic?

Phillip: [00:39:40] I was going to wrap it up by talking about our last topic, which is like we don't get into this enough, but we are remiss if we don't at least mention it.

Brian: [00:39:49] Ok. We're going to mention it and then we're going to make a commitment. I think we might have made this commitment before.

Phillip: [00:39:53] What are you talking about? What commitment are we making?

Brian: [00:39:56] To get an expert on about this.

Phillip: [00:39:58] Oh, OK. Yes. Yes.

Brian: [00:39:59] I'm sure we...

Phillip: [00:40:00] OK, fine. All right. What are we talking about?

Brian: [00:40:02] Talking about Bitcoin.

Phillip: [00:40:03] Ok.

Brian: [00:40:03] Because it doubled in the last two weeks.

Phillip: [00:40:06] Oh, my gosh. Yeah.

Brian: [00:40:07] Again.

Phillip: [00:40:07] Bitcoin 15,000 is a thing.

Brian: [00:40:09] No, no. Sixteen.

Phillip: [00:40:10] Sixteen thousand as we speak. It's actually going up by the hour.

Brian: [00:40:12] Yeah.

Phillip: [00:40:14] And you know, this is on the heels of two things that happened as at the time of this recording. Bank of America. Again, this comes over Futurism and we'll link it up in the show notes. Bank of America is trying to patent a crypto exchange. Oh, they won a patent for a crypto exchange. So it basically in brief, from the Futurism article, "In a major move for cryptocurrency. Bank of America is won a patent for a crypto exchange system. And this is part of a larger growing movement to incorporate crypto into traditional financial models." And so B of A somehow thinks that crypto is important. And then here's this. JP Morgan just put out basically a buy rating for Bitcoin of 15,000. Now, listen to this. This is what's insane. They said, don't buy at 500. Don't buy at 1000. Don't buy at 10,000. But at 15,000 it's a buy. They're putting some...

Brian: [00:41:12] Time to sell. {laughter}

Phillip: [00:41:13] This is ridiculous. And we know nothing about Bitcoin.

Brian: [00:41:19] I know, man. For being the podcast called Future Commerce, we really failed this category.

Phillip: [00:41:25] No. Well, we don't purport to know everything about everything.

Brian: [00:41:28] That's true.

Phillip: [00:41:28] We need to get an expert on.

Brian: [00:41:29] We really do. You know what else this came on the heels of? There was a major marketplace that was just that was just hacked.

Phillip: [00:41:36] Nicehash hack.

Brian: [00:41:36] Yes. Right. Millions of dollars of Bitcoin lost. You know what's crazy?

Phillip: [00:41:40] We're gonna just keep seeing more of this.

Brian: [00:41:42] There's what? Over 16 million bitcoins out there right now? Is that right?

Phillip: [00:41:46] I don't know. I don't know what the number is.

Brian: [00:41:48] Verify this. Over 16 million bitcoins. Let's just say 17 million.

Phillip: [00:41:54] Are you trying to, like, calculate market liquidity or total market economy value?

Brian: [00:41:57] Yeah. Exactly. Yeah.

Phillip: [00:42:00] I don't even know. I don't know that the economy of bitcoin is. It's growing by the minute.

Brian: [00:42:06] It is.

Phillip: [00:42:07] And maybe it crashes horribly. I know we talked about the halvening about 20 episodes ago, which was like an event where Bitcoin cut in half. That has been erased by an order of magnitude. What I think that this signals, though, to us is that there's a potential for government regulation to step in in a hurry.

Brian: [00:42:30] We've already talked about this. Russia's already on that train a long time ago.

Phillip: [00:42:32] I know. But the United States government.

Brian: [00:42:34] Yeah. I mean, we talked about this already, too. We talked about the fact that it can't go much longer without regulation.

Phillip: [00:42:40] Not at this rate.

Brian: [00:42:41] There's too much money involved. Yeah.

Phillip: [00:42:42] Not at this rate.

Brian: [00:42:43] Somebody is going to step in and say this is not okay. We're not allowing this anymore.

Phillip: [00:42:47] Yeah.

Brian: [00:42:48] Maybe not with this presidency, but definitely maybe with the next one.

Phillip: [00:42:51] I don't know. Anyway, maybe that's a good topic for Danny Sepulveda to talk about from a policy perspective.

Brian: [00:42:59] Oh yes we definitely have to have Danny talk about this.

Phillip: [00:43:02] Yeah. Danny has gotta talk about this. I just find this a sort of an interesting event for us as markets boom across the world, we're having incredible or our economy in the United States is booming right now.

Brian: [00:43:19] More jobs just created recently.

Phillip: [00:43:20] Yeah. It's like an incredible... Defying anything that I said a year ago.

Brian: [00:43:27] Well, I'll make a quick prediction.

Phillip: [00:43:29] Okay.

Brian: [00:43:30] No I'm not.

Phillip: [00:43:30] Well, save it for 2018 Predictions show.

Brian: [00:43:32] 2018 Predictions show is coming soon. Do not miss it.

Phillip: [00:43:34] And do you know how you don't miss it? You got to subscribe and like the podcast, so go to Apple podcast right now. We're closing the show down, by the way.

Brian: [00:43:42] Yeah.

Phillip: [00:43:43] Go to Apple podcast, Google Play, or wherever you get podcasts. Stitcher. We're on Sticher now. Wherever you get podcasts. And you can also stay up to date on everything we're doing at Future Commerce by hitting, signing up for FC Insiders where you'll get...

Brian: [00:43:58] So you can live through the retail apocalypse.

Phillip: [00:44:01] Yeah, exactly. Don't let the retail apocalypse happen to you. What was the thing that you said? {laughter}

Brian: [00:44:09] {laughter}

Phillip: [00:44:09] This is the most amazing episode we've ever done.

Brian: [00:44:12] Oh, man. We gotta do more in person episodes.

Phillip: [00:44:13] So be the first to know about news trends in retail tech, and you can do that by subscribing to FC Insiders at That's it.

Brian: [00:44:19] With that retail tech moves fast.

Phillip: [00:44:22] Future Commerce is moving faster.

Recent episodes

By clicking “Accept All Cookies”, you agree to the storing of cookies on your device to enhance site navigation, analyze site usage, and assist in our marketing efforts. View our Privacy Policy for more information.