Transcript of this episode:
Terrence: Hello internet, and welcome to the Engadget Podcast.
I'm Terrence O'Brien, and joining me this week, on the far left,
senior editor Chris Velazco.
Chris: What's popping, slimes?
Chris: I'm gonna run that one into the ground.
Terrence: Okay, I'm gonna, I'll just let you have that.
Also joining us, managing editor Dana Wollman.
Terrence: And reviews editor Cherlynn Low.
Cherlynn: Y'all are weird. Hi.
Terrence: They are real weird, right?
Chris: Can we also just go on the record really quick and just confirm that Dana's
last name is not Wolfman?
Cherlynn: Just on the record, yeah. That really happened.
Terrence: Wait, what happened?
Cherlynn: That was today, you missed it.
Dana: Were you in the room when I got a call, a no caller ID thing, and I picked up and the guy's like, "hello Miss Wolfman?"
And I said "it's Wollman," and I hung up.
It was beautiful. It was beautiful moment. You had to be there.
Chris: Can I speak to Danna Wolfman please?
Terrence: Danna Wolfman? I like that, that's--
Chris: I'm never calling her anything else.
Terrence:Yeah, no, that is your new name. Somebody out there please, make like a nice gif of just Wolfman.
Dana: Wolfman. I mean here's the thing, Terr-ance.
I've been called Wolfman my whole life.
Terrence: You had a really awesome childhood.
Dana: Yeah. Ish.
All right, you wanna get right to it, guys?
Dana: Yes, please.
Terrence: Let's start the show like we do every week, with flame wars For those who don't know, this is a short debate segment. We're going to have Cherlynn and Chris face off against each other, you're gonna debate the biggest topics of the week, you'll have 20 seconds to make your opening statement, I'll allow you a brief rebuttal, and then Dana will declare a winner based on the strength of your argument purely. We will be keeping score, at some point, a winner will get a thing, and at some point a loser will have to do something terrible.
There will be leader boards, hopefully next week we'll have that up and running. So this is gonna be like a whole thing.
Chris: Yeah. I'm in it to win it, unless the loser gets to eat like, Shake Shack nonstop, it's a yes, in which case I want to lose forever.
Terrence: Yeah, I mean you can do that even if you don't lose, honestly.
Cherlynn: Also, you're already a loser forever, so...
Chris: Yeah, what, I've got my Shake Shack, so you know what? Or In-N-Out, whatever. Whatever that west coast burger chain that people won't shut up about is.
Terrence: I think that's In-N-Out.
Chris: Man, whatever.
Terrence: It's just Wendy's with a different name as far as I can tell.
Chris: That, nope. Nope.
Terrence: All right, we are getting way, way off topic. But that's a debate for another time. (laughing) Somebody start like a side debate in the comments or something.
Dana: And we'll join in. Yep, we'll join in.
Terrence: So I want to start, let's go with Intel. So IDF was this week. Is it still going on, Dana?
Terrence:This will be airing on Friday, it is currently Thursday.
Dana: If it is, it's on its last day.
Terrence: Okay. We're really tuned in, apparently, to the industry and its goings on.
Dana: The news flow has certainly slowed, and now we're in a position to talk about the week that was, Intel news.
Terrence: So one of the big announcements, and...
Cherlynn: I'm sorry, Siri is annoying, sorry.
Terrence: That happened on episode zero as well.
Cherlynn: She's very sassy, she wants to be a part of this conversation.
Terrence: She's very sensitive.
Terrence: One of the big thing that Intel announced this week was Project Alloy,
it's an all-in-one VR headset. So this doesn't need to be connected to a computer, you don't have to plug a phone into it or anything like that, it's just a self-contained unit. I guess one of my questions is, and this is, start with you, Cherlynn. Is this gonna be powerful enough and capable enough to actually do VR any justice? And your 20 seconds will start now.
Cherlynn: First of all, we don't know the full specs yet, we don't know what kind of processor's in Project Alloy, so I can't make any estimates there. But for the mobile and the whole stand-alone VR headset space, this is a big step forward, because for VR to be truly immersive, he wearer needs to feel like they can move around without being limited to and being tied to their computer,
or having to stay in a room with all those cameras.
This is great.
Terrence: You are out of time.
Cherlynn: Oh, you bitch.
Terrence: Pow. That's cold.
Cherlynn: Kinda went a little far there, I'm sorry, you still hired me, so...
Terrence: Yeah, it's all right.
Terrence: I'm used to it. All right, so you, Chris, you're not super into this, you don't think this is a great idea?
Chris: Not sold.
Terrence: Okay. So your 20 seconds starts now.
Chris: All right. Intel's next are gonna be--
Cherlynn: Terrence, you really suck today.
Chris: A little quick on the buzzer there, my friend. Shall I start again?
Terrence: You shall start again. My thumb slipped, I'm sorry guys.
Cherlynn: Sure, your thumb slipped.
Chris: Intel's next are getting pretty powerful, I don't doubt that they'll have the horsepower, I'm just concerned they'll not get the experience right.
You've got to have a battery on this thing, you've got to have, like how comfortable could this possibly be for long periods of time?
Also, is it upgradeable? Like are you gonna have to buy a new one of these things every couple years Like, unless Intel can answer these questions, this is a curiosity, but I don't see it as being a whole lot more than that.
Terrence: All right, and you came in under your 20 second mark, so that's pretty good.
Chris: But I already got buzzed, so.
Cherlynn: You did, you already got...
Terrence: I got that out of my system a little early.
I guess you do make a good point, which is, you know, there's a lot of hardware that you have to pack into this thing that's on your head.
Cherlynn, do you see any issue with having to like, wear a battery, a full computer, and a display, and all these sensors to track like, hand motion?
Cherlynn: I believe that this is an early stage device and in the future, like smart phones have evolved over the years to become thinner and lighter, this Project Alloy headset will get smaller components that are lighter and feel more comfortable around the head. So I think as a preview, this is exciting.
Terrence: And Chris, I mean, would you accept, at least in theory, that you know,
for VR to really like, reach its full potential, it does need to be a standalone device? Or do you think like, at the end of the day, it still needs to be tethered to something?
Chris: I think it does need to be tethered, I think you get much more horsepower out of it, and here's the thing that Cherlynn kind of glosses over, cords are long, okay? Like, sure, you can be in a room and be tethered to a computer, it doesn't mean you have to be within four feet of a computer, right?
Cherlynn: Yes, but to feel like you're still tied to something with a line, with a cable, and also processing power just kind of depends on the cable's latency too, right? Like, if you're using a wire, that's super slow or is like, frayed at the ends or something, you're gonna lose that connection.
Chris: Yeah, but these companies have standards, right?
Cherlynn: Of course.
Chris:They'll be like, okay, use this cable and if you don't, you're gonna have a terrible experience. And then just like, don't be an idiot about it, right?
Cherlynn: Instead of using a cable that has to meet a standard, why not make the whole thing meet the standard?
Terrence: I think we're getting real hypothetical at this point there. Dana, I think it's time for you to declare a winner in this one.
Dana: Point Cherlynn, this is exciting as a preview, and Cherlynn is correct that no one has to put a first generation computer on their head if they don't want to. You might not even want to do that. That might not be advisable.
Chris: Who are you people? This is what we do. We strap first generation computers to our faces.
Terrence: Yeah, I don't know about you guys--
Dana: So other people don't have to.
Cherlynn: But then we hate on it, too.
Terrence:I certainly walked around with a PCAT strapped to the side of my head in the 80s, it was great.
Dana:You sure it wasn't just your hair?
Terrence: Just don't worry about it.
Terrence: It's okay, I'm old, we've established this on past episodes. If I refer to an IBM PC Junior, would you know what that is?
Chris: I have heard of it.
Terrence: Okay, all right. Let's just move on. You're making me feel bad about myself.
Cherlynn: Thanks Dana.
Dana: All right.
Chris: That was...That was concerning.
Terrence: Yeah, I think there might be some collusion going on.
Chris: All right, let's--
Chris: Dana's smiling, by the way, infinitely preferable to Dana sneezing.
Dana: That's not true. I prefer sneezing. Anyway.
Terrence: So, earlier this week, McDonald's basically ruined many children's days by replacing the toy in their happy meal with a fitness tracker. This is, but to be clear, also not a super fancy fitness tracker in the vein of Fit Bit, this is a pedometer that you just wear on your wrist. They have already pulled these, by the way, because of irritation concerns, so we're talking about this in a kind of, is already a story that's basically over. But still, I thought it was super interesting.
And Chris, I mean, clearly this is well-intentioned, they're trying to encourage kids to be active, but is this something that can actually make a difference?
Chris: I think so, I think individual choices are shaped by lots of factors around them, and you've got to figure, McDonald's is an American institution, like it or not. They're gonna sell millions of happy meals whether we gripe about it on a podcast or not. At least they're giving kids the option to be like hey, well, I've got this thing now, maybe I can make a healthy choice. Maybe I can walk for a little bit. And honestly, it kind of beats having like a crappy 20 cent toy anyway.
Terrence: You watched me do it and you just like, cut yourself off.
Chris: I know. I know.
Terrence: You were waiting, waiting, waiting. Cherlynn, you kind of think this whole thing is ridiculous, I guess?
Cherlynn: Mhmm. Chris, you're wrong. First of all, these fitness trackers are not good enough, like he said, to do anything productive for the kids. Also, are they gonna look at their steps and distance? Kids are not sophisticated enough to understand that, oh, distance traveled and calories burned are good for me in this way, this way, and that way.They are, it's too young,their target audience, happy meals? It's kind of off for a fitness tracker.
Terrence: That is your 20 seconds. That's pretty good.
Cherlynn: I'm done. Done.
Chris: Done. So basically your argumentis that kids are dumb.
Cherlynn: Yes. (laughing)Kids who eat at McDonald's.
Chris: Which is every kid! Like, every kid eating happy meals!
Terrence: Are you gonna take issue with the basic premise that children are dumb?
Chris: Not necessarily.
Dana: Not dumb so much as miniature dumb people.
Cherlynn: They are.
Chris: But to your earlier point, like yes, these aren't sophisticated, but like,
frankly you don't need a super sophisticated fitness wearable to be like oh, I'm walking more. This is good. Like, these aren't complex arguments you're having internally, right?
Cherlynn: Well first of all, you don't need a very complicated fitness tracker that McDonald's is giving out because you can just use your phone.
Chris: Yeah, because these kids have phones?
Cherlynn: These don't need to exist.
Chris: These kids have phones?
Cherlynn: These kids could use phones. No, some of these kids have phones. Kids are getting phones at a younger and younger age, and a lot of these have pedometers, that aren't as cheap as crap. So there are lot of budget phones these days.
Terrence: So, but do you... I mean I guess, I understand what you're saying in that, you know, these are very basic devices and you know, kids being dumb and not sophisticated enough to understand this I might say is a little bit of an oversimplification, but a fun one, to be fair. But don't you think that there's something to the idea of just encouraging kids to be more active in general? Whether, you know, even if it's through this cheap piece of plastic that comes in a happy meal.
Cherlynn: Right. I'd rather McDonald's work harder on its nutritional information and education as opposed to just dumping a kid's toy in a happy meal.
Chris: Sure, but there's no reason it can't do both. Like, nutritional transparency has been a thing in fast food restaurants for years. Whether or not it works is
up to the people who buy it and buy into the stuff. But like hey, you're giving
people an extra tool. Like, there's no downside to that.
Cherlynn: I wonder if it's kind of like almost, not a placebo, but kind of an effect where it says here, we're giving you this so it's a little bit healthier, so eat more of our food. You know what I mean?
Chris: I mean, it's super cool to be like, cynical about corporate motives. And that's not incorrect either. But at the same time, like I've said people are gonna buy their kids happy meals anyway, right? Like, that's a decision that isn't, in a lot of cases, being made by the child. At least they now have this thing where they can say hey, okay, I can do something about this, if I choose to, or I could sit in front of the TV, it's my choice. Like, it's offering information, and what you do with that is sort of your own prerogative.
Terrence: All right, Dana.
Dana: Point Fiasco.
Chris: Fiasco with a Z.
Dana :McDonald's does have some responsibility to help correct the obesity problem which it helped blow out of proportion in the first place, and I would agree that kids aren't dumb, per se, and they're not, it might not be purely their decision to get McDonald's in the first place, so why not have something that encourages them to move around a little more. Even if it does give them rashes.
Chris: Also, it's a cool looking watch. Like, what kid doesn't like watches?
Terrence: Is it cool looking?
Chris: I don't know.
Cherlynn: It did not look cool.
Terrence: Yeah, I'm gonna go with no on that one.
Cherlynn: Not on the looks. Okay.
Terrence: All right. It is now tied one-one. Let's move on. So, Elon Musk said a thing this week, that made me scratch my head. He's taking his Open AI project and he's going to teach it how to speak and understand language by forcing it to read Reddit comment threads. Yeah, well, I'm just going to throw it out there that I'm already with you, Chris. But I'm gonna let Cherlynn go first and have an opportunity to convince me and the rest of the people at this table that teaching a machine to speak using Reddit is actually a good idea.
Cherlynn: It is, because Reddit's one of the places where, you know, fresh new internet jargon is dropped all the time, so where better to pick up this vocabulary for a machine? Also, it's a very, it's an existing publicly available library and database of words and how people communicate with each other. You don't need to intrude on someone's personal data, like going through the messenger app to find how people talk.
Terrence: That's pretty good.
Dana: I also want to hear the freshest new internet jargon.
Terrence: Yeah, we'll come back to the fresh internet jargon in a second. Because I don't think I know it. Chris.
Chris: All right, so I'm going to avoid the easy argument. I think it's easy to say that we're all on Reddit pretty frequently, sometimes Reddit is terrible,
let's move beyond that. I think what's really interesting here is that the demographics of Reddit skew very heavily towards young, generally
like liberal guys, and that's cool, I don't know that it's necessarily how I'd want an AI to sort of interpret and speak.
Terrence: Cherlynn? Any rebuttal there?
Cherlynn: I wasn't paying attention. I was looking for fresh new internet jargon. I'm sorry.
Chris: I've got it for you. What's popping, slime?
Cherlynn: No, no. Like TIL, TFW, MFW.
Terrence: Oh, you mean the things that I have to Google literally--
Cherlynn: didn't have them, but they're pretty frequently there, yeah.
Terrence: So, I mean...
Dana: Can we teach a bot to say "says who?"
Dana: In response to everything?
Dana: That's the new internet jargon.
Terrence: That's the new internet jargon.
Terrence: Says who?
Chris: We're, like, Cherlynn and I feel super old right now.
Cherlynn: Yeah, we are.
Terrence: Says who?
Dana: Says who?
Cherlynn:Wassup? That's you guys, right?
Terrence: Yeah, I think you're right, Chris, in that, you know, it's too easy to make that argument that like, Reddit is a cesspool filled with horrible people.
Chris: And like, let's give it to the Open AI team, like they're not gonna actively feed this thing like racist stuff, right? Like, they'll filter out the worst of it, presumably.
Terrence: Well, here's the thing is, I don't know that that's true.
Chris: I don't think any of us knows that that's true. I'm really just giving them the benefit of the doubt. Because if they don't...
Terrence: So well here's what I'm gonna say, and I wanna hear what you have, your response to this Cherlynn. Like, I honestly don't think they're going to filter it. I don't think there's any way they can, on the scale they plan to do this, appropriately filter out potentially offensive content. So… Isn't there like a real danger here, that basically you're just getting an even more complex TAI at the end of the day?
Cherlynn: Yeah. I mean, I'm hoping that they teach it to learn not just using Reddit, but also other sources of information, because Reddit is kind of, like we have touched on, a deep dark wonderful world. But it's... But I believe that they need to not filter out anything. They need to let it learn every possible thing, and then assign meaning to it afterward. So for example, if it stumbled on a thread where people were being racist. It needs to absorb all of that and then understand how people are reacting to racist comments, and then learn to avoid that. It's just way too sophisticated for this.
Dana: Is there room for me to ask a question?
Dana: What about on Reddit, guys, all of the specialized Reddits for people who have very intense, particular interests, and a lot of time to spend on the internet? These people aren't necessarily saying racist things, they might be delving really in depth into nerdy topics. What about those people? Is there anything to learn from all the varied discussions they're having?
Chris: I think there's a lot to sort of dig into as far as sort of conversational tone and syntax and structure, and I think that is crucial. I think that's what the Open AI team is really really going for here. I think on the flip side, it is sort of easy to sort of look at that as all valid stuff, and a lot of it just isn't. Like there is, basically they're trying to teach this thing how to interpret and sort of speak and sort of inform its approach to language, which is great, but doesn't always lend itself well to just sort of listening to everybody. Like, I dunno, I think it's maybe fair to say that some people are worth listening to more than others as far as how you want to educate this thing.
Terrence: You're talking about just kind of avoiding that idea of like that false equivalency thing.
Chris: Right, yeah. Like, not all things are created equal, not all viewpoints are equally valid. Some of them are just batshit stupid. So let's maybe not feed those to an AI.
Cherlynn: Well, where else would you suggest this system learn how to communicate, learn how people are communicating?
Chris: I'm not saying don't use Reddit, I'm just saying be smart about it. Like, filter out shitty subreddits, like don't--
Cherlynn: Kappa should be out.
Dana: Should we be using written exchanges at all that aren't necessarily, what's the word I'm looking for? Immediate real time. So, should we instead be using live, real time chats? Conversations between humans.
Chris: If you want it to get real, real nasty, real fast, I would say yes.
Terrence: Well, I... (sighs) I think we're getting a little bit off topic, but I think to Dana's point, there is some truth that, you know, there's only so much about naturalistic interaction that Open AI is going to be able to learn from Reddit. These are not real-time conversations, these people are not reacting to each other in...I can't find the word that I'm looking for, but you know, it's not... It's often...
Terrence: Delayed, thoughtful--
Dana: It's like a turn-based game.
Terrence: Yeah, you sit there and you think about what you're gonna say.
Dana: Wait half an hour to think about how I'm gonna--
Chris: Yeah, but is that not better? To sort of have inputs that are based off of like, thought as opposed to just like, gut instinct reactions?
Terrence: Well I mean, I think, I think in some ways yes, but I also think it's less natural. So I think, you know... You're not going to build a chat bot or a virtual assistant that's going to respond to you in a natural, believable, almost human way simply by studying comment threads. Like, it's just, it's a very different form of interaction.
Cherlynn: I mean, there are some comment threads that are faster than others, such as like in an AMA, they tend to respond a little faster. Reddit supports that system as well. But I see what you're saying, and like I said, it can't be the only source of where the AI is learning. Reddit can't be the only place.
Terrence: All right. Before we go any deeper down that AI hole, Dana?
Dana: Point Fiasco and also Terr-ance. Ha. (laughing)
Terrence: I don't think I can get a point.
Chris: Does that mean I get double points?
[Dana and Terrence] No.
Dana: So internet speak is like a second language for people who already speak human. I think if you're a robot who is learning human speak as a first language, none of that internet speak, none of the, the way we speak in comments is not going to end up in its proper context, the way it would if human language were your first language. I don't think that's exactly what you said, I think it builds on a lot of what you said.
Cherlynn: Dana just said it better.
Chris: Yeah, that's fine. (Laughing) That's why she's my boss.
Terrence: All right, last topic of the week. Google Duo. So... (Laughing)
Dana: She's like, nope.
Terrence: Yeah. No, I mean, so for those who don't know, was it May? Was Allo?
Chris: Allo, yes.
Dana: Back in May.
Terrence: Back in May, Google announced Allo and Duo.
Terrence: Two new communications apps to kind of sort of replace Hangouts. Allo has not launched yet?
Chris: No, it has not.
Terrence: But Duo launched earlier this week, I forget, Tuesday? Monday? Does it matter? And that's the video chat portion of the equation. It's super simple, I guess it's supposed to be their competitor to Skype or whatever.
Dana: Or you know, Face time.
Terrence: Yeah, Face time, I guess that's a better equivalency. V, I know you think that Google is finally kind of on the right track here.
Chris: I do. Is it perfect? No. It's very simple, I think that's by design. It has a clear sense of purpose. As Nate pointed out in the story, I think what's really important to take away here is that the core, the foundation of Duo, works really well. It's a really simple to use video chat service that works even on low-end devices, so everyone can use it. Like, if the foundation's fine, they can make it pretty and fun later. It works. (buzzer) (laughing)
Terrence: You really got a kick out of it.
Cherlynn: I did, you suck.
Terrence: Bring it.
Cherlynn: Mhmm. Google doesn't need yet another communication app or service because they already have Hangouts, which may or may not be replaced by Allo and Duo, but I don't need two apps replacing one. Why can't they embed the video feature into Hangouts and call it a day? Also, the knock knock feature, I know it can be turned off, but having it turned on is very annoying, because you don't want all the unsolicited dick pics you're going to get with that preview video. (buzzer)
Terrence: I'm going mad with power. (Laughing)
Cherlynn: You are. You already are, actually, so...
Terrence: It's true. So, can you, before I give Chris a chance to offer his rebuttal, can you explain to people what the knock knock feature is real quick?
Cherlynn: The knock knock feature is when someone's calling you, on your lock screen you can see a preview of their video feed, so you see what their camera is showing, before you even accept the call, which I find to be kind of rude and kind of an invasion of privacy, because I get a lot of calls on Face time from random strangers out of nowhere, and if they were able to just show me whatever the crap their camera was seeing before I even said yes or no, that would be very annoying. Of course, it's good that you can turn this feature off, but... I don't like that it even exists.
Terrence: Okay, so I have a quick question that's gonna derail us slightly. You get random calls from strangers on Face time?
Cherlynn: Yeah, I do.
Terrence:Dana? Has this ever happened to you?
Cherlynn: It's really weird. I don't know, I used to get that all the time,
and it'd be like a really long string of numbers, so they probably are in China or some European country, and it will show my, like, front camera, it'll show what my front camera sees, because it's in the shading of Facetime—
Dana: I just get traditional voice calls from strangers calling me Wolfman.
Cherlynn: There you go.
I guess, I don't have... Dick pics, thank god I don't have those yet. Jesus, don't send me any.
Terrence: You really just entered a world of trouble. We should edit that.
Chris: I mean, like I'm sorry that that happens to you on Facetime, because that's effing ridiculous. Like, that's hot trash. I mean, as far as I know, Duo sort of limits itself to your circle of contacts, so unless you have, just like really terrible friends,
which we all kinda do.
Cherlynn: Or former Tinder contacts.
Chris: Which, so at that point, they should probably not be in your contacts anymore. But beyond that, I think it's easy to say, like, sure, this is one of many, and do we need all of these separate apps that kind of aspire to do the same thing? But I mean, it's Google. They kill products, they killed Wave, they killed Buzz, they killed all these things. It's not only not... Basically, I'm trying to talk a little too much for not a whole lot. But they're gonna kill a bunch of stuff, they're gonna consolidate, like, it's a natural course of action, I firmly believe it's gonna happen. Like, it's not gonna be as cluttered an ecosystem as we have now within a year or two, guaranteed.
Cherlynn: Right, but then they shouldn' have pushed it out as a new app, they should've pushed it out as a new featur as part of Hangouts.
Dana: Is Hangouts savable?
Dana:Yeah. I think, I mean, okay, so neither of you have said this yet. Hangouts is kind of shitty.
Cherlynn: I know, I've moved away from Hangouts.
Chris: Wait, what's your beef with Hangouts?
Terrence:Yeah, am I the only person who like, likes it, use it...
Cherlynn: I use it.
Chris: I use it for like, certain people who won't use anything else, but I don't actively dislike the process.
Dana: I don't love it on either mobile or desktop, I especially dislike it on desktop.
Chris: I hate it on mobile. Like the one thing that gets me is that, like this is a Google service, you should be able to search your chat history on your phone. –
Terrence: That is my one big issue with it, honestly, otherwise I'm kind of fine with it. It does what it needs to do. I kind of felt like Allo and Duo were just like, one of these things that just kind of don't need to exist. I mean, sure they might be fine apps.
Cherlynn: I think to answer your question, Dana, if Allo's features and Duo's features were embedded into Hangouts and they did fix some of the problems that we already have with Hangouts, Hangouts could still be saved, because it's very prevalent. A lot of people have it, all Android phones come with it, it's a very popular system.
Dana: Hangouts, by its nature though, is nowhere near as simple as, or pared down, as Duo is. So how would you make Hangouts like Duo, but without sacrificing a whole bunch of features?
Cherlynn: I would just replace it, the existing video calls on Hangouts function with Duo.
Chris: Yeah, I mean I think, I mean this is the one part where we kind of agree, right? Like, I think in a lot of ways it's easier to look at Allo and Duo as sort of like test balloons. Like sure, these could find their way into core Google services eventually, now they just want to see how people like them, right? And you could fold that into Hangouts, but there is a ton of architecture there already that you'd have to re-engineer to kind of get things working as simply as Duo within video calls for Hangouts. Like, that's just a ton of work. They can do it, they just need to know if people are gonna use it. If it's good, you know?
Terrence: All right, Dana.
Dana: Well this is tough, because you guys agreed on a lot of things.
Chris:We're actually secretly kind of friends.
Cherlynn:Okay, let's not tell anyone that.
Chris: Oh, sorry.
Terrence: You guys hate each other.
Cherlynn: We hate each other. That's the official statement.
Terrence: Just uncontrollable hatred.
Chris: We've been fighting on Twitter the whole time we've been talking. I don't know if you knew that.
Dana: Oh, this is tough.
Cherlynn:Don't eenie-meenie-minie-moe this.
Dana: Fiasco again, I'm sorry Cherlynn.
Dana: I appreciate the point you made that the app was designed to work on a variety of devices. Low end ones including with poor connections. That is a useful base to start with. Even if this does, as you say, end up in the Hangouts app later on. Although it doesn't sound like Hangouts, in the future, will be meant for consumers like us so much.
Cherlynn: We'll see.
Terrence: Sounds like it's gonna be primarily business focused, that's their line.
All right, so, we end, Chris wins this round three to one.
Cherlynn: I'll have all the Shake Shack. I will eat all the Shake Shack that you won't get to eat.
Terrence: Don't worry, there will be plenty of time to catch up and crush him.
Cherlynn: It's okay, I'll beat Devindra, it's fine. I'll beat Devindra. Or Nate.
Terrence: Or all of them.
Cherlynn: Or all of them.
Terrence: Physically beat them.
Cherlynn: I'll physically beat them.
Chris: I mean, I'm also fine with us just being like, tied for first place.
Terrence: You can also do that.
Cherlynn: Oh wait, no, we're not friends.
Chris: Oh. I hate you. And everything you stand for.
Cherlynn: Your face loses.
Chris: I don't like your hair.
Cherlynn: My hair's beautiful.
Chris: That's not true, your hair looks good today.
Cherlynn: I know.
Terrence: And now it's time for little bit that I like to call The Worst. I'm sure most people out there are at least somewhat familiar with the idea of using the
backspace key on your keyboard as the back button in your browser.
This is something that a bunch of Internet browsers have had for a very long time, and in May, Google made the long overdue decision to remove that feature, because it makes no sense. And, for whatever reason, I guess to placate the weird, angry people out there who were super upset that, yeah, and we'll get to you, don't worry.
Chris: I’m just making a face but go on.
Terrence: That this feature was pulled out of Chrome. They created an extension
that adds it back in. Why, why, why? Who wants this feature Let's…let’s be super honest. Backspace is the back button is super outdated. It predates web apps.
It's probably older than Internet browsers. I'd have to look. Somebody out there will probably tell me it comes from eMacs or something, yeah. So, but like, it's way too easy to just accidentally go back a page, empty out web forms. I've actually lost concert tickets because of this feature where I failed to click cleanly into the text entry field, hit the backspace key, and lost tickets to, I forget what at this point, and it sold out in the time it took me to refresh the page. I have lost posts that I've written for the site.
Dana: Yes, I have too.
Terrence: What monster thought that this was a good idea? What insane person decided, "You know what we're gonna do? "We're gonna have a one-button keyboard shortcut, "a button that you use on a fairly regular basis "to edit text, which is a lot of what you do on the Internet "is enter text into things, "and then make that this trigger "to lose everything you're doing." Beyond that, it's just poor design. Keyboard shortcuts should not be something you should be able to accidentally do, and especially not one that's so destructive. Like, Control + Alt + Delete, you're not gonna accidentally do that. Command + W, you're not
accidentally doing that. Any one-button keyboard shortcuts are keys that you're not using. Why does F5 for refresh work? Because F5 has no other function
when you're in a web browser. It's terrible, terrible design. It's a stupid feature that should have been removed a long time ago. It's basically just the worst thing in the world.
Dana: How much redder his face got.
Cherlynnn: And I was gonna say, you're so red.
Chris: Fired up! Terence do you use Chrome?
Terrence: I use Chrome.
Chris: I use Chrome. You know what's never happened to me?
Any of the things you just said. So is it an issue with the design of the
browser itself, or just
(in unison: Chris, Dana, Cherlynnn)
Dana: You've never opened up our CMS to, let's say correct a typo in a headline
and you hit the backspace key too quickly, and then you're just launched
out of the post that you're trying to correct?
Chris: No, no! And also by the way Terence, to your point about --
Dana: Are we just living in a personal hell?
Terrence: I think so.
Chris: Wait. Possibly, and to your point about writing a post in the CMS and like hitting backspace accidentally and like going back and losing it, why are you writing in the CMS in the first place, man?
Terrence: I know, look, look, look.
Chris: I think you told me that my first day at work!
Terrence: Yes, and I learned that the hard way when I first started here.
(laughter) I was saving other people the trouble of learning, don't write posts directly in the CMS.
Dana: I still write directly in the CMS.
Cherlynnn: Me too.
Terrence: Why, you guys?
Cherlynn: But I never made that mistake of the backspace, I know exactly where to put my cursor all the time. And I know where to have my active fields and my inactive fields.
Dana: Well sure, Lynn never makes mistakes.
Cherlynn: I never do you guys.
Chris: You know why?
Cherlynn: Although I did just lose flame wars, so.
Terrence: Yeah, you've never just like missed, like clicking in the text field cleanly and like --
Cherlynn: Once or twice.
Terrence: And didn't notice or like bumped the track pad with your palm, I mean, it's also not like it's super hard --
Terrence: To just like hit Alt, left., is it really that hard? Or, if you're on a Mac
and you have a touch pad, like, it's a gesture, like three fingers or two fingers or whatever.
Cherlynn: In the same way it's also annoying because I accidentally swipe left all the time when I'm trying to scroll up and down and I go back and forth sometimes by accident. So I do make mistakes, but I've never used the back button mistakenly.
Dana: You are in for a world of hurt.
Cherlynn: Man, Dana's gonna edit me so hard now.
Chris: I'm really concerned for the people out there who have parents that use
computers and have been doing this for decades. Because as you point out,
this is a long-standing practice. So I wrote the story about the Chrome extension yesterday. I did a little research because it's really like strangely fascinating, and so, using backspace to go back did not exist in most of the very first actual web browsers. It didn't exist in Netscape Navigator, the first version. So far as people have been able to discover on the Internet, it showed up in Internet Explorer maybe around '95 probably because Microsoft decided to use backspace as a hot key to go back up one directory in Windows Explorer, and that's kinda why it's there. just for consistency. So, if we're going off of that history, it's been a thing for like 20 years.
Terrence: Um hm.
Chris: People have been using it, this is, this is just their expectation
of how a computer works, and, if you're a parent who's doing this and it just changes on you, like I would hate to be the kid getting that phone call at three in the afternoon saying my backspace doesn't work anymore fix my computer please.
Terrence: I get that, and I understand the consistency thing, to an extent but yo, things change, we don't use the same keyboard shortcuts for absolutely everything and I am, also, to be clear, a big fan of keyboard shortcuts, I use them constantly, I think people who don't know keyboard shortcuts, and I watch them like use like the context left click menu, (laughter) drive me nuts, but like just from moment one, that was poor design basically, and it's something that shouldn't have survived and shouldn't have been propagated. Yeah, by the late '90s most keyboards that like came with your Gateway computer had like a dedicated browser back buttons. My Think Pad has a dedicated browser back button, my old Think Pad, which is now 10 years old has a dedicated browser back button.
Chris: Those are gross.
Chris: I hate those.
Cherlynn: Why are you --
Terrence: I'm not saying that I like them but I'm saying, they're there like on most keyboards at this point. Why did I need to use back space to do it? You know what backspace does? It erases, dang it erases text! That's what it's purpose is. (laughter) It's not to go back! It serves two different purposes, God!
Chris: So, like I'm never gonna change your mind, I'll throw this out there. Some people like to navigate with just the keyboard. God knows when I was like 12 years old, reading fan fiction on the Internet, I would put the mouse over there,
I would use my arrow keys to scroll up and down if there was a lot of text I'd use the space bar or page up and page down and backspace to go back and it's just how it worked.
Chris: Maybe this is an emotional argument because I loved that that's my childhood.
Terrence: I'm with you, I mean I'm the guy, I don't scroll I hit space when I'm reading the web, I like page up and page down. One of the the things I loathe about Mac Books is there is no dedicated page or page down button like a thing that on a Windows laptops just fills me with joy, but even still, backspace as a back button doesn't makes sense. And if you're somebody who is dedicated to using the keyboard to navigate, is not super hard to hit two buttons at the same time which would be a smarter, better designed keyboard shortcut.
Chris: We're never going to agree on this.
Cherlynn: We're never going to agree.
Terrence: We're just gonna keep stomping --
I hate to agree with Velasco but I agree with him. I think it's because how and when we grew up.
Chris: Oh, so you think it's a generational thing.
Cherlynn: It might be a generational thing because we, like your point before, we grew up being used to it. I avoid backspace key sometimes.
Terrence: So are you saying that because I predate the Internet I hate it?
Chris: And this is and this is an argument like no one younger than us will like ever have.
Terrence and Cherlynn: Yes.
Terrence: I can't wait for the hate mail to roll in for this one.
Chris: We got some really, so Nick first wrote the story about Google thinking about doing this back in May.
Chris: The comments on that story were so, so, good it was an even split too.
Chris Some people were terrified and some people were like, yes!
Terrence: Now it's time for our round table discussion. It is time for group chat. This is where we talk about the big topic of the week, and this is big, big.This is like real heady stuff, guys. We're gettin' real weird here. Real deep.
Cherlynn: When do we not get weird?
Chris: Yeah, have you been to our office?
Terrence: Yeah, I mean well usually we get real weird off camera, and off the air, and we're gonna get a little weird and on camera and on microphone now.
Chris: Exactly what we've always wanted.
Terrence: So if you have been reading the site this week, you may have noticed that we are doing a week-long celebration of artificial intelligence. This is Engadgets AI Week, and we've been covering it from all sorts of different angles, and I think, one of my favorite pieces that has gone up this week, and I love them all, to be clear, is a piece from Jess Condit about whether or not yo, this idea, we even understand what artificial intelligence is, because we don't necessarily know what intelligence is, and so I kinda wanted to take this step back and look at this from a really big picture perspective, and I guess let's start with that like first fundamental question, Cherlynn, do we understand intelligence enough to create an artificial version of it.
Cherlynn: No, we don't. I don't understand intelligence. I'm not even intelligent enough to understand your question, really. (laughing) I mean, sometimes I code, sometimes, well, the human mind is kind of limited. These big conversations, chicken or the egg, like so many things, is there a God, these things are similar to the question you're asking me, basically, do we understand intelligence? Do I understand intelligence? I don't. It's beyond the,
Chris: Speaking of not intelligent.
Terrence: Siri really wants to take part in this conversation.
Cherlynn: Beyond the biology of how a brain works, I don't think we understand how, how to program intelligence, how to create intelligence.
Dana: I think we know how to program it, I don't think we understand what it means to be a sentient human being. It's kind of like a game of Whack-A-Mole, right, you can think of all sorts of algorithms and if this then equations that can make a computer smarter and make better decisions but all of that relies on input, and it is a game of Whack-A-Mole, because you can think of one,
but there's still an infinite number of unaccounted for possibilities
Terrence: Do we need to even understand them? Do we actually need to fundamentally understand what intelligence is and how the human mind works to create an artificial version of it, or is it, like you're saying, is it just something that we can eventually get through through Whack-A-Mole algorithms and just like not even truly reverse engineering, just kind of like Band-Aid upon Band-Aid.
Cherlynn: I think we need to live with the fact that we're not going to build something that's completely intelligent. That's what we need to--
Chris: Well it all sort of goes back to what your definition of intelligence is, and the thing is, there are as many definitions of intelligence as there are people to offer opinions on what that definition is. We're not gonna come to a consensus as to what it is. Whether or not that informs how we approach AI construction is another story, but I think Dana brings up an interesting point, like (sighs) there's a lot,
I'm getting kind of fired up here. I think it's a little anthropocentric.
I think it's a little silly to just assume that the only way to build an AI is to understand how our brain works, because we're working, when it comes to intelligences that can understand and interpret abstract information we have a sample size of one. It's us. And sure, that's great, and we can build off of that, but it feels presumptuous to think that that's the only way to do it, and like there's no, I don't know.
Terrence: So you think we're using the wrong benchmark, perhaps? Even to begin with, because we have just that one.
Dana: Or the wrong species?
Chris: Or the wrong, like sure, we might be better off building as based off
of like whale thought. Or not.
Terrence: Well, I mean I think that's an interesting question isdo we have to fundamentally rethink whatever our definition of intelligence or something that we'll hopefully get to in a little bit, or like creativity, because our sole understanding of it is human intelligence and human creativity and human consciousness, and do we have to create a separate definition for machines?
Is there a different understanding for machine-created works, or--
Chris: I think it's fair to say that's the case. I think some AI researchers already hold sort of a different definition of intelligence. I don't know that we'll ever be able to create something that sort of feels and acts and emotes humanly, because I think there is, I couldn't speak to the biological mechanisms, but there's stuff inside us that doesn't—
Dana: It won't emote because it won't have feelings.
Chris: Sure. But I mean, one could argue that that's, that feelings are like a cascade of brain chemicals, and once we understand how those function, we could theoretically figure out how to replicate that.
Dana: Well, and this is getting really deep.
Terrence: That's the point of this conversation.
Dana: Would we have feelings without a sense of self and individuality?
Dana: And is that something that you can replicate on a machine?
Chris: I don't know. I think what is eventually gonna happen is scientists will continue on their path of trying to replicate the human brain in silicon and in code, but we'll also see people just sort of like trying to piece stuff together, and in search of hitting that tipping point where all these connected systems sort of transcend the system and become this sort of cohesive whole, this sentient--
Cherlynn: Is that dangerous, though?
Chris: Who knows? I mean, I don't think any of us are qualified to say it, but it's interesting, and it's something we're gonna do anyway.
Terrence: Yeah, I think the danger question is probably barely worth considering at this point.
Cherlynn: Oh, thanks.
Terrence: Well, no, no, it's not like its, I think yo, Elon Musk and all these people who are really sounding the alarm are clearly super-concerned about how machines are going to enslave us and turn us into like house cats or something, but I think Jess makes this really important point in her piece and I think others have made a similar argument which is not only do we not understand intelligence, we don't understand artificial intelligence, and it's probably too early to really be passing judgment and getting super concerned about becoming slaves to the machine. Quite literally. I think you kind of touched on something interesting though, there, Chris, which is this idea of worrying less about mimicking the brain and trying to recreate this, and there will be people who do that, and more about building this patchwork of things that's, gets past the system, as you put it, and I think at the end of the day it's, maybe I'm misinterpreting what you're saying, it's more about at that point mimicry, and achieving the end goal of appearing like conscious, if not actually being it, and one of our other editors, Aaron, did this great piece in November where he tried to teach a neural network to write a Engadget story, and then he did a similar experiment this week for AI Week in which he fed a whole bunch of Engadget stories to another system called Wordsmith, which is what the Associated Press and a couple of others use to generate sports stories and finance stories and stuff like that, and while the neural network is arguably the more advanced system, it kind of tries to mimic the way the human brain works and make all these connections, and it did spit out like what would qualify as English sentences. It could not write a news story. It didn't know how to piece together--
Dana: It was not a post we were proud to have on Engadget.
Terrence: Yeah, no, I mean, we, to be clear ran excerpts of it in line, in the story, for a reason.
Dana: We did post it on the site, but with a disclaimer, I believe.
Terrence: Oh, no, no, I'm talking about the one from November. Sorry, no, the one we ran, the one from this week was written, it wasn't written by a neural network, it was almost like a programming language, basically, where Aaron fed it a bunch of conditional rules and tables of data and then it wrote this post about the Note7, which we did publish and I will make sure that that ends up in the description and you can read it and read Aaron's piece which--
Dana: But even that we weren't terribly, it wasn't a messy post, but it also wasn't a very smart one.
Chris: No, it's very like Engadget 1.0. Like hey, here's the thing, this is what it does.
Dana: Here is this thing, these are these facts.
Cherlynn: Exactly, I heard it like that in my head.
Terrence: But it did manage to pull some context out. It wrote what is, at the end of the day, a coherent story. It wasn't like a pile of gibberish. I guess I'm still not worried about the machines taking my job, but then again, I think as we've shown this week, it's not an easy thing for a machine to do is kind of put context around these things.
Cherlynn: It was kind of scary good how coherent it was, though. I looked at the post.
Chris: I read the story and was like, okay, this isn't a great story, but hey, Toby Golby, okay, you're not awful at this.
Cherlynn: I also thought it was written, before I knew it was written by a bot, I thought it was written by our database editor, Kris Naudus, because that's some of the posts she does, and I was like, "Oh yeah, "this is totally Kris writing this."
Chris: Yeah, do we have a disclaimer in that post saying this was in fact written by Wordsmith?
Terrence: Yes, we do.
Chris: Okay, well I glossed over that part. I'm like, okay. Toby you know, maybe.
Terrence: We are almost out of time, unfortunately but I do want to ask one more question. I have a whole bunch of things that I'd love to address that we just simply don't have time, but I guess, one of the questions that this brought up for me, this whole experiment with teaching it to write is did that neural network, or did Wordsmith create this post? Did it write the post, or did it simply generate the story, and is there an important distinction between those two things?
Chris: There is an important distinction.
Chris: It generated the post, Aaron poked it sort of along the way to get it, like he wrote the first sentence. He gave it a seed to work with, and then I think, and I'll have to double check in the story, but he does kind of guide it along. There are a couple places where I think he intervenes and makes it slightly more readable in certain ways by interacting with Wordsmith directly, which to me just feels like, like this is a thing that anyone could do. It just took some words and put them in a framework that sort of makes sense. Is that creation? No. And I think that the job that we journalists have to do is write stories that these things can't, but even if, even with that said, the stories that I write and the stories that you all write and the stories that we like to read on Engadget are the ones that have nuance and context and soul, and unless you can convince or teach, I don't know, Wordsmith or something like it.
Dana: How to have an opinion.
Chris: How to have an opinion and like f-ing stick to it, you know? You will never, ever get an AI to write a story about the backspace key that will be as fiery or as fun to read as what you just did 10 minutes ago.
Terrence: (laughs) I will take that as a compliment. Dana, any last thoughts?
Dana: No, not for me.
Terrence: No? All right, Cherlynn?
Cherlynn: I wish I were intelligent enough to be a creator of AI. That's all my last thoughts are.
Terrence: Yeah, we are unfortunately, probably all at this table not nearly as smart as any of the people who are doing this work, and I think.
Cherlynn: But if anyone could do it, they could.
Terrence: Yeah, it's super fascinating, but honestly, I at times barely understand it myself.
Chris: Yeah, but you know what? Here's another difference between us and AI, like we'll always want to figure it out.
Cherlynn: They'll stop, yeah, you're right.
Terrence: And now for an awesome transition, the other thing I do want to figure out is I want to figure out what you guys want. Please tell us, send us feedback, leave comments, hit us up on Twitter @Engadget. Email us at podcast@Engadget, or hit us up individually on Twitter. Chris, where can the find viewers and listeners find you?
Chris: Twitter.com/chrisvelazco, V as in Victor, E-L-A-Z as in zebra, C-O.
Dana: Oh, on Twitter?
Terrence: Yeah, where can our fine readers and viewers and--
Dana: It is my full name, Dana Wollman
Chris: Dana Wolfman.
Dana: Dana Wollman, stay out of my mentions unless you are nice. Thank you.
Terrence: Cherlynn, where can they reach you?
Cherlynn: @CherlynnLow, or just go to one of their Twitters and find me.
Terrence:(chuckles) Just go to one of theirs. Look in the people they follow.
Chris: I'm always fighting with her, so you can just find her via me.
Cherlynn: Yeah, we're always fighting.
Terrence: And you can find me on Twitter @TerrenceObrien, that's lots of E's, no A's. Please join us next week. Thank you for watching. But before we go, I want to leave you with the comment of the week, which comes from EClinton about price or death to this article?
Cherlynn: Death to you.