Get in touch

Postlight joins Launch by NTT DATA! Learn more.

We finish our conversation with Dean Hachamovitch, and while we start out making fun of Clippy, we end up discussing conversational interfaces, security and privacy, and the responsibilities of software development. As Rich puts it, “I just want to congratulate everyone here for smoothly weaving Clippy into some NPR-ish conversation.”

If you like people talking about big software problems and how to solve them in giant matrixed organizations, you’ll enjoy this.


Paul Ford: Hi, I’m Paul Ford and this is Track Changes from Postlight, the little web and app agency that could, located in New York City. We build your websites and your apps and all the big platforms and crazy stuff that it takes to get things happening on your little phone or your browser. I’m sort of going down a hole here. I’m here at the studio with Richard Ziade.

Rich Ziade: Yo, Paul Ford.

Paul: He’s my co-founder and we’re joined again — this is really cool — by Dean Hachamovitch.

Rich: Yes.

Paul: Dean, hi. It’s good to see you.

Dean: It’s good to see you. That was almost a Steve Ballmer impersonation on your part.

Paul: What is it sound like when Steve Ballmer says your name?

Dean: It’s really loud.

Paul: You probably would break the microphone if you did it.

Dean: Absolutely.

Rich: Coming soon, Steve Ballmer ringtones brought to you by…

Paul: Dean Hachamovitch is formerly Microsoft, now out in the world exploring technology on his own terms. So it’s good to have Dean back. We’ve got a lot to talk about.

Rich: So there was this feature that was added to Microsoft Word.

Paul: I think Word 97.

Rich: Is it 97? I remember, Word was a … I think this overlaps with your time there. In my view, it was a work of art at one point. I forgot which incremental version it was and all of a sudden, my computer felt faster. You probably know which it was. It just felt like I could scroll to the bottom of a document real fast all of a sudden. It’s like, “Whoa. This is cool. I didn’t have to go to my same …” I love the opening my, whatever it was, 386 back then and putting new chips here and there. And then Intel lost it’s mind. It was a big bar, you sort of plugged in and it wasn’t, t was just this whole —

Paul: This is when you say, “Get off my lawn.” Just letting you —

Dean: Yeah. Exactly. I’m thinking to myself, “This thing is awesome” and then …

Paul: Word has good years and bad years. You had a good year with Word.

Dean: I mean I’ve pretty much been with Word.

Paul: Yeah.

Rich: …forever?

Paul: It’s one of the longest-term relationships of your life, right?

Dean: It’s probably one of the longest … We still talk.

Paul: I’ve been with Microsoft Word for 24 years probably at this point.

Dean: I mean it was … I was a Word Star guy before Word. I just skipped WordPerfect, because WordPerfect tried to go GUI and they just lost their minds.

Paul: That was bad.

Dean: It took two and a half hours to install that thing. It was just like, all right this isn’t going to work.

Paul: 5.2 was the last great WordPerfect.

Dean: Yeah.

Paul: Then it kind of went downhill.

Dean: I’m always excited for the new version of Word even though I don’t need anything anymore. I just still was excited about the new version. And this little paper clip shows up and says something along the lines of, “It looks like you’re trying to write a letter to your mother.”

Paul: Just a letter, it didn’t know mother.

Dean: It didn’t know mother?

Paul: No, it just knew letter.

Rich: It was just your version. That was just your version that we targeted.

Dean: You sent me the Mother Version of Word, you saw some issues there. And I was like, well this is interesting. First off, it was animation, which in of itself, I’m in a word processor, usually life is not that exciting.

Paul: I mean most people don’t really want that many animated characters in their word processor.

Rich: No, but it was still novel and weird to me.

Dean: It was trapped in a window.

Rich: He couldn’t really walk around the screen. There was that. But I was just like, I’m good right now, dude. I’ll talk to you in a bit. So I would send him off.

Dean: Close box.

Rich: Close box, was it … close box. Then fifteen minutes later he comes back. He says, “Are you trying to print, dude?!” I’m thinking to myself, “Whoa.” You ever go to a cocktail party and you sort of position yourself to not be noticed by that guy or that girl so they don’t come and talk to you again?

Paul: For me, it was just like, “Wow I have a roommate.” [laughter] I thought I was … It was a personal computer. It was just me and this computer. Then there was, somebody else was living in there.

Dean: So we’re being cryptic here because this is fun and nostalgic but the feature was called Clippy, or the character I guess was called Clippy, and they took a paperclip and they used the bottom loop and turned it into a smile and they put two eyes on it and it would hop around and do its thing. They had different characters as well. There was Einstein if I recall correctly.

Paul: There was a puppy and a wizard.

Dean: It was the Assistant.

Paul: The Office Assistant?

Dean: The Office Assistant. It may have resembled Einstein from some people’s point of view. I think the estate might have had some issues. I think it was actually called “Genius.”

Rich: You never say the word Einstein.

Dean: I also usually refer to Sticky Notes as compared to the copyrighted name. There’s some residual scar tissues from some lawyers.

Rich: Did you give birth to it? Did it come to life while you were handling Word?

Dean: I was working on Word 97 and the folks working on the core Office team on all sorts of user interface parts were working on the Assistant.

Rich: OK.

Dean: Or the friendly character TFC, which was the code in the actual code for the product. It was referred to as TFC. TFC Cur for the current one, and you know.

Paul: I could see that standing for other things too.

Dean: There were all sorts of things some people might speculate that TFC would stand for. Now it came out of the office team. It showed up in Word and Excel and Powerpoint and everywhere else. The intentions were so good, and if only there was some way I could not sound defensive in describing this, if only there was a way I could not sound like I was trying to make my brother-in-law look smarter than he is. [laughter]

Paul: Could I describe how I see it?

Dean: Go for it.

Paul: What I see is that Microsoft Word and Office in general was very very complex at that point, probably one of the most complex pieces of consumer software that ever existed —

Rich: — and getting more complex quickly.

Paul: Right. There’s been, we were just talking about AI, there’s always been a strain of thinking in software where people go, interfaces need to become more conversational and more friendly. We’re actually seeing that now. There’s a huge resurgence in bots and conversational interfaces, driven by things like Slack, that’s happening in the world today.

So I’m assuming, the way that I interpret Clippy historically, not as a user but as someone who’s interested in software, is that Word had gotten, and Office had gotten so hard to use, or so big, not hard to use, but just big —

Rich: I think that’s the word.

Paul: — there was just so much stuff going on, and we were worried, as the software industry was worried, that regular users couldn’t make sense of this world.

Dean: So a great example of this is when we would look at feature requests. People would write physical letters or send email or call up. The fraction of requests for features that were already in the product, when those started to dominate, we thought “Aha! We might have a discoverability problem.”

Paul: OK.

Dean: There’s a question of, OK, what do you do? Do you rearrange the menus, we had toolbars, do we do something beyond toolbars? Are there floating palettes? There are all these interesting ways of trying to present the space of commands.

Paul: What’s interesting is I think the modern conventional wisdom is why didn’t you simplify, but you actually are getting these feature requests.

Dean: Right, and even the point of simplification, your crucial features is my bloat.

Paul: Right.

Dean: There’s this other interesting thing playing out called, “I just need a word processor for simple documents. I don’t need something with all that fancy lawyerly stuff that does table of authorities and blah blah blah.” Well, until you get a document from your lawyer and you want to open it, and then you do want that. Or, I just need a simple one but then you’re kid is writing a paper with footnotes.

Paul: Well see you lose interchange if you simplify.

Dean: Right, and so if you start forking and fragmenting, things get weird. We’re like, we’re not going to go down that path.

Rich: Just a quick question. Did they ever think about like Word for Attorneys?

Dean: You know, that’s like an amazing April Fool’s episode waiting to happen, like what would you put into Word for Attorneys.

Rich: I saw a paralegal once use Word and it was a whole other universe. She was incredibly versed in it, in this whole suite of tools that were just very specific to legal.

Dean: Absolutely.

Rich: It was just stuff I’d never seen before.

Dean: As compared to the suite of tools for people who were doing form design and building applications on top of Word and Excel.

Rich: Right, right.

Dean: Just to try to accelerate through this, there was this notion of well, we have all this help material, what’s wrong with help? There was a really cool breakthrough using some early statistical work and language understanding work, and it was called answer reserve. The point was, how can you get, if people were willing to type in a question in English, “How do I blank,” we could get a better answer back. There’s a challenge around Office 95 around how to get people to type more in?

The fun example of this was watching this user trying to figure out why the paragraph marks, those backward Ps, were all over the screen. The user had accidentally hit some button, and view hidden characters, view formatting. This user typed into the help box, “How do I get rid of the elephants?”

Paul: Sure.

Dean: Because that paragraph mark glyph looked like an elephant.

Paul: Got a little trunk hanging off —

Dean: Yeah! Absolutely.

Rich: That’s amazing.

Dean: And we’re sitting here going like “OK. We really need to re-think how we approach this.” There was all this really cool work done with how do people talk about what they see on the screen. They actually had technology where if you would write in, how do I get rid of the elephants, or how do I print sideways, these great examples where it actually worked.

Paul: Because nobody would know portrait or landscape, they’d know sideways.

Dean: Right, what do these words mean? Me want print sideways. It would come back do you want a whole page to be sideways or just this part of a table to be sideways? There was all this great technology, but the challenge was how do you get people to type this in?

And in Office 95 there was an Answer Wizard tool bar that was grey and rectangular and did its best to encourage people to type into it. That didn’t really go very far. There was this whole bunch of interesting research at a Stanford, I think the researchers were called Nass and Reeves, around social interface, and a couple of smart people ran with it and we got the Office Assistant and I think the default was Clippy but there was also Rover, the yellow dog, and Genius. In Japan there was Office Lady, who would bow a lot, and all sorts of interesting characters all over the world. Merlin, I think, there was a very wizard-like creature.

Rich: Yup. There was a wizard.

Paul: One thing too for listeners to this podcast may not realize what a big deal help was —

Dean: — pre-internet days.

Paul: Pre-internet. You couldn’t Google. There was nothing. It was the manuals and the apps had gotten so big that you actually couldn’t publish all the help that was being generated. I can still remember, you know, .chm files. There was a whole ecosystem, a whole world, around creating —

Dean: It was even pre .chm files, so .hlp files back in the day.

Paul: This is like, there were tools to create these, tools that were kind of proto-hypertext-y —

Dean: Rich, pre-html. So I want rich text with a link that I can click that will jump and a back button, but it wasn’t a webpage.

Paul: I mean, millions of hours were spent reading Microsoft help files over the many years. I mean it was a real format onto itself. There was this huge base of experience and expectations that you were coming into and you were trying to help and save people time exploring and searching through those documents.

Dean: I should be clear that when you say “you” here, you don’t mean Dean.

Paul: No. Microsoft.

Dean: You mean the Microsoft collective. There were some very amazing people who decided to go with Clippy and work this out.

Rich: Now I’m feeling bad because it’s fun to to make fun of Clippy. You think, “Oh what’s the big deal? Some people decided to throw up an animated paper clip.” But this is a massive amount of research and work went into this thing.

Paul: This is always, we were talking about this earlier, you’re always in a position when you do this work, feedback never comes in gently, and it always starts with good intentions.

Rich: You know, there’s a, I think it was, I may be misquoting, but I think it was Don Norman —

Dean: Great UI person.

Rich: Great UI person. He said, “Nobody thought that at the yellow light humans would speed up.” [laughter] It’s supposed to slow you, you know, hey, settle in, slow down, and everybody speeds up. I think, you just don’t know what you’re going to get. This is what I’m really curious about. All this beautiful work happens, Clippy goes out, and what happens?

Dean: Well, it’s like a whole lot of other pieces of technology. Humans interact with it and then make their own set of decisions. There’s another great example: in browsers, there’s in private mode, or private, or incognito, there are a variety of names for this. When we first came out with this, the press person who broke the news called it “porn mode.”

Paul: [audible shudder]

Dean: Which is [laughter] essentially pre-positioning and it’s, I guess it is a use case —

Rich: This is The Village Voice reviewing a browser, ladies and gentlemen.

Dean: No, I’m just saying, whatever the functionality is —

Paul: Right.

Dean: Right? There are different ways of, “Hey is this empowerment and good or is this an attempt to get back at us?” DRM is another great one. DRM is —

Paul: Digital rights management.

Dean: — is an amazingly politicized, and crypto now, Crypto’s fantastic. The entire privacy back and forth and I’ll give you the cheat sheet on privacy. If you include features that help protect privacy, then you’re accused of enabling very creepy, unpleasant people to hide from law enforcement.

Paul: Sure.

Dean: And if you remove them and make finding people easier, than you’re rapidly accused of making vulnerable groups even more vulnerable, whether it’s political dissidents in foreign countries who end up using US technology, or families that are fleeing from an abusive other family member. It’s just one of these, there actually is no win situation here. There is just a piece of technology and how different people are going to try to use it.

Paul: So do you, if you’re working on IE and you’re getting that feedback, are you taking that case by case? Is there some moral framework that you fall back on? What do you do?

Dean: Fundamentally, I think you have to have a set of principles. The only way anything ever scales is if you have principles. Of course you get to have exceptions and guidelines and apply thought to these principles.

Paul: Sure. That’s why we can amend the constitution.

Dean: Ideally. We can hope on that one. Off the cuff, when I looked at incoming issues, first and foremost there was respect the laws of the lands in which Microsoft does business. When a government or legal issue came in that had a certain level of urgency at the top, you could argue the airway breathing circulation scheme that was OK we want to respect laws. That’s very important to us.

Right behind that: is this a threat to customers and partners? Is there an active security exploit, does this put their data at risk? Then you slowly work down all the other color codings, so that you can be rational in allocating the scarce human engineering time towards any of these incoming issues. You just kept that framework up and it helped you stay sane.

One thing that was interesting is, a long time ago, IE was relatively unique in terms of how many people are using it, how are they using it. It’s kind of wild now how commonplace that is. I’m sure there are a lot of people now who are like, “oh yeah, 100,000,000 people, sure,” how many people go to

Paul: That was an unprecedented number at the time.

Rich: Of anything.

Paul: That’s the thing, Windows was so big, and so global.

Dean: It was unique in it’s reach for a window of time. What we get to get used to now is how many things we rely on that deeply. One thing that I find from somebody who works in technology and still enjoys technology is how much we lose track of the underlying layers. I flew out from Seattle yesterday, we had a power outage because of high winds. I’m going, “Wow, it’s 2016, and I lost electricity? Really?”

Paul: Because of wind. Because of a zephyr.

Dean: Right, we should be on top of this one. You suddenly realize, or look at what happened in Detroit with water, and it’s one of these, wow, it’s amazing. I can be on an airplane, using WiFi, transacting — this is wonderful, but it’s very inconsistently distributed. That’s a recurring thing. It’s very easy to lose sight of, now, how many people use Uber at any moment in time or how many people rely, you know, we were talking about US digital services and a lot of the great stuff coming out of the White House, like how essential that is for emergency responders.

Paul: You’re alluding to, there’s a great William Gibson quote, which is that the future is here. It’s just unevenly distributed.

Dean: Absolutely.

Paul: We’re living with that. That’s good and that’s valuable and that’s very high-minded, but —

Rich: I just want to compliment everyone here for smoothly weaving Clippy into some far-reaching, almost NPR-ish conversations here.

Paul: We need to find out how it ended for Clippy.

Rich: It’s just…

Dean: Clippy had multiple deaths. Stage 1 was in Office 2000. Office 2000, two changes happened. One: Clippy got out of the box. We got rid of the containing box with the title bar and a closed box. He was free. He had sorts of interesting transparencies. He could live more natively on top of your spreadsheet or —

Rich: Could you drag him around? I think you could move him around.

Dean: Yes. In fact, I think the most popular command is when you could right click and choose animate and just, when you were on hold on a phone call, you just sit there and animate and watch him cycle through all the different dances he could do.

Paul: They had all these different animations. Yeah I remember.

Dean: It’s very important, like there was one where paper was coming out of his head or something. [laughter] He lost the containing box and he was turned off by default. Then I think I left the Office team after that release to go do something else. Over subsequent versions, I think he may have come back from the dead, and then there was an ad campaign where they were putting Clippy out of a job, and they just kept milking Clippy. Clippy was a very polarizing historical figure.

Paul: What actually happened when, under your watch, was that Clippy went from default and on to opt-in and —

Dean: Off by default.

Paul: Off by default.

Rich: Which is probably the end of Clippy.

Dean: There were some people who really liked Clippy. Letters —

Rich: Just a rainy day, kind of lonely.

Dean: Well, Paul didn’t want a roommate but clearly some people…

Paul: Did a significant percentage of users do you think actually used Clippy for Clippy’s intended purpose, to learn and explore Microsoft Office and become more empowered users?

Rich: It turns out that the users who did engage with Clippy ended up having a better experience because it turns out that to have fun and do the generic, it’s a three-tier architecture, look, the important stuff was under the hood, there was a much better set of explanatory text if you were trying to show the elephants or hide the elephants, to go back to the paragraph marks, or print sideways, or whatever it was that you were trying to do. There was better text there.

How could we get you to actually ask a question using more words in English so we could understand what you were trying to do? Well, then the tip of that iceberg was there was an animated paper clip with big eyes and pumping his eyebrows. Perhaps not the greatest moment in technology history. Where we got from there, if you think about it, as you said, when you look at Slackbot, if you set the pixels aside, this notion of, ‘type what you want or you could argue a marginally more enlightened command line interface. Back in the day it was you need to learn how to type exactly the sequence of characters for things to work.

Paul: Sure.

Dean: We got more forgiving in what you could type. Then we tried to dress up where you were typing to encourage you to be a little more liberal in what you would type.

Paul: Got it.

Dean: I’m trying to tell you to think about Clippy as a proto-AI entity.

Rich: I mean it’s a bot.

Paul: It’s a bot. We’re coming out a world where, in very recent memory at that point, your interface guide was a keyboard overlay. You’d actually put a piece of plastic on your keyboard and it would tell you which commands you could use to do things. WordPerfect was famous for that.

Dean: Shift-F7–6. Print preview.

Paul: Exactly. Some of them, alt-F3 is still in my brain for I think reveal codes on WordPerfect. That was really only about three years away, you know, four years away for most people by the time you’re animating this little guy.

Dean: There’s this interesting compression of history. People are coming from very different points of view. Let’s make this modern for a moment. There are a whole bunch of people who have really never used a mail client, an email client, on anything other than a mobile device.

Paul: Sure.

Dean: On an iPad, or an Android tablet, or on their phone. There is a mail client, there is a mail program. On their Mac or PC, they just use the browser.

Paul: Sure.

Dean: So generationally, there are a whole bunch of people who have no idea that it could work any other way.

Paul: That whole category of application is gone.

Dean: Just skipped. Well, there are a lot of people who still live by Outlook. What I’m saying is there are a whole bunch of people who just skipped that.

Paul: Gmail has a billion users now. I mean there’s a, for tens of millions of people, there’s no email application on their computer that they care about or know about.

Rich: There’s a wave now that doesn’t even know what a desktop is. I mean, their email and their messaging is all happening mobile or tablet, probably mobile, and that’s that. There’s no dock, there’s no tray, there’s no downloading of apps. I mean there’s apps, but there’s no notion of file systems. A file system is alien to a thirteen-year-old.

Paul: Well an app is also now a pop-culture artifact that costs a buck.

Rich: Right.

Paul: That’s a huge change from $179 in a box, in fifteen years we’ve gone from —

Dean: Well you know sometimes $350, but the competitive upgrade is only $99.

Paul: Right.

Dean: Distribution is still a shrink-wrapped piece of cardboard.

Paul: And now if you can a $1.79 you’re killing it.

Rich: That’s rough.

Dean: Well the question is $1.79 up front? Or is it free and you have some in-app purchases or the app is free but there’s a subscription, like Netflix. Netflix app is free!

Rich: There’s still, I mean I think it’s $3.99 is the tiny scan app that will take a picture, you take a picture of a piece of paper, does a good job flattening it into two-tone or whatever, and then it’ll turn it into a PDF. Then send it off to 10 different storage services or email it and attach the PDF. Real work. Somebody worked pretty hard on this for a number of months. $3.99.

Dean: I have no idea how many people will buy it. This goes, you know, Paul, you were talking to me earlier about this whole commodification, right? You used to be able to differentiate by writing really good print drivers. Nope, everyone expects printing to just work as free, there’s not a way to differentiate. Or typing. Having a really good place to type, used to be a great way to differentiate your product. Ah, you know I think it’s pretty commodified.

Paul: Take it for granted.

Rich: I have a closing question.

Paul: OK.

Rich: No, but go Paul.

Paul: No I —

Rich: No, this is your show, dude.

Paul: This is your show, too. We’re co-founders. What’s your closing question?

Rich: Well, you have a question, so let me not make it, this is a pre-closing question because Paul gets to close the show, everyone.

Paul: No, no, no.

Rich: Paul capital F-o-r-d.

Dean: I can leave the room.

Paul: My! God!

Rich: Stay, stay with us. So you obviously a unique perspective. You got to sort of be in the middle of the hurricane in some of the most important years. What’s exciting today? What do you get excited about today? You can say nothing. Arrogantly just say, no, this is all bullshit.

Dean: No, this is, it’s so much more exciting now than it was before. On so many different levels. Again, you just view this as compound interest on technology. You look at the new mobile devices and what they’re capable of. You look at the sensors that are coming online. Look, every time I go online and I log in to Amazon Web Services. It’s just like…wow.

Rich: AWS is amazing, yeah.

Dean: Right, and it’s just, people have bludgeoned the following truth so many times at the beginning of their talks at Google IO or anyone else. It’s the “back in the old days, you would have to call up Dell and buy a lot of computers and find electricity and cooling and hope that you hadn’t bought too many or too few.” Now basically computation is vast and free and you can rent it by the minute. Which is amazing. Storage has effectively —

Rich: — like electricity —

Dean: More stable, if you live in the Pacific Northwest, it turns out. Cost of storage has depleted fantastically. I can’t remember if it was The Onion, there was this headline around computer science being renamed “Googling Stack Exchange.”

A lot of this information if you go back and read some of the original amazing books, like The Soul the New Machine by Tracy Kidder, that whole notion of the high priesthood that protected computers from the rest of us, or the famous 1984 Apple ad. That whole holy priesthood that was keeping people away, that’s all disappeared because of what we’ve gotten to on the Web.

I’m giddy at the stuff that can be written for fun just to understand what our modern technology, and how does programming work now. A couple of my kids are classics students, so they study Latin and Greek. For fun I wrote them a website where they could type in something in Latin and it would do all the appropriate look ups and bring them back something useful.

It was fascinating to me that I could build this at zero cost using just what came for free with my Mac and a whole bunch of web searches and the APIs are out there, some group has this interesting API/CARPC on top of this corpus, and I could just build that and it was useful for them and it worked on one kid’s Android phone and another kid’s iPhone and it worked on my desktop.

Rich: Dad of the month right there.

Dean: It’s, the thing is that it is, it’s just nuts to think it’s not exciting.

Rich: It is, that those building blocks are just out there.

Dean: Yeah and they’re just out there. You know, we talk about the exciting apps like, we’re at this stage where we just get to apply software learning towards making other things better.

I want to be clear, it’s not infinitely better in all directions at all times. You gotta talk about Uber. You’ve got to talk about AirBNB. The general problem in economics is there is this amazingly expensive capital good. How do I get more value from it over time? Well my car can sit in a parking lot or I can drive it around and earn money taking people. Wow, that’s unbelievable liberating. Was that a software problem? No, software’s a heck of a good tool.

Rich: Yeah.

Dean: Anyhow, so it really is beyond exciting. My kids at this point just know how to roll their eyes with one of these, “OK you’re going to tell me how to used to be.” Really? No, really.

Rich: Well it sort of establishes an appreciation I think. I think it’s the ability to ship, right? You can piece something together, put it out, let the world tinker with it, without the distribution chain.

Dean: The distribution chain and, for the sake of the pessimists, there’s a lot of other stuff that’s come with all of this good. I mean, we still have spam problems or junk mail problems. We have identity theft, we have amazing —

Rich: Security issues.

Dean: Amazing security issues, amazing privacy issues.

Paul: So let’s close it out with a little bit of pessimism. [laughter] You built a web browser.

Dean: I was responsible for keeping a whole bunch of really smart people who built a great web browser happy and defended from internal randomization.

Rich: He was on email and PowerPoint mostly.

Paul: Excellent diplomatic clarification. We built a lot of websites and apps, and steadily as we talk to people, we’re asked, “Is there a future for the web?” We’ve got not just that maybe mobile will take over, I don’t think it’s just that, but also that the web is increasingly locked into a handful of platforms. Facebook is something that now people experience as “the internet.” Many people don’t know that there’s a web beyond Facebook, as far as I can tell. You know, things like Twitter, LinkedIn, whatever, these very very large networks that are really starting to lock people in.

Dean: Dave Winer calls them silos. He writes very eloquently around beware of these silos. We’ve seen this before.

Paul: We have, but where does this go? Are we going to — is it, maybe it’s cyclical, is there an endgame for the web where there’s only a few silos? What happens?

Dean: I’m going to make some popcorn —

Paul: Yeah!

Dean: And watch the show. Look, we, again, back when giants walked the earth, it was fun to complain about AOL. There are all these people trapped in AOL, they don’t know what else there is. And for a while that was Yahoo. There was a window of time when Yahoo was another great walled garden. I think that we have no idea where we’re going in terms of, what is the web and what is the apps?

To geek out for a moment, what you really want to do is write a service that runs happily in the cloud. What happens then is people are going to interact with that on a variety of devices. They can interact with it on their phone, on their pad, in a browser, from their Raspberry Pi, from their Amazon Echo by saying, “Alexa, go do something.” You’ve seen these Amazon buttons, right, that you just hit the button and detergent comes to your door. That’s just a great example, there’s a web service doing the right thing.

So I don’t really see an endgame or a demise of the web. I see a whole lot of evolution.

Rich: Yeah. You know what I feel like? I feel like the web, where there used to be a lot of residential and commercial, will become more of an industrial zone, where it’s really the piping to a lot of this stuff, but the consumption in the windows are going to be these other places because Facebook is just such a —

Dean: Think about it as fixtures. There are fixtures a top of really, like the plumbing’s going to get a lot better. The fixtures are going to get different.

Rich: Right. Exactly, exactly. And the gravitational pull of Facebook is what it is. It’s just our reality today. If you really, if you’ve got that idea, it’s itching you, and you want to do something interesting and disruptive, the tools are still out there. You can’t knock on Facebook’s door and say I’ve got a great idea and I’ve got a tool belt on, you mind if I just come in and just build this in your world? No, it doesn’t work like that. The builders and the people who want to mess around still go to the web. It may not be “the web consumer,” the browser web, but the piping and the infrastructure, all the pieces, are still out there for anyone to play with. I’m hopeful. [laughter]

Dean: I think there’s still a ton of upside.

Rich: I think so too. I don’t know how this ends. I don’t know if it does end. I’m trying to be optimistic.

Paul: I don’t think it does end, and hopefully we get to bring Dean back on in the future and ask you about how it all turns out. It’s been great to have you in here.

Dean: Thank you.

Paul: Come back and visit us.

Dean: I’ve really enjoyed it.

Rich: Thanks, Dean.

Paul: Postlight has a newsletter! You can subscribe. Just go to and you’ll see the subscribe link right there. Every day we send out something interesting and engaging to people who work in, around, or with technology. Lots of links, lots of stuff. It’s very nerdy but sometimes it’s not. And you can go straight to and see all the newsletters that have ever been.

Thank you everybody for listening. It was great to have Dean Hachamovitch here in the studio for Track Changes from Postlight. If you like this show, it makes a huge difference if you go to iTunes and say so. Give us a good rating and feel free to send us an email to if you have any questions at all. We’ll see you next week.