While every app rushes to add an AI assistant, iA Writer’s Oliver Reichenstein asks, “Why should I bother reading what you haven’t written?”
Here he talks tells about the deep thinking behind the latest version 7 of IA Writer and the new feature that dims the text you paste from AI tools, keeping track of what is yours and what isn’t.
Watch on YouTube:
Or listen below:
Show Links
Oliver
iA - https://ia.net
iA Writer - https://ia.net/writer
iA Presenter - https://ia.net/presenter
Writing with AI - https://ia.net/topics/writing-with-ai
Oliver on Mastodon - @reichenstein@mastodon.net
iA on Mastodon - @ia@mastodon.net
Andy
Website: https://www.polaine.com
Newsletter: https://pln.me/nws
Podcast: https://pln.me/p10
Courses: https://courses.polaine.com
LinkedIn: https://www.linkedin.com/in/apolaine/
Mastodon: https://pkm.social/@apolaine
YouTube: https://www.youtube.com/@apolaine
Transcript
Note: This transcript is machine-generated and may contain some errors.
Andy Polaine (00:00:03): Welcome to Power of Ten, a show about design operating at many levels of Zoom from thoughtful detail through to transformation in organization, society, and the world. My name is Andy Polaine. I’m a design leadership coach, service design, innovation, consultant, educator and writer. My apologies for the late start. We had both some tech problems and also the city have decided to cut down a tree just outside my window, so I’ll be hitting mute judiciously throughout the interview to try and hope that you’re not going to hear that there. It’s so while every app is rushing to add an AI assistant and add ai, this and everything will help you write on LinkedIn. You can go online and start to write a post and AI is like, would you like to meet Ed? Write your thought leadership for you. My guest today, Oliver Reichenstein says, why should I bother reading what you haven’t written? Oliver is the founder of information Architects at the company behind the very popular markdown based writing app, IA writer, and more recently, IA presenter is here to talk about keeping it real when writing with AI and why and how they built in this authorship mode into IA writer Oliver, welcome to Power of Ten.
Oliver Reichenstein (00:01:20): Great to be here. Thank you for the invitation.
Andy Polaine (00:01:22): So it’s been an age since we actually kind of chatted person to person. I can’t remember. We met many years ago at a conference and hung out. Our two families hung out together and obviously I’ve seen IA writer evolve over the years. One of the first apps that I used on the mobile space actually I think is probably where I first started using it as the iPhone came out and plain text was a thing and markdown was a thing before you could kind of really do anything more interesting. So first of all, before we get into the latest bit, how has it evolved over the years for you and how many years has it been?
Oliver Reichenstein (00:02:02): It’s been 13 years now. The idea is very, very old. I used to earn my money as a teacher of Microsoft products, Microsoft Office, and in particular Microsoft Word where my job was to teach people how to use Word in particular to write scientific papers for university or for high level schools. And I used to teach at Art School Basel, and I noticed there that the art school pupils had a similar problem as I had where instead of writing, they were just choosing fonts and colors and line heights. And by the time they were done with the right layout that the lesson was over. And of course that was bothering me as a teacher, but I noticed that I had a similar pattern every time I open Word, I first have to make sure that everything is correct before I get started. And so at the time I started using a typewriter myself to see how that compares to work and I found that I write much better with the typewriter even though it’s very bothersome when you make a mistake and all that, but I found the way it slows you down, the fact that it hurts when you type that you think twice before just typing anything you can correct it later actually led to a much better writing just simply because I was thinking before moving my fingers.
(00:03:33): And so that idea was born in the nineties, somewhere mid nineties, but I always carried it with me and by the end of the zero years or the naughties maybe, I think they’re called in English, not quite sure
Andy Polaine (00:03:50): They do we say the naughties?
Oliver Reichenstein (00:03:51): Yeah, it sounds nasty, but
Andy Polaine (00:03:53): Well, it’s the English dub.
Oliver Reichenstein (00:03:57): Yeah. By the end of around 6 0 7, we started to make concepts of an electronic device that we wanted to build in China by ourselves. It was half serious to serious, but I started to work with a Finn designer that used to work for nasa, did toilets, they dump toilets for spaceships because Americans didn’t like making toilets. It’s dirty. So they got the finish guy that had no experience with anything space, but yeah, he’s a funny guy. He lived in Japan and we worked together on a concept for an electronic device, an actual device where he could only write, and then the iPad came out and things went happened very quickly. We were lucky to get one of those iPads early on because we were working for and decide, and the moment where Steve Jobs presented the iPad, we were on the job, so we were prepared with the design conditions of the iPad. At first, we printed out paper iPads and moved little boxes and pieces of paper on top of it. And then I think in May we got the first iPad at the office. We had a little, hes start because I think it came out, I don’t remember exactly when it came out, but we were really, the timing was perfect for our app. And yeah, we had a couple of good ideas from the very start. I think we came out with focus mode, the idea of focusing on one sentence at the time. Like a
Andy Polaine (00:05:42): Typewriter, right?
Oliver Reichenstein (00:05:43): Like a typewriter, exactly. It was very really a translation of the typewriter, what we’ve done. But we also said it’s silly to just force people to write on a typewriter where you can’t edit because that wouldn’t be medium appropriate. It’s a basic design principle that we follow and everything we look at what we do now mean in the physical world and how does it translate, how can we improve? It? Had reading time and we spent a lot of time on the typography of it as we always do, and we started thinking about something called reading typography where we discerned the writing from the reading typography. I think it’s a new thing on computers because before we didn’t have much choice on the typewriter, what organic typography use, and we found that actually Monospace had a lot of benefits. I’m not going to get into each and every point there, but that’s where we started.
(00:06:41): And the vision was always the same to create a writing app that is made for writing because word is not like that word is the default writing app, but it’s more like of a layout thing, but you can do this and that and versioning and all kinds of stuff. It has a visual basic editor in there. I know each and every corner of the nineties word application because I taught that and I know each and every bug from the nineties, the latest versions may be fantastic, but from what I hear, there’s still the same box around. You can’t get out of lists, especially if you have a nested list. Once you’re on a nested list, you can get out. All these things still
Andy Polaine (00:07:22): Present. It’s hugely buggy. I mean, I rant about Microsoft on a regular basis. I mean mostly about their irritating account authentication. But yeah, no, I mean having been working in a consultant, well, there’s two use cases that still just irritate me enormously. One is having worked, having written a book, there comes a point where because copy editors want track changes, which we’re going to get to, then you’re stuck. Then however you’ve been writing, and I write in Mark down and I have done for many, many years as pretty much as soon as the spec came out, then you have to go into Word or you can kind of do it in Google too. But it’s horrible. It’s horrific, and I hate it. When I wrote my PhD, I avoid using, I used another app called Elle back then. I couldn’t face doing a long form. It’s really bad for long form writing where it just starts chugging really well. It feels like I always feel like it. I’m using an emulator, which I kind of suspect I am. And the other thing is then in consulting obviously is the world of PowerPoint and talk about the number one place where you can faff about and lose enormous amounts of time on formatting. That’s the place. Now you guys, before we get back into IA writer, we probably just talk about IA presenter a little bit. So tell us about IA presenter.
Oliver Reichenstein (00:08:53): So yeah, I haven’t worked at PricewaterhouseCoopers or wherever you’ve been,
Andy Polaine (00:09:00): Well, it was Fjord, right? Fjord
Oliver Reichenstein (00:09:02): Right of used to be a design agency, but Accenture,
Andy Polaine (00:09:06): That’s such a burn, but yeah, okay. Yes.
Oliver Reichenstein (00:09:07): Yeah, sorry. I don’t know about the insider jokes in those
Andy Polaine (00:09:14): Consultancies,
Oliver Reichenstein (00:09:16): But yeah, as a design agency also you live in PowerPoint. I used to work, which is not that far away from the really bad evil consultancies, but
Andy Polaine (00:09:31): Purely FA nice guy. But yes,
Oliver Reichenstein (00:09:33): Very nice. No, they’re very nice. Yeah, they just belong to McKinsey or what was that
Andy Polaine (00:09:39): Again? They belong to Accenture, which I probably would. Yeah, I have to stop. Not the worst, I would say. I think McKinsey probably pretty high up there. So anyway, so I
Oliver Reichenstein (00:09:52): Have to start with the ego jokes, but yeah, so what is it about You live also in PowerPoint and especially if you work for a bigger consultancy, you just make one PowerPoint after the other. And that’s pretty much what I did at the, I just lived in PowerPoint, felt like very basic camping actually. And then there was a phase where every web designer had to be at the design conference and I followed the try. And so I was giving one presentation after the other until there was a point where I asked myself, do I really need to use PowerPoint here as well? Do I really need to have a presentation? And I started experimenting with presentations without PowerPoint and I liked that a lot. And I remembered on weddings or Christmases, when your talented uncle stands up and hits the glass and then starts talking, people paid a lot of attention.
(00:10:55): Everybody was happy and looking forward to these. And most of these conferences, I’m sorry, but most people don’t really enjoy these presentations. They don’t. They’re just boring and they have the memes and they have the videos and so on, but there are very few people that really care about what they say. And yeah, I kind of ended my whole career at conferences because family was starting to complain. You fly to Australia, you fly here, you fly there. Well actually you have two kids at home starting to say no and staying with the family. But I ended that career more or less by sticking to non PowerPoint presentations and just talking. Even though I was talking about design, I always have this tendency to philosophize about design, which is the reason why I accepted these conferences because that gave me the opportunity to talk about Hayek and Plato and Aristotle, which clients at the time were not.
(00:12:02): So, but yeah, that was pretty much it. And I’ve been mostly working on the application for the last 10 years except for some select clients that I still work with and I didn’t need much PowerPoint anymore. And then one of our employees started working on a markdown version of PowerPoint quite some time ago. And developer very, very talented, extremely talented, extremely efficient, interesting personality. And I was very skeptical. I was like, first of all, I don’t like presentations and then just making PowerPoint in markdowns kind of pointless and I hate that. I hate presentation apps viciously. And then his presentation app wasn’t big commercial success. And he said, yeah, but I built it. I have all this stuff and maybe I think what is needed is some design. And he felt like making it look nice. And I don’t take that personally.
Andy Polaine (00:13:18): Oh, could you add, can you put some nice design sprinkle on top of my code? Is that you’re
Oliver Reichenstein (00:13:24): The bit, like I said, he’s a really cool guy. He totally understands what we are actually doing, but I think he might have, no, I’m being unfair. He knows what we’re doing and he knows that it would have probably gone a bit further, but we planned three months to see what can happen if we add some design. And then these three months turn into three years because design is not on top, it starts on the bottom. So we started asking what is really a presentation? How should it be done? We started researching, we went to ro, I don’t know how to say in English, that’s just something I always read in English. Cicero, I
Andy Polaine (00:14:07): Think Cicero, I say Cicero,
Oliver Reichenstein (00:14:09): Cicero, maybe even native speakers. Don’t know how to say Cicero because they always read it. And anyway, so Cicero and all the traditional rhetoric theory, and we found some very nice, very useful tips like the canon of rhetorics where you start with the idea and then you work on the structure and then you elaborate your speech and then you add detail, you try to remember it. And in the end, Axio is not unimportant that you actually, you do a good speech. This is something you need to train and you need to focus. And all of this is completely left out in PowerPoint. You just focus on moving boxes around all the time. And so study getting excited about it and getting so excited that we worked on this for three years until we had the structure where we say, okay, so our presentation, you start with the idea.
(00:15:13): If you have nothing to say, maybe you shouldn’t do anything. You start with the idea, you write your speech, then you add visuals where it makes sense, not because you have a slide where you need to have something and your speech will not be something that everybody needs to see that you put on every slide. So you don’t forget when you present, but it’s something that only you see. So it’s split between what you say and what you show is a fundamental part of it. And the fundamental part of it is also that you don’t use stock imagery and stuff. So we actually, we have some secret functionality that we don’t show. So we have Unsplash actually built into present, but we never unlocked it because we feel like it would be detrimental to the idea. So we have a sort of, I would say a Protestant, a Protestant tendency in our company where we don’t do things that would be possible.
(00:16:15): We also have hidden in the code, we have graphics, how do you say, charts and graphics and so on that you could write, but we haven’t unlocked it because it’s not good enough. We have back links in IA writer, but we haven’t unlocked it because it’s as good as everything else and we want to unlock it when it’s better. So we’ve had that for nine months now in writer. We’ve had backlinks, but we don’t download because it’s messy, it’s nasty and it’s as good as other apps, but we feel like it’s below the level of what we have in I writer. So that Protestant thing of we keep it back, we could do it, but we don’t because not to make other people suffer like the Protestants, that’s the Protestants idea. You always have to suffer a little bit, otherwise you don’t go to heaven. That’s not the idea. But we kind of sometimes a bit frost frosty with these new features, but I think it’s good. So to make sure that we keep the original vision. And that brings you back to ia, right? The vision is always that to make a tool where that is made for writing and that means that where you enjoy writing as well. And I noticed lately that means where you enjoy thinking, which is almost a paradox because thinking is very painful. Yeah,
Andy Polaine (00:17:41): I want to get onto that. I’m going to get onto this. Just one sec. I’ve got a question here from Nicole. Hi Nicole. And she asked whether it uses reveal js under the hood I present
Oliver Reichenstein (00:17:52): When it comes to technical matters. I’m very, I think yes, you’re not, I’m pretty sure. Yeah, I think I heard reveal before. I leave that to the developers and I’m 99% sure, but sometimes I say things technically and they tell me no, but I’m pretty sure it’s revealed 99%.
Andy Polaine (00:18:12): Alright. Alright. So yeah, just to finish off that, I think one of the things that used to try, and I guess when I learned this, but it’s the idea that the deck is not the presentation. I think this is a thing in consulting that gets conflated all the time and for a lot of people when they’re presenting, and it’s partly because you start with the deck and usually in consulting you don’t start with an empty deck. You go, here’s all this kind of other stuff that we really have and you make a Franken deck out of all those different decks and then you’re kind of stuck with all that. So let’s move on to IA writer. So I want to touch on the kind thinking thing actually the, there’s a designer called John who’s one of tomato. He’s been a very long experienced designer and he often said to me he missed the times when Photoshop was, well, machines were slower and you’d apply a blur filter in Photoshop and go off and make a cup of coffee whilst it was doing it chugging away the kind of blue baring because it gave me time to think about what I wanted to do next.
(00:19:30): And so there is this though, there is this connection between writing and thinking. One of the things I learned as a writer was keep the structural stuff of, and that’s also formatting, but also keep the structural stuff like the outline and what you’re planning to do and editing separate from the actual act of writing the generative thing. Because when you combine those together, and this is always my message to my students, writing their thesis is often the first big bit of writing they’ve done. If you do the editing and writing at the same time, you get caught in that thing where you spend two hours writing the first sentence. And I used to write a column for Design magazine and have to knock out a thousand words every month. And I always knew the first two or three paragraphs they were just, I had to let myself write junk.
(00:20:19): I knew I’d go back because halfway through or towards the end I would realize, oh yeah, now I know what I’m talking about. And actually a journalist friend of me said, gave me a tip once. So take the last paragraph you wrote and cut it and paste it at the beginning and because at that point actually what you’re going to write about and that’s the introduction. So I think there is a kind of parallel there between this idea of thinking and writing. And I know there will be a bunch of people who are watching this who are from the PKM person knowledge management community who use obsidian. And for me I use both. And the way I think about it is I use obsidian for notetaking and for thinking and the outlining and I use IA writer to do the writing. If I want to do some writing, I’ll use that.
(00:21:04): And that’s for me the kind of combination of those two. One of the things that does also start to happen though is you, as I was saying before, you sort of write to think, right? So yes, you’ll think about what you want to write, there’s a load of stuff that happens as you’re writing where you start to realize and then, oh, this is what I actually mean and so forth. Which kind of brings us to the whole AI assistant thing, which is obviously part of the way you can use that is just to do the writing for you. And I’m pretty sure I read a thesis the other day that was mostly chat GPT, hard to say or you can kind of use it differently. So let’s go there because as you’ve built this into IA writer, God, the AI and IA thing is really annoying, isn’t it? You must be so annoying about that. You’ve kind of thought about this beyond just, oh, we’re going to add this feature or going to add this stuff. So maybe we can talk about you and what you’ve been thinking about in terms of writing with AI and Well, you wrote several pieces actually about what is it good for and when is it useful and when is it not?
Oliver Reichenstein (00:22:19): I mean this whole notion that writing is thinking this is just not my thought at all. This is wildly and widely shared in writing guides. Yeah, William Sin said it just exactly like that writing is thinking, but I’ve read that in various versions. I really love to read books about writing and they’re written under the condition that they have to show that they know what they’re talking about. And that pressure often leads to good examples. I think some of the best writing from Stephen King is about writing, actually not even, it’s even called like that. The book.
Andy Polaine (00:23:11): Sorry, go on. It’s great. It’s great. I’m about to That’s great. I was just say his thing was that writing is an act of telepathy, right? That’s my favorite line from That’s
Oliver Reichenstein (00:23:21): Great too. That’s true. But there are also various versions of this formulation. Writing is time traveling and so on. But yeah, when ti came out a year ago, I was taken by surprise how well it simulates human language at this point. And the big shock for me was not so much that statistical approach actually works and simulates human thought to the point where you often are not sure is it even understanding what it does because it can’t without the body I’m philosophically convinced you can’t understand. But it was shocking how well it can simulate thinking. And there were two big shocks. I tried it with some philosophical literature. We both studied philosophy. That’s also something that our, your breeders maybe your views maybe know. But it’s a funny coincidence that we also found out only a bit later when we talked back in the day.
(00:24:34): Now what was shocking is how well it understands complex philosophical texts. So when you read the critique of the pure reason in German, even now after almost 40 years, for me it’s really difficult. I need to fully focus and I’m trained in the matter, but I have to fully focus and I can consume maybe a page per hour or so when it gets the company. But with chat GPT, I can ask it what does that mean? How does that relate to transcendental perception or whatever? And it gives me pretty good answers. And of course you’re suspicious and often it hallucinates, but very often it’s spot on and it can help you. So if you use it as a dialogue partner to read complex texts, but just as you deal with a human dialogue partner, you’re not like, oh yeah, I’m not that much, must be it.
(00:25:35): But as a dialogue partner where you discuss with the person I used to, my best friend at university Miana was my dialogue partner back in the day and we used to read Ecker and all these impossible. We used to read Hago like the worst stuff. And it was fun in dialogue and I can use it like that. As you know, there are few people, like my wife doesn’t want to talk to me about philosophy. She’s majorly annoyed by that waste of time. And other friends, I don’t know, the few people really have the desire and have the background, the education and the will to discuss philosophy for many hours. Chachi PT does that. So that was a nice surprise. I have someone I can talk about the metaphysics, which is something I’ve been studying for two years now. Again with great passion. And I love this book now after many, many years where I used to hate it and with Chachi I can do this.
(00:26:40): That was a positive surprise. The negative surprise was that the bullshit you get from cha is very similar to the bullshit you get from corporations and from marketing talk and from management. Consultancies is almost indiscernible and I realized that we are all less original than we think we are because when we use language, we often use cliches and things we heard somewhere and we forgot that we heard it and we repeat this and we think we came up with it ourselves. We also work a lot of Jet GPT language works like that. However, there’s one major difference if you are really serious about communicating with other people, what you want to do is you want to transfer the impression you have inside yourself, find an expression that matches that impression as much as possible so that other people on the other side can pick up that expression and make an impression and incorporate it actually into an impression inside themselves.
(00:27:59): I think this is what’s really happening when we communicate, we have something inside something more or less vague, we call it feeling, and we try to find a shape for it, which usually is a form of language, can be visual, can be verbal, can be music, can be, when we say language, we often just think about verbal language. But this is what happens when we speak and we mean what we say. We try to find a shape for what we sense for what we feel, for what we think by whatever want to say and put that into a shape so other people can have the same impression. And actually dealing with chat pt, I started thinking about this in a very, very intense way together with chat pt. I think my first longer conversation was with it was if it can understand and at first it said it can and I almost convinced it that it can’t understand because it doesn’t have a body.
(00:28:58): And I tremendously enjoyed that conversation, which was slightly annoying because I didn’t expect that to happen. But it made me think because I thought that this is very, very helpful in many ways for writing. It’s helpful for preparation, it’s helpful for testing what you want to say. And then for me as a non-native speaker, it’s often very, very difficult to write English and not make mistakes that I just can’t see. I can’t spot them. Commas missing something is misspelled and then people read this and I feel like people, I’m very stupid but I just can’t see it and Jack CT can always spot these mistakes. So there are many things, but it’s really great. But all the downsides were very obvious to me from the very start. It’s in fact most people are going to use it like business consultants and just splattered the bullshit all over the place.
(00:29:54): And you said you just read the paper that was very likely written with J gt, we develop a sense for it, we develop a sense for the visual ai. When I see this lollipop stuff, I’m not impressed at all anymore and we start seeing it with language too, but it’s much, much harder because our senses for language are not as sharpened as our senses are for visual products. Where now most people I think that have a little bit of training in visual perception will see 99% of AI produced imagery and will be able to recognize that. Yeah,
Andy Polaine (00:30:36): That kind of, I talked about it when we were talking earlier, excuse me, where there’s the people know the uncanny value, this idea that A 3D rendering or it comes from robotics actually, but something that’s either the closer it gets to being human, it goes through that value of the more it’s a little bit off and a bit creepy and uncanny actually something that’s more abstract like an emoji or whatever or feels more we are more empathetic to than again, it has to be really super exactly human for us to go back through that valley. And I’m sort of getting that quite a lot with AI imagery but also with chat GPT, I’m kind of reading stuff and I think at least, so this is the challenge and we will get to authorship mode is I think I’m reading it and I think I’m thinking there’s a shift in tone here and it kind of feels like this has not been written by the same person and my suspicion is chat GPT. And obviously when I then put in a prompt around the kind of subject matter that I’m reading and I get kind of the same text back and like, oh, okay, yeah, I’m sort 90% sure. I do worry that I’m probably missing quite a lot else. So you’ve built this idea of you’ve built this functionality into IA writer seven called authorship. So tell us a bit about how that works and the kind of thinking behind it. I’m going to put the video up on screen whilst you were talking about it too.
Oliver Reichenstein (00:32:06): Yeah, so there’s this initial perception of how well it can imitate us and what it could lead to in a positive, in a negative way. It strengthened more and more the idea that it’s good to use JG PT as a dialogue partner when you don’t have a human being you can talk to about things and not everybody has an editor to work on their texts and not everybody has 20 philosophy friends. They can talk about the critique of the pure reason if they want to read it outside of university. I think it’s a very positive use case, but the ability to pretend that you think, to pretend that you’ve written, to pretend that you understand it’s huge, and this will be the standard use case. A lot of voices started popping up about, yeah, we need to have some sort of water marking some fantasies that there will be AI to recognize ai.
(00:33:10): It’s such a ridiculous idea technically. I mean you don’t need to think very hard to know this is just an arms race that will never be won by anyone. We thought that this sounds awfully familiar, this problem about originality and is it really you and so on. And it led back to a long, long internal discussion we had about copywriting copyright, right? Yeah, it’s a huge complex, delicate, difficult, dangerous topic to talk about because you’re in proximity to big corporations that use copyright and patent laws, which just exercise power, but it’s also very, very real, especially if you’re a graphic designer or indie developer and you see your stuff being used by these bigger corporations often that then claim copyright over it. It’s a huge problem and it’s practically impossible to solve because whenever you say something, you call someone else, you sound like a whiny little
Andy Polaine (00:34:22): Animal. Oh, you were on YouTube, can’t say that, but there we go, done
Oliver Reichenstein (00:34:25): It now. Whiny little animal of the can kind and there’s just no way to deal with this just from this perspective and we usually just avoid it. But as a designer you also know that sometimes you’ve stolen stuff and sometimes that you’ve done things that you shouldn’t have done. And I’ve done things I shouldn’t have done plenty in my life as a designer and as a human being and I’ve stolen designs that shouldn’t have and I vividly remember these, especially one case where I really shouldn’t have done two cases actually, but I really shouldn’t have done it and I still feel very bad about it, extremely bad to a point where the two people that I’ve stolen from, they probably like, yeah, it’s a long time ago and you corrected it and it’s fine. One time I’ve been called out about this and I corrected it right away and it was so embarrassing because I knew I stole content from another website and the guy came to me, he’s like, you’re a young man and you may not be fully aware of that, but this is not good.
(00:35:43): And I apologized and took it down and the other cases where I stole a design and I really ripped it off one-to-one because it was so great. It must have been like 2006 or something. I’m not going to say exactly, but it was terrible. It was just absolutely horrific and that stayed with me. I was like, yeah, the person you steal from kind of knows, kind of knows. It’s not like a headache where only the person that has the headache knows, the person knows as well, but you really know and you know shouldn’t do it. So we were like, you know what, this is the right approach for copyright. It’s about author, it’s not just about the person you steal from. It’s a bad thing and the person will get offended because you steal their time because to design or to write something to find shape takes so much time, it takes no time to copy stuff, especially in the digital reel.
(00:36:40): And so we were like, this is somewhat connected. What if the right approach for AI also is not about policing and patrolling and trying to catch people that are using AI like you were kind of aiming at before. It’s about them and maybe if they’re not just complete assholes, they feel worse, they feel even worse. It’s, it’s a very embarrassing thing to steal from other people and being called and called out and so on, but it’s not about being called out. You feel bad to begin with if you’re an honest designer, if you’re an honest person when you bullshit people. And we do that inadvertently sometimes, but when you say what you don’t understand, when you say what you don’t mean, if there’s any humanity left in you, you feel really bad about it. And I thought maybe this is the right angle that we do not try to create something that scans for chat CPT typical expressions because we have, they
Andy Polaine (00:37:47): Don’t work very well anyway.
Oliver Reichenstein (00:37:48): It wouldn’t work really well. No,
Andy Polaine (00:37:50): I’ve tried them and everything I put through it went a hundred percent human. I was like, I’ve just taken this generated from chat GBT and it doesn’t work.
Oliver Reichenstein (00:37:58): You could go down this path, but whenever you do something like that, there’s always a countermeasure as well. But it’s ethically the wrong approach. Ethically the right approach is saying you have to decide for yourself and you can use Chachi PT for spelling for grammar. Here we go again. There you go. For spelling, for spelling for grammar. I think editing’s fine Journalists have editors and they change a lot of text. I think it’s fine. You don’t need to have the academic levels of scrutiny where you’re not allowed to actually just transcribe a thought, but in the end it’s up to you and as an author and so we want you to build something like that. And we said the main problem we have currently is that if you are an honest person, but you use Jack CPT to write in the most honest way, you don’t want to pretend to say what you don’t understand or express what you don’t mean. You don’t have a possibility to discern what you got from Jack CPT and what you did to yourself. And then we have a technical problem. How the hell do you do this in markdown? And now we have a solution where people are like, yes. Some people say like it’s just marking off. It’s no big deal. It is super, super hard to do. We didn’t want to You didn’t want that signal. I’m speaking too long. No, no,
Andy Polaine (00:39:27): No. It was your video and I forgot to turn the audio off. I know. Yeah, it was this.
Oliver Reichenstein (00:39:34): There you go. It was super hard to do and now it’s kind of obvious. So then the circle closed where we were like, holy shit, now we have that copyright problem again. People are going to copy it. And our friends that obviously and our friends at Obsidian, they’ve been just ripping off our design up and down just to the pixel, how to reconstruct I writer and Electron. We like the team at Obsidian and obsidian great software does fantastic impossible things. My favorite function is how you can lay out stuff like an illustrator and stuff. This is all good, but ripping off our design, it piss me the fuck off because it’s not cool. It’s cool if you do it for yourself, but it’s not cool if you put entire leverage out there how to. It’s really not cool. To
Andy Polaine (00:40:27): Be fair to the obsidian team though, it’s not them, right? It is the plugin and kind of theme. Oh
Oliver Reichenstein (00:40:33): Yeah, I’ve talked to ano about this. When you started logging someone for being so innovative to invent syntax highlight, I was like, dude, this is not innovative. It’s just another rip off. And he was like, yeah, I’m sorry. It’s true. He knows the feeling because he’s also a developer and he gets ripped off all the time. But it’s also not an attitude honestly to say, well, it’s not our fault. We want to be flexible. Actually, I think, and I have to talk to them again about this, I meant to do that. I think they should tell their developers, Hey, please don’t make clones of other indie apps. If you make a clone of word you can punch up like this. But clones of other indie apps and because we don’t punch punch down or on the same level, but obsidian with being a free app has a pretty big audience at this point. I don’t think it’s punching down anymore when I say that, yes, have policy for what’s cool to clone and whatnot. If it’s pixel perfect colonial, it’s just not cool. It’s not cool and we haven’t said anything about it until the show, but I think it’s just fucking uncool this stuff. And it’s also back clone novel because what we do can’t, it’s not a native app and the degree we go to without typography is insane, but I’m not going to get
Andy Polaine (00:41:58): Into details. Well, I was just going to say I counter, I think there’s may be a fair counter argument to this. I’d be interested to see what your response is because I have seen people say, well, I like this bit, but I don’t like all the rest and I don’t want to be. Because I think it’s probably fair to say, and I think you were kind of one of the early, you said it at the beginning, but also sort of one of the early voices of the idea of opinionated app design where IA writer has, there’s opinion in there. There’s a lot of detailed thought and design philosophy that goes into it and it’s not going to be for everyone. And right down to whether you show the markdown and whether you don’t and all that kind of stuff, these are decisions I know you’ve thought long and hard about. I don’t think they’re, they’re not just random, but I was thinking of the German word IC and I can’t think what the English word of that is.
Oliver Reichenstein (00:42:53): I think random,
Andy Polaine (00:42:56): But at the same time people might go, well, I really kind of like the focus mode, but I don’t really want all the rest or the other stuff. The way you handle the kind of markup is not for me or whatever. Do you think that it’s fair then or would your answer be, well yeah, sure, do it for yourself, then write a little CSS snippet for yourself to enable that. But is your point really it’s not that. It’s actually the whole constellation of trying to customize and use a whole bunch of plugins and CSS to try and completely clone the exact look and feel within an
Oliver Reichenstein (00:43:32): I have absolutely no problem whatsoever. You can clone anything you want for yourself if you do it for yourself privately. I do that all the time. You learn a lot. It’s great. And maybe this particular person’s like, oh, the bold needs to be gray. We have, like you say, we have a good reason why we don’t do that and we just say no, we thought about this love and heart typography kind of something. Yeah, we’re very opinionated in a way, but we also know what the hell we’re doing there. And so yeah, if that is so important for you, you can replicate that in whatever app you want. That’s fine. If you do things for yourself, it’s totally fine. What I think is absolutely horrendous is to make full packages that allow you everyone to clone our apps functionally design wise down to the pixel, which actually is not possible anyway, down to the pixel in another free app.
(00:44:39): This is totally onco. And people thinking like, oh yeah, it’ll be free advertisement for you or something, fuck off. No, we don’t need free advertisement, we’re doing fine. That’s not the point. It’s about stealing someone’s design and make it available for everyone. It’s called counterfeiting and if you want to, it’s illegal, but that’s not the point. It piss us off because it’s our work and you’re not supposed to distribute it freely to everyone. Now if you’re not happy with a certain way that we design our app and you want pretty much everything that we have except for some asterisk risks that want to be great, well yeah, do what you can and build, rebuild the app in Xco if you want a native feeling. No, I’m not going to provoke people now with this. No, no, it’s fine. It’s fine. The problem we have is, and we didn’t realize that for a long time because like I said, usually we stay silent, don’t say anything. And actually these plugins are, they’re not even that popular and they’re being left there and they decay and it’s not a big commercial problem for us, it’s, it just pisses me off because the attitude is the wrong one. This is not something you can just take and distribute freely to other people. It’s not yours point.
Andy Polaine (00:46:09): You have bigger problems in the app store though with things like AI writer and things, using your logo and everything and just kind of switching the letters around.
Oliver Reichenstein (00:46:17): It’s not a commercial. We’ve solved 3 million apps in 13 years. You have 500,000 users. Everyone has paid our app. I don’t know. I don’t know. We have made some experiments with a free tier on Android. It was an absolute disaster. It attracts completely the wrong people for the type of app we have. Anyway, the last word on opinionated, I hate that word. I
Andy Polaine (00:46:42): Thought it might.
Oliver Reichenstein (00:46:43): And the problem is I’m not a native speaker, so I don’t have so many context for the word opinionated. The only time where I hear opinionated, it’s negative and I know it’s been used.
Andy Polaine (00:46:55): No, you just used quite a lot. Yeah, yeah,
Oliver Reichenstein (00:46:58): Yeah. People have been trying to use this in a positive way, but every time people say opinionated, I’m like, no, it’s not opinionated.
Andy Polaine (00:47:04): No, I think
Oliver Reichenstein (00:47:05): I have opinions, but we don’t do that because we think our way is the only right way and you have to follow because
Andy Polaine (00:47:15): We know. Yeah, but that’s the idea. I mean when people talk about an opinionated app, what they’re saying is, this is my opinion. I mean, is it log sec? I think it’s that one, which it takes that view as well, which is this is the approach I’m going to take to this. I have opinions about the style or the mental model or the UX or whatever it is about this, and I’m going to apply those to the thing I’m making, which is fair enough, I’m making a thing. And you have to have an opinion. The worst way is down the other end I think where you get the design by committee. A lot of enterprise software is like this and a lot of open source stuff is like this actually where it’s kind opinion less and it’s generally really watered down or incoherent and inconsistent because of this, I think. So from my view, I think you have to have an opinion about we’re going to choose this because an opinion helps you make a choice. It’s a filter. Say we’re going to do this instead of this because our underlying philosophy is based around that. So that’s what I mean by that. But I know .
(00:48:29): I want to get back to this idea of interoperability because partly it’s in my head because I’ve been listening to and reading Cory Dr Hours, the internet con and the heart of it is this idea of interoperability or avoiding it. And that’s what big tech has done is obviously is tried to see it going on with the Confederated thing at the moment, tried to avoid interoperability in order to create this huge ward gardens and monopolies. And so one of the things briefly, I guess one of the things that authorship mode does in IA writer is when you are pasting stuff in, it will keep a reference of what you’ve pasted and who the author is of that stuff. Which for me is great because one of the things that happens I find when I’m working with other people is we lose track of who wrote what.
(00:49:24): And as you mentioned in the article, track changes in Word or using, I know there’s a school of working or writing which uses Git for writing. It’s pretty horrendous really in the end as a kind of user experience. And so you’ve come up with this approach to do this, and in it you sort of grade out stuff that gradually you, as you rewrite the AI written stuff or the other author’s written stuff, it kind of turns black. And so you get the sense of how much is yours and how much is written by someone or something else. But you’ve also released this on GitHub as markdown annotations. So given the conversation around features and not being ripped off and so forth, why release this as a on open source as a spec?
Oliver Reichenstein (00:50:22): So this was a heated debate. This was a heated debate because we also run a business. We do, and the first thing that happens when we do something new is we get ripped off regularly. And I know how whiny and how little animals this sounds, but it just, it’s a reality. It’s unbelievable. And you don’t know. You have no idea how much that happens because we never say anything, but we see all this mostly because people report that when we came out with ia, right? We had different males every day of people pointing to this and then other, and you develop a thick skin, but it’s also annoying. And then from a business perspective now, it would’ve been much clever to just keep this for some time until someone copies it or copies something similar and then it’s not interoperable anymore. But the reason why we like markdown most of all is because of its interoperability.
(00:51:24): It’s fantastic. I copy paste stuff all the time, and this is something that a lot of people don’t understand about markdown, like, oh, can you have an export here, an export there? But actually it’s about copy and pasting stuff I copy and paste it from, I write to WordPress chat GPT and back, and it always works. Also chat GT uses smart dams. Fantastic, right? It’s really cool. There’s interoperability for that. So we were like, okay, so there’s the interoperability aspect where we can’t just ignore this and rather than people doing something similar, but then it’s not compatible anymore, we should actually publish it. And then there’s the other aspect where we say, well, author ship’s up to you. You need to decide. So we said, so why don’t we formulate it like that? Well, it’s not something clear cut like the legal texts. And like I said, it’s not about legality.
(00:52:16): When we talk about copyright, it’s about morality and it’s not clear cop, you can do this, but not that. And the other thing, because design doesn’t work like that design, when you copy design it and the copy person kind of knows it as well. So we said, well, let’s do it like that. You can use the functionality just as it is. Well, if you can help us developing it. And we got a lot of really great feedback on that as well. And we will develop it based on that feedback, which is fantastic. We didn’t expect that because usually when do so something open source, you don’t get much feedback. People just use it and that’s it. We seen that with the phones, right? Anyway, I’m not one, I don’t want to complain too much. What we want is people to use this and do their own thing with that. And if they do their own thing, it’s fantastic and it’s completely up to them. And that doesn’t mean you can’t abuse it and just say, well, you serve it up to me so I can cheat. That’s really not the idea. They have to get into the reality of the creative business where you need to find a way with yourself to come to peace with how much you copy and how much you don’t. And then you need to live with it. That’s our take. And we’re not going to call out people.
(00:53:38): I’m not saying categorically not calling out people because if you go too far, you go too far. But we reserve the freedom to say something if it’s just ridiculous. But very likely we’re not going to say anything, even though we don’t agree. It’s up to you. The thing though is if you are not sure, you can always talk to us. And that’s also a reality that few people grab that when they do something and they’re not sure, will this piss off the other person? They talk to you, and I always appreciate that,
Andy Polaine (00:54:16): But I think to end up, but I’m really pleased you to have released it because I find the interoperability part is the critical bit of plain text, the markdown, the whole fact that can plaint text is very robust. There was a little conversation about it on Mastodon. You can have several apps open at the same time accessing the same plain text file. And even when they’re, most of the time they don’t clash with each other, obviously just move from one side to the next for whatever you’re doing.
Oliver Reichenstein (00:54:50): However, it’s not very realistic that this is going to be a standard or something even. Well, I
Andy Polaine (00:54:59): Hope I really
Oliver Reichenstein (00:55:00): That a lot of people are going to use it. I’m not very hopeful on
Andy Polaine (00:55:03): That. You’re not. I am, I am. No, it’s
Oliver Reichenstein (00:55:05): More,
Andy Polaine (00:55:07): Well, because the perennial pain for anyone who’s writing and having to collaboratively write and like I said, and you’ve got an editor and stuff, is at some point having to go back into Word or Google for track changes. I know this critic markup, I’ve had a go. Maybe if I was a proofread a bit like the way proof readers know all those little symbols, they used to kind write on a manuscript by hand. You kind of remember all the markup. I just found it too much for me to remember at the time. And I really liked this as an approach. So I hope it does because I really want the sort of collaborative markdown writing tool that doesn’t involve some kind of weird gymnastics game between putting stuff up onto Google and then exporting again. And I’ve tried it. None of it ever works. Even there’s a sort of ether pad kind of plugin for obsidian that sort of attempts that, but it doesn’t really work. So that’s the holy grail for me. And so I hope that it does become a standard. Hey, we’re coming up for time. Where can people find you online? Are you basically on IA net? Is that where they find you? Yeah,
Oliver Reichenstein (00:56:21): I moved away completely from Twitter. We have some bad conscience tweets here and there sometimes because we think, yeah, few good people that are still left on Twitter and there are some good people still left. This is by the way what we did. But we will move completely over to saying subscribe to a newsletter or come to Mastodon where I am right now. I think I feel very comfortable on Mastodon. We are at some instance on Mastodon and at Mastodon, what I used to do on Twitter. Now we do there. It’s great. The audience is perfect for in the developers and for people that think alike. Still. One more last word about this, about whether it becomes a standard or not. I’ve been dragged into that word being forced to use word again and again myself. I’m not naive. I know what the standard defines.
(00:57:24): It’s the most used apps, Google Docs afterward. But I think it’s really important that we do what we say and that we act accordingly. And so putting this out there, of course there’s a little hope that more people will pick this up and we would be able to use authorship in different places, but mostly we have to offer the possibility to do that. And we have to stay true to what we say about authorship and copyright and allow people to make their own decisions in that way. And if we said, no, this is our standard. You can only use it here. It may be profitable short term, but it won’t be very useful. What we did instead is we tried as much as possible to look at this back and forth. If you copy a text in that is different, you can say, where does it come from? And we spend a lot of time on that as well. But if other people use it and evolve it and find another shape of it, maybe even a better one where we would be tempted to copy and improve it, because that’s what creativity is. You copy and improve to the point where you don’t necessarily recognize the original at one side, then that would be really, really cool.
Andy Polaine (00:58:42): But it would be nice, I mean, I think one of the gifts to the world of John Gruber was to kind of not try to own the markdown spec, especially as there’s been other variants of it and there’s all sorts of different kind of flavors of it now. And he’s resisted a lot of pressure to change his original spec. This is the spec if you’re going to go off and do something else and go off and do something else. But I think he’s changed one thing I think, I can’t remember which one it was. I dunno if it was images or footnotes. I don’t think footnotes are in the original spec and it’s become this great thing that’s everywhere, even in chat, GPT. So the show Power of 10, it’s named after the Ray and Charles Eames film about power. It’s called Power of 10 Powers of 10, sorry, that’s what it’s called. And it’s about the relative size of things in the universe. And I’ve always really liked it for that kind. Different levels of mental zoom and the repeated patterns at different phases. So on that front, the final questions always what one small thing is either overlooked or could be redesigned that would have an outsized effect on the world.
Oliver Reichenstein (00:59:59): Wow. It’s not, I never watched your show, which I should then I would’ve known this. One thing that I think, I don’t think it’s a little thing, but one thing I noticed is the trash can. We separate the trash. We have organic trash and plastic bottles and cardboard, and we have a big box now of different trash elements and we do our best to separate that. And yet for some reason, whenever I look at my trash can, I feel like what bunch of evil people are we? Every time I look at my trash can, I’m like, yeah, you’re not trying as hard as you can because there’s still too much stuff in there. It still fills up too quickly. And I wonder, I don’t know how, I don’t have a solution, but I wonder if someone could have a look at this and find a way to redesign the trash can in a clever way, that it helps us to not fill it up so quickly with all kinds of crap. My brain starts working already on it about labels. I think about small, making it smaller or whatever, but making a smaller opening. I’m not sure, but I think this would change a lot.
Andy Polaine (01:01:18): Okay, well, I’m
Oliver Reichenstein (01:01:20): Looking away way too much stuff.
Andy Polaine (01:01:21): I’m looking forward to the IA trash can. You’ve got these beautiful notebooks you just made. We didn’t even talk about those. I’m looking forward to getting my hands on one of those too. Oliver, thank you so much for being my guest on Power of Ten.
Oliver Reichenstein (01:01:35): Thank you as well. It was a pleasure
Andy Polaine (01:01:37): You have been listening to and watching Power of Ten. You can find more about the show polaine.com, P-O-L-A-I-N e.com, where you can also check out my leadership coaching practice online courses, as well as sign up for my pretty irregular newsletter doctor’s note, although I will be doing an end of year one. If you have any thoughts, you can put ’em in the comments below. If you’re on YouTube, you can find me at apolaine, A-P-O-L-A-I-N-E, and on pkm.social on Mastodon. All the links are in the show notes. Thanks for listening and watching and I’ll see you next time.