Latest / The Two Authors’ Podcast / LLMs for the Assist with MJ Caan

Transcript

LLMs for the Assist with MJ Caan · The Two Authors’ Podcast

• 0:00
  There's something about stretching yourself into an unnatural position in a really hot
• 0:04
  room covered in sweat and, you know, that just makes you, I don't know, limber for hours
• 0:09
  and hours and it feels so good.
• 0:12
  Doug said the same thing on a trip in Singapore, actually.
• 0:15
  Yeah.
• 0:19
  Welcome to the Two Authors podcast, an entertaining chat show hosted by bestselling mystery and
• 0:25
  thriller authors Douglas Pratt and Nicholas Harvey in lively humorous and informative
• 0:31
  conversation with distinguished guests who share their thoughts and experience about
• 0:35
  forging a successful author career.
• 0:41
  Welcome to this week's show brought to you as always by the fabulous most fastest formatting
• 0:46
  solution in the world.
• 0:48
  Who's that?
• 0:49
  It's vellum.
• 0:49
  Who?
• 0:49
  It's vellum.
• 0:50
  Who?
• 0:55
  I don't know why they don't know you know what we should talk them into letting us stand
• 0:58
  out front of them at authorization just going it's vellum.
• 1:03
  I'll bring the yoke.
• 1:04
  Kind of like the Ricola guy.
• 1:07
  Vellum.
• 1:08
  Ricola.
• 1:09
  Yeah, let me stand in the corner.
• 1:11
  You know I was uh people would skirt like go all the way around the other side of the
• 1:16
  hole just to avoid test then.
• 1:19
  Maybe maybe vellum could pay us to go stand on the other side of the room and send people
• 1:23
  their way. Yeah. You know, they, I was just talking, we were, we were talking the other
• 1:30
  day with some people about the, you know, so much translations and stuff. And they've
• 1:34
  done this cool, they've expanded the more of their languages, like they've got not built-in
• 1:39
  versions for like German, French, Spanish, Italian, Dutch, Portuguese, I bet they have
• 1:45
  Klingon.
• 1:46
  Klingon.
• 1:47
  I bet they could.
• 1:48
  Yeah.
• 1:48
  They'd be brilliant.
• 1:49
  It's an expanding language.
• 1:51
  You know, they do a duolingo for it.
• 1:52
  So why not?
• 1:53
  So it's really cool.
• 1:55
  I could just get built in the information.
• 1:57
  So you have to worry about like the, like the chapter title page, all those words
• 2:01
  that comes up in the correct language.
• 2:02
  It comes up right.
• 2:03
  It's kind of cool.
• 2:04
  So I like that.
• 2:05
  Yep.
• 2:06
  They do make everything so easy over there.
• 2:08
  I just, that's the biggest thing about Bellam that makes them so
• 2:11
  fantabulous is how easy they make it.
• 2:13
  Yeah.
• 2:13
  Yeah.
• 2:14
  It's super, super easy.
• 2:16
  But yeah, also a big thank you to all of our Patreon supporters.
• 2:19
  We are very appreciative of you and thank you so much.
• 2:22
  And if you guys feel like you're getting anything out of this show,
• 2:25
  then please come visit us over at patreon.com/thetwoauthorspodcast.
• 2:31
  We would like you to buy us a beer.
• 2:32
  It's only $3.
• 2:33
  It's a cheap beer, like a Pat's Blue Ribbon.
• 2:36
  That's not per episode.
• 2:37
  That is per month.
• 2:39
  That's right.
• 2:39
  So you would break that down, though.
• 2:41
  I don't know.
• 2:41
  That's a math I don't want to do, but please check us out nonetheless.
• 2:45
  So, but hey, we're going to have a really cool show today.
• 2:48
  What we got? What we got? What we got? We got a good friend of ours, M.J.
• 2:51
  Kahn, who writes science fiction, paranormal fantasies.
• 2:55
  And he's pivoting to a more fast paced, character driven thrillers under his new pin
• 3:00
  name, Imber Scott. How cool is that name?
• 3:02
  I don't know. Imber Scott.
• 3:04
  It's a slow burn, people.
• 3:06
  Yeah, very slow. I'm Imber Scott.
• 3:09
  That should be a character name, almost.
• 3:11
  I know. He kind of blew that on his author name.
• 3:14
  He's got the coolest author name.
• 3:15
  He'll never have a character name as cool as his author.
• 3:18
  And think about how cool it would be like if he had a website and Ember Scott like was burning like into the font
• 3:24
  I mean hey Doug I got news for you
• 3:26
  He has a website
• 3:29
  But oh my gosh, but he's I think it does have a little flame thing going okay, then that's cool
• 3:34
  That's a he he's see he's smarter than us. I'm kind of sort. He's a lot smarter than us as for sure
• 3:39
  Yeah, he's gonna join us and but firstly a listener question from somebody even smarter than that and that was our recent guest Maggie Maher
• 3:47
  So Maggie Ma, thank you so much for sticking around to do a listener question with us. It was a wonderful
• 3:52
  Chat with you. We had a few few weeks ago. So now you get to turn the tables. What do you got for us?
• 3:57
  What if I got for you? Okay, I have a question
• 4:00
  What are you finding to be the best bang for your buck as far as marketing right now?
• 4:09
  Right now isn't a day. Yeah, absolutely. I'm always interested in how independent authors are marketing
• 4:17
  Really good question.
• 4:18
  - It's an outstanding question.
• 4:20
  And if you asked us last month
• 4:21
  and if you asked us in a month,
• 4:23
  they would probably move around.
• 4:25
  - I'm gonna lean toward Amazon ads
• 4:28
  or probably the best right now.
• 4:30
  For bang for your buck though,
• 4:32
  I'm gonna actually lean,
• 4:33
  I'm gonna back it up and say my newsletter is still my,
• 4:36
  for my bang for my buck,
• 4:37
  I think I could get more or money out of my newsletter
• 4:39
  than I do out of,
• 4:40
  if I throw my grandkids in there,
• 4:42
  then it almost doubles how many sales I do.
• 4:45
  but he only has a few grandkids and he's already sold two of them.
• 4:49
  So he's running out of grandkids to sell,
• 4:51
  so it's not a great business model.
• 4:53
  - We've got a fourth one coming right now,
• 4:55
  so we're- - They can't push him out fast enough.
• 4:57
  - We're generating them pretty quick.
• 5:00
  - And they're wondering where they're going.
• 5:01
  They go say who the grandma and grandpa
• 5:03
  and they never come back.
• 5:05
  All right, so I'm gonna agree with Doug.
• 5:07
  I think it's a really weird time right now, right?
• 5:09
  You know, there's some people certainly killing it
• 5:12
  with Facebook ads still,
• 5:12
  and there's some people killing it with Amazon ads.
• 5:14
  I think some of the extra reach you can get through certain methods with Amazon that's
• 5:20
  available to some people, there's some opportunity there.
• 5:23
  I would say my Amazon ads are being more productive than any Facebook ads at the moment.
• 5:27
  Facebook is becoming really, really tricky.
• 5:31
  You have something that works and it just stops working.
• 5:33
  It's really, really difficult.
• 5:35
  So I feel like it's back to basics.
• 5:38
  It's back to the release regularly, get a new book out there, build your newsletter
• 5:42
  or as organically as you can and keep in touch with your audience and keep giving them what
• 5:48
  they want. You hate to use the term "tread water" because you don't want to ever tread
• 5:53
  water, you always want to be moving forward. You're better off treading water than pounding
• 5:57
  down the wrong avenues at the moment. I feel like there's a lot of change and a lot of
• 6:01
  things that are going to sort themselves out. So experiment, try stuff, but try carefully.
• 6:05
  I would say don't be too scared of it like that. A lot of people are worried about it,
• 6:08
  and including myself. I worry every time I get on Facebook trying to do one of their ads.
• 6:12
  Now I've got them all turned off, but you know a couple of weeks
• 6:14
  I'm gonna turn them back on and try again because that's just the way it is
• 6:17
  So I think you have to be in flux, you know, we always talk about pivoting
• 6:22
  I think we have to pivot at the drop of a hat too. So yeah, be ready. Yeah tricky times great question
• 6:27
  Yeah, thank you. Thank you for the answers. Thanks for sticking around. We appreciate it
• 6:35
  MJ how you doing man? It is great to see you. Oh man, it is great to see you guys. It's been a while
• 6:41
  Well, it's not too long.
• 6:43
  No, no, a couple of months, we got to hang out together
• 6:46
  in Texas a little bit and that was superb.
• 6:48
  So, and right there, and we talked about getting you on
• 6:51
  to talk about some of the stuff.
• 6:52
  So, Starter List of Soft was a little bit of background
• 6:55
  on how you first got into writing
• 6:56
  and when you were first published.
• 6:58
  Okay, yeah.
• 6:59
  So I've been writing, gosh, a little over a decade now.
• 7:02
  I started back in 2012, 2013.
• 7:06
  So just after the, I guess, maybe the first gold rush,
• 7:10
  back in the days of like cable words, you know,
• 7:13
  that's where I cut my teeth on, you know,
• 7:15
  that wasteland there, you know,
• 7:16
  you just kind of venture into that.
• 7:18
  (laughing)
• 7:20
  - Are they even there?
• 7:21
  I mean, I think they are.
• 7:22
  - I think they're still there, but you know what though,
• 7:24
  everybody who didn't get in on that,
• 7:26
  they just don't know what they were missing out on.
• 7:28
  You know, you took you a digital knife
• 7:29
  and you just waited into the mess with that.
• 7:31
  (laughing)
• 7:32
  - I can't imagine now it's a nightmare
• 7:35
  'cause the people that are in there now are probably like.
• 7:38
  - Yeah, but that's where I learned about, you know,
• 7:40
  self-publishing, you know, because before then, of course, you know, you had the gatekeepers
• 7:43
  was trad publishing, you know, I would like type something out and print it off and mail
• 7:49
  it in the mail to, you know, I had my big giant writer's market book that I'd look
• 7:52
  stuff up from and, you know, try and do that. And I guess I just, I've always wanted to
• 7:58
  be involved in storytelling. It wasn't necessarily writing maybe, but maybe from like the youngest
• 8:02
  age, you know, I just wanted to be involved in, you know, creation, creating stories.
• 8:08
  I love to read.
• 8:10
  I feel like I could read from as far back as I could remember.
• 8:14
  I started off reading and fell in love with the X-Men.
• 8:16
  Comics was my first great love, and I fell in love with the X-Men, and just wanted to
• 8:21
  create that.
• 8:21
  And then I kept reading more and more as I grew, and it got to the point to where I
• 8:27
  was working in the corporate world and decided that, "You know what?
• 8:32
  I wanted to go back to my first love, which was creation."
• 8:35
  So I started writing discovered writers market to figure it out.
• 8:38
  You know what?
• 8:39
  I was reading books.
• 8:40
  I'm like, I can write a better book than this.
• 8:42
  We don't say that.
• 8:43
  Now we can't do it.
• 8:44
  Maybe we can't do it, but in my mind, I'm like, I can write a better book than this.
• 8:47
  So, you know, the book is yes, there are some that for sure we can write a better
• 8:51
  book than even I can write a better book than some.
• 8:56
  Well, it's not stretch, but you know, okay.
• 8:58
  All right.
• 8:58
  So.
• 9:02
  - Bam, that's fair.
• 9:04
  - So I did, I wrote my first big novel, 100,000 words,
• 9:10
  and it was a sci-fi, almost like a space opera,
• 9:15
  genetic engineering, cross mashup of everything.
• 9:18
  And so I got into K-boards to figure out
• 9:21
  what do I do with this now?
• 9:23
  That's why I discovered editing, cover art,
• 9:26
  all of how to format something.
• 9:28
  I actually had to format with Word
• 9:30
  and then format with Scribner
• 9:31
  because Bellum did not exist.
• 9:33
  - The dark days.
• 9:34
  - Don't even tell you.
• 9:35
  - The dark days.
• 9:37
  - The dark days.
• 9:37
  - Formatting a word almost made me quit writing.
• 9:40
  I hated that so much.
• 9:41
  - It would take longer than it would to write the book.
• 9:43
  Just to sit there and format all those chapters
• 9:45
  and get them ready to go.
• 9:46
  - And you had to go put it on the Kindle
• 9:48
  'cause they didn't even have the preview.
• 9:51
  - The preview on the KDB dashboard.
• 9:53
  - I had to actually upload it to the Kindle
• 9:55
  so that I could see what it looked like on the Kindle
• 9:57
  and then go, "Ugh, there's lines."
• 9:59
  - Do it again.
• 9:59
  - There's spaces all over the place.
• 10:01
  It was a nightmare.
• 10:02
  - Yeah, so I did that and this was back
• 10:06
  when there wasn't a lot of, there was competition,
• 10:08
  but not a lot of competition.
• 10:09
  There was no such thing as ads.
• 10:11
  It was really just word of mouth.
• 10:12
  Other authors would push you, you'd push them,
• 10:15
  you'd talk about it and just, I think Amazon
• 10:17
  just really pushed everything back then
• 10:19
  because they wanted to get the readers reading
• 10:22
  on those Kindles.
• 10:23
  So if you put something up, it's sold.
• 10:25
  You'd make you a Microsoft paint cover
• 10:27
  and do whatever you had to and it would sell out of the box.
• 10:31
  So that's how I got started.
• 10:32
  And so since then, I have kept the,
• 10:34
  I've used the MJ Khan, my pen name to write science fiction,
• 10:39
  but then also, paranormal was where I really broke out.
• 10:42
  I did paranormal and urban fantasy.
• 10:44
  And luckily I had some things that were already on the Kindle.
• 10:49
  And then when the pandemic hit
• 10:51
  and everybody was locked inside reading,
• 10:53
  I just saw my paranormal stuff just really break out.
• 10:57
  I wrote paranormal women's fiction,
• 11:00
  which just happened to be at the same time
• 11:02
  that a group called the FAB13, some amazing authors,
• 11:06
  created a Facebook group and they launched
• 11:08
  all of their stuff in that genre.
• 11:10
  Mine was like right there with it
• 11:12
  and the first book that I had that I launched
• 11:15
  in that genre, I think it hit like 700 in the store
• 11:18
  somewhere around there.
• 11:19
  And there was just an audience waiting
• 11:22
  for anything paranormal related.
• 11:24
  So I just kept writing and that led to me
• 11:26
  becoming a full-time author and here I am now still.
• 11:30
  Right in the way.
• 11:32
  So you're making a shift now towards the thrillers.
• 11:35
  Yes.
• 11:35
  So probably a couple of years ago,
• 11:37
  I started to look at that because, you know,
• 11:39
  I wasn't burnt out on paranormal,
• 11:41
  but I was starting to see that there's a definite ceiling.
• 11:43
  There's a ceiling in all the subgenres, you know,
• 11:45
  and they're getting crowded.
• 11:48
  But, you know, it was that creep that I saw coming in
• 11:51
  to where there were a lot of man chest covers that, you know,
• 11:55
  I'm like, that's not really the genre,
• 11:57
  but it was taking over the genre, do you know what I mean?
• 12:00
  And I was like, you know, I don't really want to play that game.
• 12:03
  And then I was just needed to find something that I still really enjoyed
• 12:06
  writing and understood.
• 12:07
  So, you know, what I did was, you know, I looked around, I said, you know,
• 12:11
  what do I really, really enjoy other than things with magic and aliens in it?
• 12:15
  And, you know, I looked at my Netflix queue.
• 12:17
  I looked at my one clicks.
• 12:19
  I looked at, you know, my Amazon Prime subscriptions.
• 12:22
  And I realized that just about everything had to do with mystery,
• 12:27
  thrillers, action adventure, that thing.
• 12:30
  And it's amazing what you will absorb
• 12:33
  in your subconscious when it comes to understanding tropes,
• 12:35
  pacing, everything about a genre that you really, really love.
• 12:39
  And this was one that I loved.
• 12:40
  So I kind of had an intrinsic knowledge of how it worked.
• 12:43
  So I started writing that.
• 12:44
  I started a new pen name, Ember Scott,
• 12:46
  that I published my first Jesse Knight thriller
• 12:48
  under that name.
• 12:49
  - Best pen name ever.
• 12:51
  - Yeah, and just.
• 12:51
  - Best pen name.
• 12:51
  Yes, we raved about your pen name.
• 12:54
  I was like, that is a great pen name.
• 12:56
  - Well, Ember's my dog's name.
• 12:57
  So, you know, it's perfectly good.
• 13:00
  Yeah.
• 13:01
  But yeah, it still has a very smoldering sound to it.
• 13:03
  Yeah, I like it.
• 13:04
  You know, so it's just like I published my first thriller under that and it did, you
• 13:08
  know, it did fairly well for a first book with no following, no news list.
• 13:11
  I didn't have an existing, you know, base of readers and it's just grown from there.
• 13:15
  And I have loved every bit of it and have no, you know, no issues with what I've chosen.
• 13:21
  One thing you've done that's smart and we also do is you're looking at what you like
• 13:26
  to read and what you like before you start writing.
• 13:28
  I love that description.
• 13:30
  Yeah, as you went through all the things you looked at to go,
• 13:33
  what is it that I'm attracted to?
• 13:35
  Yeah.
• 13:35
  You could have gone, you know, a lot of people were making a lot of money writing a hockey romance.
• 13:40
  And but that's not something I would want to write.
• 13:42
  So I think my hockey romance would probably be pretty terrible.
• 13:45
  He's like, yeah, your hockey taranty romance would be terrible.
• 13:47
  So I was going to say I could write.
• 13:53
  I could write it, but it would be it would end up with one of them killing the other.
• 13:56
  with the hockey stick. So you know those those those skates that they have on my
• 14:01
  I think of all kinds of things to do with those things.
• 14:04
  And then like oh but all the the words you can have at the end I gotta go
• 14:08
  fuck yourself. So you know that would be and if I'm like if I'm gonna do that
• 14:15
  and I might as well that's what I enjoy you know so you know like I said I knew
• 14:19
  the tropes I knew the universal fantasy you know even if I didn't know it
• 14:24
  exactly. It was in the back of my mind because I read it all the time. And I think the hardest
• 14:28
  thing about thrillers a lot of times is that who done it aspect of it and the pacing for it. And,
• 14:34
  you know, a couple of the books that I've written have been not so much a who done it, but a how
• 14:38
  they did it. So, you know, I might start off where they're working their way through how it was done.
• 14:42
  But that's just something you can feel your way through. You know, I didn't have to study it. I
• 14:46
  didn't have to learn it. It was just there. And so it just came out of me. Actually, I can write
• 14:50
  fast and it just, yeah. You've been watching it. You've been read. That's what you're reading.
• 14:54
  You already understand that buildup of the story and stuff. So that does help a lot, I think. So
• 15:00
  and now we're here from this episode sponsor, the author engine, how to write, launch and scale
• 15:05
  like a seven figure publisher. Tony Lee, writing is Jack Gatlin has nearly half a billion page reads
• 15:10
  25 crime thrillers written for Marvel DC and the BBC. He has built the machine at author nation.
• 15:17
  He teaches you how to plan launches backward from pub date and treat every book as an IP asset.
• 15:22
  Publish your first thinking not writer first.
• 15:25
  Author nation.live join us November 10th through 13th in Las Vegas.
• 15:30
  One of the things that we've spent a lot of time talking about and you've been at the forefront of
• 15:34
  this is using LLM, so large language model like Claude Chat, GPT. And I believe you settle at the
• 15:41
  moment on Claude and I'm always hesitant to say that because give it six months, give
• 15:46
  it six weeks, give it six days.
• 15:49
  Don't talk too loud because it could change right now.
• 15:52
  Exactly.
• 15:53
  Claude heard you talking and goes, "I'm not working."
• 15:55
  Yeah, I'm done with you.
• 15:57
  Since I started the sentence, the one that's doing better is changing.
• 16:00
  So it's such a lively environment with so much evolution going on.
• 16:07
  the strengths of one, outweigh another right now, and then in two weeks it can change.
• 16:12
  But I know you've settled it at the moment, you're working with Claude.
• 16:16
  So how did that evolve?
• 16:19
  How did it go from just using it as a search tool to expanding into using it as a lot bigger
• 16:26
  tool and an assistant in your writing?
• 16:28
  When OpenAI first created ChatGPT, I was just immediately captivated with it.
• 16:33
  And I was like, "Okay, it's kind of a cool little toy."
• 16:37
  see what it could become. But at the same time, you know, I knew that it was creating a firestorm.
• 16:43
  There were a lot of people to either side of it. And I think that in the artistic community, you
• 16:48
  know, there was a lot of controversy around it because, you know, people were starting to use
• 16:53
  it to write, which, you know, I saw that and I was like, you know, that's just nothing against it.
• 16:59
  You can do with whatever you want to with it. But it just did not write very well. And in my
• 17:03
  In my opinion, it still does not write very well.
• 17:05
  But I'm like, okay, well, it's interesting.
• 17:08
  But then they created, I think they created Canvas, which was kind of like a work sandbox.
• 17:13
  You could create projects, that type of thing.
• 17:14
  I'm like, well, now you're starting to give me something that's a little more that I can
• 17:18
  work with because I started to see it more as being a useful assistant where I could
• 17:22
  put stuff and then start to develop ideas and start to develop ways that it could help
• 17:27
  me with my business.
• 17:29
  Claude came along shortly after that.
• 17:30
  I immediately could see the difference with Claude.
• 17:33
  It was more conversational even than ChatGPT was,
• 17:37
  but it was geared more towards business.
• 17:40
  So that's what I liked.
• 17:41
  It had a lot of business analytic tools
• 17:43
  built into it from the beginning.
• 17:45
  And what we're doing is you both know we are a business.
• 17:51
  People may hate to hear that.
• 17:52
  I'm a creative, you're a creative,
• 17:54
  but we are running a small business.
• 17:56
  Most businesses, no matter how big they are,
• 17:58
  have employees. Well, you know, we have a virtual assistant, a PA, you know, someone
• 18:03
  that helps us with this, someone that helps us with that. Some writers may not be to the
• 18:06
  point where they can do that. What I started to realize is this tool, especially Claude,
• 18:11
  could become that. You know, I could start to create a virtual assistant that, you know,
• 18:16
  was there 24/7 for whatever I needed, you know, to bounce ideas off of, to, you know,
• 18:20
  help me with analytics, to help me with whatever I might want to do around my business. The
• 18:28
  do a lot of words. Once I'm going, I can really get going. And I love to write. But what I
• 18:34
  hate doing is I hate marketing. I hate looking at anything spreadsheet related. Some people
• 18:40
  live for that. It makes my eyes glaze over. But I've realized that, you know.
• 18:46
  There's one person in this talk that likes to do that and the other two being you do
• 18:50
  not. So. Yeah. Love me a good spreadsheet.
• 18:53
  - Big, he gets that spread boy where he spreadsheets
• 18:58
  just go crazy.
• 19:00
  (laughing)
• 19:01
  He folds out the spreadsheet, he goes,
• 19:03
  "Look at it, it's got 300 lines."
• 19:05
  I'm like, yeah.
• 19:05
  - Oh, that would just be like, yeah.
• 19:08
  A little piece of me dies on the inside
• 19:10
  just looking at that.
• 19:11
  (laughing)
• 19:12
  But so as quality advanced,
• 19:16
  and it moved into here recently,
• 19:19
  agentic AI, which is agents that can actually
• 19:22
  operate and perform tasks on their own on your behalf, that's when I realized that, you know what,
• 19:28
  this is a game changer. Again, it's not going to really do anything for your writing as far as the
• 19:33
  actual writing, the creative process, but it freed up my time to where I can write more
• 19:38
  and I could create better books because of some of the tools that I was able to create.
• 19:42
  So, you know, I think that what we're entering is we are entering an age of personalized software
• 19:48
  that you build to do the things that you want it to do. You don't have to take something that
• 19:54
  someone else built, some big company, and then customize it and tear it down and figure out
• 19:58
  how to make it work for you. You can create it on the fly, something that is just yours, works
• 20:04
  just for the way that you want it to do and does your task to a T. And that's what I have realized
• 20:09
  that, you know, running this business, I can do so much more from a creative standpoint if half of
• 20:15
  my mind is not always focused on what did my Facebook ads to do? What did my AMS ads do?
• 20:21
  What's the market research for this genre? Is this going to work? Is this a trope that
• 20:25
  works? Does it not work? There's so many things that we can now do. And that's where I've
• 20:31
  dived into, clawed, and using the co-work and now code to create personalized software.
• 20:39
  So two questions I want to jump back to, Rolf and one, if we could talk about the difference.
• 20:43
  said agentic. So for some of us who are older and dumber, and I'm not going to point it, Nick, but you
• 20:48
  know, please don't pretend. No, yeah, it'd be rude. But could you give us a, you know, the quick
• 20:54
  version of what exactly is when we talk, what is an agent and what does the agentic mean when you
• 20:59
  say it like that? So
• 21:00
  so agents are individualized bits of programming that are specific to whatever task you assign them. So
• 21:13
  create an agent, which is just like in Claude,
• 21:16
  it's just a piece of software that's part of Claude's
• 21:20
  programming that I interact with that what they are trained
• 21:24
  on and what they know is everything that is marketing
• 21:26
  related.
• 21:27
  You know, if I want a spreadsheet specialist,
• 21:29
  I can create a spreadsheet specialist that that's all
• 21:32
  that they would do is sit there and dissect spreadsheet.
• 21:35
  - Then we wouldn't have a Nick Harvey, and if you said.
• 21:38
  - I could create my own Nick Harvey.
• 21:40
  (laughing)
• 21:41
  I just have him live on my little computer
• 21:44
  and he would just live there and do my bedding
• 21:47
  with a spreadsheet.
• 21:48
  - I don't know that Claude could be as snarky as his.
• 21:51
  - No, no, no.
• 21:53
  (laughing)
• 21:53
  That could be a good thing, I don't know.
• 21:54
  - It seems like pretty polite.
• 21:55
  He's like, oh, Claude's always like,
• 21:57
  "Oh, no, you're right, you're a good job."
• 21:59
  And Nick never tells me I'm right.
• 22:00
  - Yeah, they're good at, the LLMs are really good
• 22:03
  at apologizing.
• 22:05
  - And you know, there are ways around that too.
• 22:07
  I've gotten a reminder does not do all that anymore,
• 22:09
  But, you know, but again, you have to look at it as what is it that you need, you know,
• 22:15
  they are not going to be all things to everybody and they shouldn't be.
• 22:18
  Just like as an author, we don't want to be a master of everything and try everything
• 22:23
  that you hear just because it's the newest and greatest thing.
• 22:25
  It may not be applicable to you.
• 22:26
  Well, it's the same thing with LLMs and, you know, agents.
• 22:30
  You need to see what are your pain points and then you create something that solves
• 22:34
  that pain point.
• 22:35
  You know after we had that conversation a couple months ago
• 22:37
  We were we have I've kind of played with Claude a lot more and I've got to say Claude and I are becoming fast friends
• 22:43
  So we have a couple of you know headbutting times which feels on par for everything
• 22:48
  But but overall I got to some of your advice and was like let's see how I can work this so
• 22:53
  you know kind of as
• 22:55
  Somebody who's if I were stepping in as somebody I'm not really touched my toes into Claude much
• 23:00
  Can you kind of give me an idea of basics of what if somebody might want to use it how they might want to set up that?
• 23:05
  in the projects and we talked about projects and agents and all the fun
• 23:10
  things that are in there. So if someone is new to you know an LLM you know an LLM
• 23:15
  has a basic chat box just like a text messaging box that you can type into it
• 23:20
  a query a question whatever you want and start a conversation with that LLM you
• 23:25
  know they all start with that as a basic function and I think that you know
• 23:29
  people you know maybe are intimidated by that because they think well I don't
• 23:33
  know what do I say to it? Well, you would say the same thing to it.
• 23:36
  Do I need to know how to code?
• 23:38
  Yeah, do I have to know how to code? You don't have to know any of that.
• 23:40
  Basically, it's like you've got another person sitting across from you.
• 23:43
  You ask it the same thing. So you just talk to it to start off with.
• 23:47
  You know, you can introduce yourself and say, Hey, I'm Doug the author.
• 23:51
  Here's what I write. What do you think about that?
• 23:53
  And you just have a conversation back and forth.
• 23:55
  And then what you can do is you say, this is my business.
• 23:59
  You know, I am an independent author. I publish on Amazon.
• 24:03
  I have audio, whatever aspects of your business that you have,
• 24:07
  you tell it about that.
• 24:08
  And it's absorbing all of that.
• 24:11
  The more that you talk to them, the more they get to know you.
• 24:13
  And I don't want to say their personalities,
• 24:16
  but they will alter their actions in accordance
• 24:19
  to how they're interacting with the person driving it.
• 24:22
  So you tell it about your business, you tell it everything.
• 24:25
  And then you tell it, ask me questions
• 24:28
  that will help you understand what I do and what pain
• 24:31
  points I may run into.
• 24:32
  And it'll just start asking you, how do you do this?
• 24:34
  How do you do that?
• 24:35
  What do you think about this?
• 24:35
  What's important to you?
• 24:36
  What do you want in five years?
• 24:38
  You tell it all these things
• 24:39
  and it will start to understand what you're doing
• 24:42
  and where you want to go and how you wanna grow
• 24:44
  as an author.
• 24:45
  And you just kinda go from there and start building it.
• 24:48
  And then once you're comfortable with that,
• 24:51
  you start to say, this is what I don't understand
• 24:54
  about my business or what I don't like to have to deal with.
• 24:58
  And it will tell you so much
• 25:00
  that you've never thought of outside the box
• 25:02
  about dealing with that.
• 25:04
  And from there, you can start saying, okay,
• 25:06
  I wanna start building my own personalized apps
• 25:09
  or my own software to help me fix some of this.
• 25:13
  You know, and in my case,
• 25:15
  I have built at this point a stack of apps.
• 25:18
  You know, I've here built two websites
• 25:21
  that are, you know, for my business,
• 25:24
  for both of my pen names,
• 25:25
  and I built them, hosted them,
• 25:28
  They're up live on the internet,
• 25:30
  all in a matter of about a half an hour from scratch,
• 25:33
  without understanding how to code.
• 25:34
  I don't understand a word of code.
• 25:36
  I've never done it before.
• 25:38
  I was able to get them up and they look good too,
• 25:40
  because the one thing that I thought about was,
• 25:42
  what is it that is coming down the road?
• 25:46
  And we all know people talk about discoverability.
• 25:48
  That's always been an issue.
• 25:49
  As more and more people start to talk to their AIs
• 25:53
  and asking for recommendations,
• 25:54
  there's a lot of conversations about,
• 25:56
  I just read the greatest book.
• 25:58
  I just read "Forthwing," what's another one that you recommend that's just like that?
• 26:01
  I need another book like that.
• 26:03
  Those AI agents are not operating like the old Google.
• 26:06
  They're not doing SEO.
• 26:08
  They have their own way of interacting with what's on the web and returning information
• 26:12
  in a conversational way.
• 26:14
  We want to be part of that conversation.
• 26:16
  The reason that I built a new website was because I thought, "Okay, you can take an existing
• 26:20
  website retrofitted and jimmy tree rig everything into it to make it so that it works with AI
• 26:27
  discoverability, but then I thought, well, why can't I just build one from scratch? So I built
• 26:32
  my author Ember Scott website, let's say, from the ground up, and I told the AI, I said, I want to
• 26:37
  build it with AI discoverability in mind. So all those little tools that the AI does for their
• 26:44
  GEO, Generative Engine Optimization versus SEO, I said, I want them built in from the ground up.
• 26:49
  And it did it. It built it all in, it built in all the stuff that the AI's read on the back end,
• 26:54
  the way that they crawl the web and it gave it access to AI crawlers.
• 26:59
  So you couldn't do that before.
• 27:01
  You would have to pay a developer a considerable amount of money.
• 27:05
  It would take a lot of time,
• 27:06
  a lot of testing even if that developer used WordPress plugins,
• 27:11
  they still have to take time and it's going to cost you money.
• 27:13
  I don't know anything about that.
• 27:15
  So I'd build it from scratch.
• 27:17
  It did all of it as code.
• 27:19
  It's not plugins, it's all code,
• 27:20
  it's all stable and it took a half an hour to do it.
• 27:24
  and I got it up and hosted it live.
• 27:27
  And the total cost was,
• 27:29
  I think it was $21, it cost me,
• 27:31
  like $12 for the domain.
• 27:33
  And then I spent another,
• 27:35
  yeah, the domain and the registration,
• 27:36
  and another $9 in credits to build this with,
• 27:40
  using API on the, with an LLM.
• 27:43
  And that was it.
• 27:44
  And there's no hosting fees.
• 27:46
  It's literally zero.
• 27:48
  It cost me nothing per month
• 27:49
  to have this up there and live.
• 27:51
  - So let's use an example here of,
• 27:53
  Because if someone hasn't got involved in this, there's a lot of stuff that's being
• 27:58
  talked about that's probably quite daunting.
• 28:02
  And misinformation.
• 28:03
  There's a lot of misinformation out there.
• 28:05
  There is, of course.
• 28:06
  But let's run through a process here and let's use the writing room as an example, because
• 28:10
  this is a great, a great and useful tool.
• 28:13
  And you explained it to us and helped us get going with it.
• 28:17
  So I want to build what you call the writing room, which is going to be a group of agents
• 28:23
  who are gonna help me with my book,
• 28:25
  both from editing, tracking, keeping the voice the same
• 28:29
  and all these kind of things.
• 28:30
  So I'm gonna do it in Claude.
• 28:32
  I'm gonna start with what they call a project, right?
• 28:35
  Which basically puts a fence around everything within that.
• 28:40
  So the memory can source everything
• 28:42
  that's within that project.
• 28:45
  - And I gotta say, when you told me to do projects,
• 28:47
  that's been a game changer for me.
• 28:49
  So like I wasn't doing that before.
• 28:51
  I was just being wild and nuts,
• 28:54
  throwing it out there in the chat and just letting it go.
• 28:56
  - Yeah, so to kind of be clear about this,
• 28:58
  so if you just keep using the chat
• 29:01
  and keep asking things and different things,
• 29:03
  eventually it's gonna accumulate so many of these chats,
• 29:06
  it won't reference them all and link them all together.
• 29:08
  So you put them in a, am I correct?
• 29:10
  Is that- - Correct, yes.
• 29:12
  There's just so much, it will continue.
• 29:14
  Your conversation can get so long
• 29:16
  that it will lose the thread.
• 29:17
  It can remember like maybe the last 10, 15 things
• 29:20
  talked about after a while. But if you've got like, you know, a couple of million, which
• 29:24
  you'd be surprised if you've got a couple of million words in there, just for your conversations,
• 29:28
  it's not going to be able to find what you talked about 10 days ago, 15 days ago, and
• 29:33
  it will start to muddy up everything and it'll make stuff up. And that's where people get
• 29:35
  the hallucinations from because it's trying to say, okay, you need to know this. I don't
• 29:42
  remember that. So I'm just going to make it up. Here's what I think it must have been.
• 29:45
  But projects, if you create that, you know, like you said, it's a sandbox, it's a set
• 29:50
  of files within a project, and it will be able to remember everything within there that
• 29:55
  you're talking about, because that's all that it's focused on.
• 29:59
  Right.
• 29:59
  There's a cap to that we can talk about later where the memory gets in.
• 30:02
  There's words to work around that.
• 30:03
  But let's keep going on the basics here.
• 30:06
  So we're going to start a project.
• 30:07
  I want to have a writing room.
• 30:08
  I want to build an agent.
• 30:09
  I'm sitting at the keyboard.
• 30:11
  What am I going to do?
• 30:11
  I want to copy editor.
• 30:12
  Let's start with a copy editor as an agent.
• 30:14
  What do I do?
• 30:15
  Right. So the first thing that you're going to do is you talk to the LLM in the project about
• 30:21
  what you're doing. I'm writing a thriller. This is the genre. These are the tropes that I have in.
• 30:26
  This is the audience that I am trying to reach. This is the ideal reader. And so from there,
• 30:31
  you ask it. You said, "Now, I need to build a writer's room." And the idea for this came about
• 30:35
  for me after I saw, I think it was a TED talk or YouTube with Shonda Rhimes discussing how she
• 30:41
  creates these amazing shows that she does and the writer's room that she builds because
• 30:45
  you know she's not doing everything in a vacuum by herself. She has specialists in
• 30:49
  all areas and they all work together. So I'm like, I want the same thing. So I'm like,
• 30:53
  Claude, can you recreate a writer's room in digital format for me to use using agents?
• 31:00
  So I created eight agents for my thrillers. You know, I've got like the story architect,
• 31:04
  the character psychologist, an editor who does developmental and line editing. All of
• 31:11
  work together and each one has that function. So imagine you've got a table
• 31:16
  with eight people sitting around it, each one assigned one of those functions.
• 31:20
  That's what I created. So when I'm writing, I write in Scrivner, I finish a
• 31:26
  chapter, copy that chapter, I put it into my project and say, "Here's chapter one.
• 31:31
  What does everybody think?" And then what happens is each one of those agents goes
• 31:35
  through that chapter that I just gave it and gives me their feedback that's
• 31:39
  relevant to that chapter. So the character psychologist will be like your main character
• 31:43
  is not acting the way that you've established for the past three or four chapters. They drifted
• 31:48
  here. This is out of their character. The lore and continuity agent will say, "You said
• 31:54
  that their eyes were blue in this chapter, but in the beginning you said that their eyes
• 31:58
  were green." So they catch things like that. And then they put all that into a storyboard
• 32:03
  so that I've got a running graph of everything that's happening past, present, and future,
• 32:09
  weighed against the outline that I put into the writer's room as well.
• 32:13
  So everything starts with an outline.
• 32:15
  And then once they don't change anything, so they don't change any of my writing, they
• 32:19
  will just give me their output about what they think based on what they've seen and
• 32:24
  encouraged.
• 32:25
  And here's possible things that are maybe a mistake.
• 32:27
  You spelled this word wrong.
• 32:28
  You use the word wrong word here.
• 32:31
  then I can go back, rewrite my chapter based on whatever
• 32:35
  there is that they may have given me.
• 32:36
  Either I can accept it or not accept it
• 32:38
  and just keep on doing my thing.
• 32:39
  Then I give them the final chapter back again a second time
• 32:41
  and say, this is the final chapter, I'm done.
• 32:44
  They archive it all and it becomes part of their memory
• 32:46
  so they can always go back and reference it moving forward
• 32:49
  when I do the next chapter.
• 32:51
  - Right, brilliant, brilliant, brilliant.
• 32:52
  So when you're setting up these agents,
• 32:55
  so again, it's sort of like you're sitting down and you go,
• 32:58
  okay, I want an agent, I'm gonna have this copy agent
• 33:00
  continuity agent, you don't have to code a darn thing. You just sit there and talk to
• 33:05
  Claude about it and tell it what you want, right? And you give it as much context as
• 33:09
  possible. And this is a bit of a, there's one of the things I learned when I started
• 33:14
  doing it. It's chicken and egg, right? I don't really know exactly what I want, but I'm
• 33:18
  going to ask it for some stuff. And then I'm going to ask it, what else should I ask for?
• 33:21
  That's always nice. Yeah, that's really great. Yeah. That's a good answer. That's really
• 33:25
  good.
• 33:25
  Yeah, that's the beauty of it. It's like, you don't have to know how to code. You're
• 33:30
  you're literally just asking it and taking its advice
• 33:32
  and saying, hey, what kind of agents
• 33:34
  would be optimal for this story?
• 33:36
  Do I need to create this book?
• 33:38
  And it will tell you and then you're like,
• 33:40
  okay, I like this one, this one, this one,
• 33:41
  I don't really need that one.
• 33:43
  Spin those up and it will create them.
• 33:45
  And they're all sitting there silently around your table.
• 33:48
  Like let's say your table is the project
• 33:50
  and as you write and put something into it,
• 33:52
  they go to work on it and then they give you their feedback.
• 33:55
  And that's all that that is.
• 33:56
  So, you know, it's great for developmental editing.
• 33:59
  It's great for line editing for me.
• 34:00
  And everybody's got to develop it themselves
• 34:02
  and get to where it works.
• 34:03
  But I've got a couple of things that I've kind of like,
• 34:06
  as these models get smarter,
• 34:10
  they introduce a new level, a new version of them.
• 34:13
  And it seemed like they're coming out constantly
• 34:14
  with new versions.
• 34:16
  They are really leveling up what they can do.
• 34:19
  And because I have been working with them,
• 34:21
  doing the same thing for so long,
• 34:23
  even when the versions change,
• 34:25
  mine don't typically revert back to being,
• 34:27
  what they once were or give me a bunch of gobbledygook or anything like that, they're just really
• 34:31
  good about staying on task. They just get better at it.
• 34:34
  Got you. Got you. And there's a section called Instructions. I think it's labeled Instructions,
• 34:40
  isn't it? That's where... So if you're wondering where these agents go, what you do is you
• 34:45
  and correct me if I'm wrong on any of this stuff, but this is how I remember, I did it
• 34:49
  when I built them, is you get this chat back and forth and go, "Get all this information."
• 34:54
  you go, "Okay, this is the agent I want to build.
• 34:57
  I want to call it such and such.
• 34:58
  Give me the text that I need to put in the instructions."
• 35:00
  It gives you all that stuff.
• 35:02
  And you can do it all in, I did it sort of continuously
• 35:05
  until I said, "Okay, here's all the agents that I want.
• 35:08
  Here's what I want them to do.
• 35:10
  Give me the text."
• 35:11
  And then you just cut it and paste it
• 35:12
  and put it in the instructions
• 35:14
  where you can go and mess with it later.
• 35:16
  So add to them, one of them that you don't like,
• 35:19
  you can like update it.
• 35:20
  There's the instructions, there's the knowledge base,
• 35:22
  all of these things govern what's happening within that project.
• 35:26
  So if you've got an outline of your entire book,
• 35:28
  you can put it there and they will always reference that outline.
• 35:31
  If you stray from the outline,
• 35:33
  which I do all the time,
• 35:34
  and they will say,
• 35:36
  that's not what happens in Chapter 6, you've changed it.
• 35:40
  I will say, no, I'm venturing from the outline because of X,
• 35:43
  Y, and Z, they're like, "Okay, that's fine."
• 35:45
  They will on the fly adjust what's going to happen,
• 35:48
  what's been happening so that they know moving forward.
• 35:50
  Okay, this new adjustment is now what's Canon.
• 35:54
  This is what we're gonna work from.
• 35:55
  - Yeah, they'll use that word a lot, Canon.
• 35:57
  Yeah, it's like that's set in stone.
• 36:00
  Yeah, mine's got to the point now where Tracy is my outliner.
• 36:03
  She traces the outline when I talk to Tracy,
• 36:06
  gets to the end and she goes,
• 36:07
  "Okay, what do you wanna do now?"
• 36:08
  She's given up on the original outliner.
• 36:11
  She knows not to even mess with it.
• 36:13
  She goes, "Okay, what do we do now?"
• 36:14
  - Yeah, mine, it's so funny because, you know,
• 36:17
  they will have discussions amongst themselves
• 36:20
  will give me their findings. And I have it set up to where if there is a strong disagreement
• 36:25
  about something, then the lead story architect will come forward and say, hey, we think this
• 36:31
  is an issue. We don't agree on something. And then whatever I tell them becomes that's
• 36:35
  the fact. That's the fact though. This is where we are. And they accept it. But they've
• 36:39
  gone, they will ask me questions. I'll be like, Hey, why did you use these disarrangement
• 36:44
  of words instead of this, which is what you said to normally do? You know, so I have to
• 36:48
  explain that to them and they're like, okay, got it. That makes sense. It's now in our memory and then they move on to the next
• 36:53
  little thing. But it really is just like, you know, having a room full of people that you've corralled. And if there is something that
• 36:59
  you present and throw up on the whiteboard that they, that's part of their specialty to address them, they'll address it. So that's
• 37:06
  how I use it.
• 37:07
  And so like, it's great for the, because you put this all in one project. So I think you want to tell me that each book, like as
• 37:13
  I'm right. So I set my own project up. I'm writing my third book in my J. Delp series right now. And so I set my project up. I uploaded books one and two and my story Bible. And that way it's been great because if I go, well, did I do this already?
• 37:27
  Nick can go. Did I already put a character in it? Because I'm real bad character names, the worst, right? How many times I got to come up with a new character name. Did I use that again?
• 37:36
  Usually my wife goes she'll finish a book that you've used a man to like three times. So Claude's been great about no
• 37:41
  This is where your characters are. This is the name of these characters like some very side
• 37:46
  kids of side characters or something like that and it's been it's been a game changer as far as time like me having to
• 37:51
  Go back and search for it would have been some of the stuff is things
• 37:54
  I didn't even think to put in the story Bible, you know, I didn't think about putting that story about let me go find that
• 37:59
  Now let's add it to the story Bible. Yeah, and you know the good thing is too, you know for me
• 38:04
  It's helping me to create a cleaner first draft, a cleaner finished product, which then goes on to my editors.
• 38:11
  But it's just so much quicker because, you know, and I think that the books are better too because, you know,
• 38:17
  I've added a pacing and attention graph to, you know, what I have it look at too as well,
• 38:24
  so that I can actually see a visual plot line of the points.
• 38:27
  You know, it'll say, okay, if I wanted to compare this to Save the Cat or the 4x Structure, Heroes Journey,
• 38:33
  It will plot each chapter on that graph and tell me where it falls in relation to where it should be
• 38:41
  Based upon the the story structure that I'm using so if I see it'll flag me and it'll tell me hey
• 38:46
  You've got three chapters or hasn't that have kind of been developmental
• 38:49
  But you need to have some action here then I know that before you know
• 38:53
  Yes, it tells me ahead of time or you know
• 38:55
  It will plot out each chapter that I've done and say okay chapter eight
• 38:59
  Maybe you might want to ramp up this or dial down this let the let the reader breathe for a bit
• 39:04
  And I can see that that tension is really high really high really high
• 39:07
  And then it'll say maybe flatten it a little bit for this one so I can go back and retroactively do that or apply it moving forward
• 39:13
  I love the graph idea there
• 39:15
  And I just want to point out that that is a subconscious love of spreadsheets the fact that you
• 39:25
  How do you make a graph?
• 39:27
  You put the data in a spreadsheet.
• 39:28
  You get some spreadsheet nerd
• 39:30
  to come in and do your little spreadsheet, your graph for you.
• 39:32
  Then you can look at your graph and go,
• 39:34
  ah, pie chart, I'm good with that.
• 39:36
  So.
• 39:37
  - You would be amazed when you see it, you know,
• 39:39
  like as a graph that's, you know, like building tension
• 39:41
  and it's arching up, you know, the way that you want it to,
• 39:44
  you know, you get that sawtooth pattern,
• 39:46
  which is kind of how a thriller should run
• 39:48
  from a pacing standpoint.
• 39:49
  And you can see visually where you're flailing
• 39:52
  and you can kind of go back and look at that and be like,
• 39:54
  You know what, I do need to add some action here
• 39:57
  and pick it back up and you know.
• 39:59
  - Yeah, that's a, I did it not nearly that good.
• 40:02
  I'm gonna go back and change my instructions,
• 40:04
  but I did take it when I put the first two books in,
• 40:06
  I said, I want you to give me a pacing outline.
• 40:08
  I don't want you to tell me what the story,
• 40:10
  I don't want you to write the story.
• 40:11
  I just want you to tell me based on what I've written before,
• 40:13
  how do I, how do I pace the books
• 40:15
  so that I know this is where this falls,
• 40:17
  this is where that falls.
• 40:18
  And this one's a little more mystery oriented
• 40:20
  than action for me.
• 40:22
  So that always requires me to keep the clues in line.
• 40:25
  So I'm really bad about that.
• 40:27
  - Yeah, tracking those red herrings.
• 40:29
  - Tracking those red herrings,
• 40:30
  make sure I got them in there.
• 40:31
  And so, but yeah, no, I thought that so, that's great.
• 40:34
  I like this story graph idea.
• 40:35
  I'm gonna totally be playing with Claude this afternoon.
• 40:38
  You know, instead of writing.
• 40:39
  - Yeah, we jump off this and dive into fixing our writing.
• 40:44
  - Totally aware, like at this point,
• 40:46
  my writing day is shot, I'm gonna be playing in Claude.
• 40:48
  So.
• 40:50
  So one of the things that I remember you talking to us about was getting it to learn your voice.
• 40:55
  And what was the process you used to get it to learn your author voice?
• 40:59
  So the reason that I wanted to know my author voice is so that as I'm writing,
• 41:04
  I want to make sure that I don't lose, that I'm not writing something that maybe doesn't kind of gel with things that I've said before or my style.
• 41:12
  And the way that I do that is, I've discovered that you can actually extract your writing DNA
• 41:18
  by giving it samples of something that you've already written. So I would feed it like three
• 41:22
  chapters where Mike is, I've given it entire books and I said, I want you to extract a voice DNA.
• 41:28
  And it knows what a voice DNA is. And he said, but you have to tell it apply it to
• 41:33
  an author who writes. So their style, their tone, your voice becomes the DNA that it builds. And
• 41:40
  what it will do is it will break down, it'll ask you a bunch of questions based on what you feed it
• 41:45
  in the way that you answered it will then go through and analyze the sample that you gave.
• 41:49
  And you've got to give it at least like I say three chapters of your own writing, preferably
• 41:54
  more. And then what it will do is it will actually map out your personal tone, your style, how you
• 42:00
  write, how you think. And then it will apply that to the agents as well. So they kind of know what
• 42:05
  to expect from you. And if they read something and they're like, this doesn't sound like you,
• 42:10
  they'll say this is why. The DNA extraction is really, you know, key for building a better book,
• 42:17
  as far as I'm concerned, because again, it's not going to write for you, but it can definitely help
• 42:21
  you to write a much better book. I mean, the last three books that I've written using the writer's room,
• 42:26
  I mean, the reviews just started to come in, you know, immediately, and they're just like, you know,
• 42:31
  we love this, love that. And it's just more of giving readers what they come to expect,
• 42:37
  but amping it up and I can do it faster now.
• 42:41
  And now a quick word from this episode sponsor.
• 42:44
  Jamie Davis here again, show manager for Author Nation.
• 42:47
  I help run all the logistics at the conference in November and I hope to see you there.
• 42:51
  In the meantime, did you know why book store owners are so organized?
• 42:56
  You should. It's shelf explanatory.
• 43:00
  And now back to the show.
• 43:02
  What doesn't it do well?
• 43:05
  Well, I can say it doesn't write well.
• 43:08
  Oh, yeah.
• 43:08
  I'll just do it right there off the bat, which is not a bad thing.
• 43:12
  No, that's not a bad thing.
• 43:13
  But, you know, again, it just, it has its own voice.
• 43:16
  If you look at, if you read a lot and if you're, especially some of the stuff
• 43:20
  that I've seen, even in KU, you know, the explosion of books in KU, you can tell
• 43:25
  what is written by a human and what is not because, you know, it doesn't do that.
• 43:29
  And another thing that it doesn't do.
• 43:30
  They tried to tell not show, I think.
• 43:33
  I say that a lot.
• 43:33
  It does the complete opposite of what we're supposed to do. It tells not shows.
• 43:37
  And another thing that it doesn't do is it cannot keep a secret to save its life. If you,
• 43:42
  if your outline tells the ending of the book what's going to happen, at some point those agents
• 43:48
  are going to be like, well, why don't you put in here that so-and-so and make it very dramatic that
• 43:52
  so-and-so saw the gun or so-and-so's fingerprints are going to be found on the gun. I'm like,
• 43:57
  no, I don't want to foreshadow that. It takes foreshadowing to a whole new level. It's like
• 44:01
  It would wrap up the book in three chapters if it could.
• 44:05
  So that's another reason why I'm like, I can't, you know, I don't want it to write.
• 44:08
  I don't want it to rewrite anything that I do.
• 44:10
  But what it does well from an analytic standpoint, it does amazingly well.
• 44:15
  And that's what our writing is.
• 44:17
  Our writing is code.
• 44:18
  And these LLMs understand code better than anything, you know, that I've ever seen.
• 44:23
  And they just dissect it immediately.
• 44:25
  So once you learn that, and once they learn that, it's amazing what can happen.
• 44:30
  One of the things that I found is that we talk about hallucinating.
• 44:35
  You'll get them comment on something like, "Oh, well, you said only two people know that,
• 44:40
  but three people do."
• 44:41
  I'm like, "No."
• 44:42
  Remember, because we talked about that and we're not going to tell the Sergeant at the
• 44:46
  station for this reason, "Oh, you're right.
• 44:49
  Yep.
• 44:49
  I read back through and I'm sorry."
• 44:51
  Yeah, it scans, right?
• 44:52
  It has this interesting method that I don't even know how to describe it, but it doesn't
• 44:57
  read, like we think we'd say, "Hey, read through this." It doesn't read it as we comprehend
• 45:03
  reading. It scans things and pulls pieces.
• 45:08
  It skims it almost, doesn't it? Yeah.
• 45:10
  It's looking for what it considers to be an important factor detail based upon what you've
• 45:15
  asked it. So if you put your entire book in, it's not reading it word for word, because
• 45:18
  if you think about it, the token usage and an 80,000 worth, you would like burn through
• 45:24
  So what it's doing is it's grouping them together and going through those 80,000 words and seeing patterns and recognitions and things that you know
• 45:32
  you have
• 45:34
  Maybe repeat it that are un that are important and it starts to catalog those so it may seem like it reads everything
• 45:40
  But it really does not also it does remember it's gotten better like that Claude and Chad
• 45:45
  I think both have just came out with we are remembering everything now
• 45:48
  So a little bit about memory like to me
• 45:50
  I always thought it was just gonna be in there
• 45:52
  So it does have limits in each chat even that it can do before it starts to run off
• 45:59
  You want to kind of tell us a little bit like explain that to me like I'm an idiot
• 46:06
  If your imagination will stretch that if you can figure it out like
• 46:12
  And you know I think a lot of people need to realize too that one of the first things you can do is when you start working with them
• 46:17
  go into the settings for the LLM. There are ways to turn the memory on so that it remembers
• 46:24
  across chats. It can remember across projects, but you have to turn certain aspects on. It's
• 46:30
  been so long. I don't know if they've maybe built that in, but in the beginning, they did not have
• 46:33
  any of that built in. You had to go in and turn on memory so that it could remember a lot more.
• 46:38
  It wasn't automatically on two months ago because when you told me about it, make sure this is on.
• 46:43
  One of them was, one of them wasn't.
• 46:45
  Yeah, I had the same the same issue so I haven't looked again but yeah. Yeah, so you turn the memory on but you know basically again you know it's going to run out of memory you know these new ones have like a one million token.
• 46:57
  Context window so think about that a million tokens is probably around 800,000 words so that's a lot but even if you really max that out it's still going to have trouble remembering.
• 47:09
  All of that stuff in the beginning of 800,000 words to the end.
• 47:13
  So what I have found when I'm writing these, you know, really long chapters and I put them in and I get to where I've got like, okay,
• 47:20
  there's been 10 chapters.
• 47:22
  You know, I know that it's going to start to run into that memory issue where it can start to confuse things.
• 47:27
  So what I will tell it is, and I've got my setup to where it'll give me, it'll tell me ahead of time, it's like, hey, this is starting to get a little bit long.
• 47:33
  Let's start a new thread within that project.
• 47:36
  So I will tell it to give me a handoff document.
• 47:38
  So that that way, it'll create a document that says, "This is what's happened before based upon."
• 47:43
  It's still referencing the outline, the storyboard, everything. Here's what's happened before. Here's
• 47:49
  where we are. So that when I start a new thread, I don't have to go back and tell it anything.
• 47:54
  It just picks up where we left off.
• 47:56
  Do you just add that in with the other files in the project then?
• 48:00
  I don't put it into the files. When it gives me a handoff document and I start a new thread,
• 48:04
  I just put it into that chat. That's the first thing we talk about. I'm like, "Hey,
• 48:07
  Okay, we're starting, you know, I'm going to be putting in a new chapter soon, but here's everything
• 48:12
  up to this point that we've discussed in all the chapters and it just has it all there.
• 48:16
  And that's where that living storyboard comes in because it's doing a lot of the heavy lifting.
• 48:20
  And I've noticed that since I've gotten that storyboard, you know, more powerful and more
• 48:25
  up and running and the way that it uses it, it doesn't need as much of a handoff because the
• 48:30
  storyboard lives there permanently. And it just will reference that. So it doesn't take as much of
• 48:34
  a large handoff to start a new thread.
• 48:37
  And each thread, I will put 10,
• 48:39
  12, 15 chapters in it.
• 48:41
  And then it'll start, it'll tell me,
• 48:43
  like, hey, before we do another one,
• 48:44
  before we put another chapter in,
• 48:46
  you might want to think about starting a new thread with me.
• 48:49
  - That's interesting, 'cause I asked it, okay,
• 48:52
  I was worried about it.
• 48:53
  And it said, yeah, about every 10 or 12 chapters,
• 48:55
  we should do a consolidation, a handoff.
• 48:57
  And it had me put it in the instructions,
• 48:59
  'cause I asked it where to put it.
• 49:00
  And it told me to put it in the instructions.
• 49:02
  I just actually, I'm in the midst of doing it.
• 49:05
  The last thing I did before we started recording shows earlier today was get a second hand
• 49:11
  off document from chapter, I think it's 13 onto 22.
• 49:15
  And how I knew that it was starting to run out of context, because I asked it, "Can
• 49:19
  you give me a warning?"
• 49:20
  And it told me it couldn't.
• 49:22
  So I should ask that question again, because as they continually update, there's things
• 49:28
  that it can do that it didn't know it could do before.
• 49:31
  But I asked it, "Hey, are we okay?
• 49:32
  Do we need to?
• 49:33
  We're starting to run out of memory."
• 49:35
  And it told me, "No, no, we're fine.
• 49:36
  We're good, and we've got the stuff through Chapter 12."
• 49:38
  And I read through them, like, "You just said Chapter 12."
• 49:40
  I'm like, "The red flag that we do actually have a problem is the fact that you don't
• 49:44
  realize we're on Chapter 22 right now."
• 49:46
  And I went, "Oh, yeah.
• 49:48
  Busted.
• 49:49
  Sorry."
• 49:50
  I asked it yesterday.
• 49:51
  I was trying to remember if I had a side character.
• 49:54
  I was like, "What kind of card did I have in like one of the last books?"
• 49:57
  And so I asked it, "Oh, no, you didn't have any of that."
• 49:59
  I was like, "Are you sure?"
• 50:00
  I was like, I really thought I gave him like a suburban.
• 50:02
  He says, well, let me check book two.
• 50:04
  No, no, it was a Chevy Tahoe.
• 50:06
  I'm like, why didn't you look and tell me that?
• 50:08
  And Claude said, you know, if you had asked me,
• 50:12
  search Chevy Tahoe, I would have found it.
• 50:14
  I was like, if I had to ask you,
• 50:15
  how to search Chevy Tahoe.
• 50:15
  - If I had the answer.
• 50:17
  - I wouldn't have wanted to ask you.
• 50:20
  - To which of course I love his, oh, fair point.
• 50:25
  - Yeah, and then they still have a long ways to go,
• 50:27
  But for what we do for our business,
• 50:30
  people have to realize too that the LLMs,
• 50:32
  they were not created to do creative work.
• 50:37
  They were there to do programming.
• 50:39
  They're there to help businesses and things like that.
• 50:41
  And what we've done is adapted it to our needs.
• 50:43
  Because again, it comes all the way back to,
• 50:45
  we are running a business.
• 50:47
  It's just a matter of finding your pain point
• 50:49
  and fixing your pain point.
• 50:50
  And I've created now so many stacks of stuff.
• 50:55
  I didn't even know about Netlify and GitHub and some of these other places where it taught
• 51:01
  me how to get my information and my apps that I wanted to build up and running to where
• 51:08
  they're like websites online.
• 51:09
  I can use them for anything.
• 51:11
  And the good thing is they're personal, they're private on my computer.
• 51:15
  So it's like having my own personal application that I can access from my phone, my computer
• 51:19
  anywhere, just like if I wanted to go to a website because it is a website, but it's
• 51:23
  only one that only I can see what goes into it.
• 51:25
  I can take my entire KDP dashboard, which I've done, a lifetime dashboard and drop it
• 51:30
  in and say, create these reports for me and it will create apps that will do certain reports.
• 51:35
  It does my read through rates for me.
• 51:38
  It has gone through and gotten to the level where it will tell me the best day of the week
• 51:42
  for me to launch a new book on by genre, by from looking at, you know, 10 years of who
• 51:47
  buys on what days.
• 51:49
  It'll look at my AMS spend.
• 51:51
  What do you, I say, you got to find all that data, just your data then, not just in general.
• 51:55
  - Yeah, no, I just take the KDP dashboard.
• 51:57
  Like, you know, if you can tell,
• 51:58
  you can download the entire report.
• 52:00
  I'll just download the whole thing and just drop it in
• 52:02
  and it will slice and dice it and give me an app
• 52:05
  that will give me what I want to
• 52:07
  and mine that information.
• 52:08
  - This is brilliant, man.
• 52:10
  - I was gonna say,
• 52:11
  - We could keep going.
• 52:12
  - This is going to be a week of my not writing
• 52:15
  while I build apps for me to do other things.
• 52:17
  - We just touched on so many different things,
• 52:19
  but hopefully it's given some folks out there
• 52:21
  some insights on how to use this tool.
• 52:24
  if they want to go down, everyone has to choose their own path. But if they want to go down,
• 52:28
  it hears some ways to dip a toe and get involved. And there's so much more to it in all the different
• 52:34
  business aspects. But all right, we better bring it to a close. We've got a couple of questions to
• 52:39
  finish up here. And one which we ask everybody. So let's switch it back to, because you're actually
• 52:44
  an author. We did start on talking about that as an author. What do you think your author's super
• 52:49
  of Paris.
• 52:51
  I'm not afraid to pivot or to try new things.
• 52:54
  I think that so many authors get, you know, not pigeonholed, but they say they're like,
• 52:59
  they're in a rut.
• 53:00
  And they're afraid to step outside of that, even though they admit that they're in a rut,
• 53:03
  because they're like, this is safe.
• 53:05
  This is what I know.
• 53:06
  I know my audience here.
• 53:08
  I know what to expect, even though what they may expect is completely fallen, you know,
• 53:13
  all to the wayside.
• 53:14
  But they're like, it's better than, you know, venturing into the unknown.
• 53:18
  I don't want to switch genres.
• 53:19
  I don't want to deal with building a new pen name.
• 53:22
  Whereas I am like, you know what?
• 53:23
  I see that as a fun challenge.
• 53:25
  I'm like, I'm not the least bit afraid.
• 53:27
  If it works, it works.
• 53:27
  If it doesn't, the worst thing that happens is
• 53:29
  I wrote a book that didn't sell.
• 53:31
  The upside is--
• 53:31
  - We've all done that at some point.
• 53:33
  - Yeah, exactly.
• 53:33
  The upside is it could be a lot of upside.
• 53:37
  So I'm just not afraid to try that.
• 53:40
  It's an 80,000 word book.
• 53:41
  If I don't like, I'll write another one.
• 53:43
  I'll go back to what I was going to.
• 53:44
  'Cause you can always go back,
• 53:45
  but you should always be willing to move forward as well.
• 53:47
  Great. Yeah, that's good. Good point.
• 53:50
  All right, let's wrap it up with a spin spin wheel of questions.
• 53:59
  He's got it over lubricated just so you know, so it will spin for eternity.
• 54:05
  I slow it down.
• 54:06
  That's right. Okay.
• 54:08
  What is a skill or maybe a hobby or something that you possess or do that?
• 54:13
  Most people would be maybe surprised or don't know about you.
• 54:17
  A skill or a hobby?
• 54:19
  That is a good one.
• 54:20
  You know, I have really gotten into, as I've gotten older, I discovered the benefits of
• 54:27
  stretching.
• 54:29
  Oh yeah.
• 54:29
  So, I have really decided that, you know, it's amazing what hot yoga can do for you.
• 54:35
  And it's just, I don't have the aches and the pains, you know.
• 54:40
  It used to be you get up, your back is stiff, you know.
• 54:42
  I'm just like, the older I get, my joints all hurt or fall apart.
• 54:46
  But there's something about stretching yourself into an unnatural position in a really hot
• 54:51
  room covered in sweat and that just makes you, I don't know, limber for hours and hours
• 54:57
  and it feels so good.
• 54:59
  Doug said the same thing on a trip in Singapore, actually.
• 55:02
  Yeah.
• 55:04
  Came back with the same report.
• 55:06
  I told me to go to the sauna.
• 55:07
  I was like, let me check it out.
• 55:08
  I was impressed.
• 55:10
  Four hours later, you're like, okay, yeah.
• 55:13
  I couldn't walk straight.
• 55:14
  It was weird.
• 55:16
  Yeah, and he had a very large credit card bill. All right, Doug. What's your hidden hidden talent or skill hidden talent or skill?
• 55:26
  You know, most of the people that know me now probably don't know I used to water skill lot. I didn't know that. Yeah
• 55:32
  I was actually pretty good for fat guy
• 55:37
  One noise fat, so you know I'm saying I
• 55:41
  I did. I used to slalom. I used to do. I didn't know. I never did it competitively or anything.
• 55:44
  Like I did. I grew up slalom. That's, I was skiing when I was a kid.
• 55:48
  Did you have a wakeboard?
• 55:49
  I did. I wakeboard. I, uh, I kneeboarded. I one time took the worst fall I've ever taken in my
• 55:55
  life trying to impress a girl on a, with a kneeboard.
• 55:58
  Well, first of all, it's pretty tough to impress girls from a kneeboard right out of the gate.
• 56:02
  That's a, you're a lot more impressive on a ski or a weight.
• 56:06
  Actually pretty good on the kneeboards.
• 56:08
  - Yeah, but you're still on your knees to start with.
• 56:10
  There's very few things that you look particularly
• 56:14
  impressed with.
• 56:15
  (laughing)
• 56:16
  When you start by begging.
• 56:18
  - In Singapore, they were like, yeah, it's cool.
• 56:20
  (laughing)
• 56:22
  - They tell you anything you wanna know in Singapore
• 56:25
  when you give them another hundred bucks.
• 56:26
  - No, I tried to do, I was trying to learn
• 56:28
  to do a flip on one and I hit so hard
• 56:32
  when I did that one that it was the last time
• 56:34
  I tried to do a flip on one.
• 56:36
  I used to, I also used to barefoot.
• 56:38
  - Wow, that, now that, that is impressive.
• 56:41
  Yes, that's impressive.
• 56:43
  It's amazing how rock hard water is
• 56:46
  when you hit it with a velocity.
• 56:47
  - Oh, is it?
• 56:49
  - Yeah, you think landing in water, oh, that'll be fine.
• 56:51
  - Oh, that's easy.
• 56:52
  - Not so much.
• 56:53
  - That hurts like hell.
• 56:54
  Like, yeah, so how about you, Nick?
• 56:56
  You got any secret skills?
• 56:58
  - Man, I was just trying to think about it.
• 56:59
  I don't know that I have anything.
• 57:01
  I've tapped into the only things
• 57:03
  that I have any kind of aptitude towards,
• 57:05
  which was racing orientated.
• 57:07
  And then I don't know if I have a talent for it,
• 57:10
  but I've now switched careers completely to writing
• 57:14
  and I do that.
• 57:15
  I don't know outside of that that I can do much.
• 57:17
  - You don't do much outside of that.
• 57:20
  (laughing)
• 57:22
  - I have things I'm trying to do.
• 57:24
  - Mixed home alone, like his wife's left him alone
• 57:27
  and he hasn't left the house.
• 57:28
  I was like, dude, have you gone out and seen the sunshine?
• 57:30
  No, he's in the type of work.
• 57:31
  - I can see the sunshine out my window here.
• 57:34
  Yeah, yeah.
• 57:35
  Did the Sun see you? That's a better question.
• 57:38
  No, that's true.
• 57:39
  Yeah, yeah.
• 57:40
  Hey, it was there's no question that it is not a skill and people understand this by previous episodes,
• 57:46
  but I'm attempting to play the ukulele, but but I would not class that as a skill.
• 57:52
  And we all know now, so.
• 57:53
  Yeah, everyone knows from a previous episode that it is not not a talent.
• 57:57
  Hey, look, this has been it's always fun hanging out with you, man,
• 58:00
  and it's great to see you.
• 58:01
  And thank you so much for giving our listeners so much insight.
• 58:10
  Man, that's a conversation that we could have carried on like all night.
• 58:14
  I was trying not to take notes.
• 58:17
  Like I always worry about like look down and like because oh my gosh, you had so much information.
• 58:22
  What he taught us already just you and I talked to him a couple months ago has been a game
• 58:27
  changer.
• 58:27
  What he's talked about today, I think is just another level up and I'm definitely going
• 58:32
  to take some of it.
• 58:33
  - Yeah, it's evolving and changing, right?
• 58:34
  I mean, literally we were with him two and a half months ago
• 58:37
  and the stuff that he's got going on now compared to then.
• 58:41
  Yeah, it's up another level.
• 58:43
  So really cool.
• 58:44
  - I like the idea, I do like the idea of using it
• 58:46
  maybe for some marketing stuff.
• 58:48
  So there's gonna be some--
• 58:49
  - I've done a little bit.
• 58:50
  I've done a little bit.
• 58:51
  I've had it help me with some Facebook stuff,
• 58:53
  ad stuff that hasn't necessarily worked yet.
• 58:57
  But that's on Facebook more than it's on Claude, I'd say.
• 59:00
  But to have it analyze the ads, I think is going to be the handiest thing.
• 59:06
  That's going to be amazing for me.
• 59:07
  So, but maybe we'll see.
• 59:09
  We'll, we'll, I do plan to play with it a little bit, get it set up.
• 59:12
  So.
• 59:12
  All right.
• 59:12
  Thanks to everyone in Patreon.
• 59:14
  We did a really cool bonus with MJ.
• 59:18
  So if you're up for that, then dive over to our Patreon, check it out.
• 59:22
  It's only a few dollars a month to subscribe to our Patreon and you get all
• 59:25
  those bonus materials in there.
• 59:26
  We appreciate you.
• 59:28
  And big thanks to our friends at Vellum.
• 59:30
  couldn't do this without you. You guys are wonderful. We really, really appreciate it.
• 59:33
  And head over to TriVellum.com/TheTwoOrthers to learn a little bit more.
• 59:38
  And that helps us out because if they see you're clicking on them from us, they're
• 59:43
  like, "Oh, those guys are worth a while."
• 59:45
  Yeah, people are noticing.
• 59:46
  They're not just beautiful singers.
• 59:49
  I said no one ever. All right. All our links to all the stuff, I guess, stuff, MJ's links
• 59:56
  and all that kind of thing to his books and all that good stuff is in the show notes.
• 1:00:00
  And until next time, be cool to each other.
• 1:00:02
  Fair wins and following thieves.
• 1:00:04
  You've been listening to the Two Authors podcast with Nicholas Harvey and Douglas Pratt.
• 1:00:10
  Subscribe on your favourite platform for new shows coming every week.
• 1:00:15
  (upbeat music)