Full transcript
0:00All right, it is time to talk about make
0:02versus nadn for automations and
0:05specifically for automations that are
0:06useful for authors such as myself and
0:09others who follow this channel. And this
0:12is going to be an interesting one
0:13because they're both okay and they're
0:15the definitely the top two that you will
0:17hear about when it comes to AI
0:18automations. However, there are a few
0:21reasons why I just found myself
0:23gravitated towards one in particular and
0:26I'm not even using the other one
0:27anymore. So, I'm going to break down a
0:30little bit about what makes make good,
0:32what makes NAND good, and why you might
0:34want one or over the other. And by the
0:36end of this video, I'm going to tell you
0:38which is the one that I definitely
0:40recommend and is the only one I'm going
0:42to be teaching going forward as a result
0:44because I think there's not much use in
0:46teaching both. So, let's start with
0:48make. Now, if we go and and I share my
0:51screen here, if you go and create a new
0:53scenario in make, you'll be presented
0:55with this giant beautiful big button.
0:58And you just click on this and it'll
1:00give you some uh integrations and things
1:02that you can select from. And this is
1:06probably the biggest advantage that Make
1:08has is it is relatively easy to get
1:10started. You can just dive right in and
1:12start making your first automation right
1:14out of the bat. And it also has a really
1:17good-looking UI. I I find it very easy
1:20and intuitive to understand in that
1:22sense at least for something that is not
1:24as easily understood by most people that
1:28are that are less technical. Make does a
1:30pretty good job of trying to make it as
1:32easy to understand as possible. The
1:34other major advantage that I think make
1:37has going for it is that it actually has
1:39a higher level of integrations. It's got
1:40like something like 2,000 integrations
1:42or something like that. You can see some
1:44of them here. And I do think at least in
1:47terms of builtin features, some of these
1:50are actually more complex and have uh
1:53more features essentially than what you
1:55get out of the box in N8. Uh for
1:58instance, if we look at Google Docs,
2:00there's lots of things that we can do
2:03with Google Docs. We can have it watch
2:04for documents, list a number of
2:06documents like in a folder or something.
2:08Get the contents of a document, create a
2:09new document, create a document from a
2:11template, uh, insert a paragraph into a
2:14document, insert an image into a
2:16document, replace an image with a new
2:18image, all of this stuff, right? Replace
2:21a text in a document, download a
2:22document, delete a document, make an API
2:24call, make all links in a document
2:25clickable. There are a lot of things
2:27like this where it's just gives you a
2:29bunch of options of what you can do with
2:30Google Docs. And when you look at the
2:33same node in N8, you will not see as
2:36many options. However, that doesn't mean
2:38that you can't do them in NADN. That
2:40just means that there isn't necessarily
2:42a easy button click for all of it.
2:45Sometimes you have to go through a few
2:47more hoops to get to what you want. So,
2:49for that reason, I think there are a few
2:51aspects of make that make it a little
2:53bit more beginner friendly. However,
2:55there are some downsides to make that we
2:57absolutely have to be aware of. And this
3:00is going to be really important given
3:01that we're coming at this from an
3:03author's perspective where we're looking
3:05to do some big projects and things like
3:07that. Uh as well as analyze a lot of
3:09content and do a lot of steps on that
3:11content. Some reasons that you might not
3:14want make include the following. Uh
3:17first of all, it's not nearly as
3:20customizable. While it does have a
3:22couple of those extra nodes and things
3:24that you can use as I showed off with
3:26the Google Docs, you can actually do
3:28more in N8N just with a little bit more
3:31technical knowledge. And N8N has some uh
3:35abilities to connect via API to a lot of
3:38things that don't have a direct
3:39integration. So make has a lot of
3:41integrations, but N8N can actually
3:43interface with probably more things
3:45overall just through an API. The second
3:47reason why you might not like make is
3:49that it actually operates on a credit
3:51system. And now I know a lot of you guys
3:53just from having seen comments on other
3:56software that I've talked about are not
3:58big fans of credit systems. And I
4:00honestly am not either. Now I I get it
4:02in certain circumstances like that's
4:04kind of the best way you could do it,
4:06but in other circumstances I just don't
4:08like it. Which means that every time
4:09you're running an automation, you're
4:11charged on every single operation that
4:15happens. So, every step of the process
4:17uh uses up a credit and some use up more
4:20than others. Now, they're usually
4:21generous with the number of credits that
4:22you get for the price, but still I know
4:24that rubs people the wrong way and you
4:27don't have anything like that in NAN.
4:29Another one that a lot of people don't
4:31talk about, but that was a big deal for
4:33me is the fact that there is a 45inut
4:37cut off for each automation. Now, a lot
4:41of the automations that I do can
4:42sometimes take a lot of time to run
4:45because you're running large expensive
4:48prompts on a lot of content. So, there's
4:51a lot of things that need to happen and
4:53those prompts can sometimes take quite a
4:55while to generate and then we'll have
4:57multiple steps of those prompts in which
4:59case, you know, it can sometimes take
5:01more than 45 minutes for an automation
5:03to fully execute. And if we're in make,
5:06I have to keep that in mind so that I
5:08don't let it go for too long. Otherwise,
5:11it'll hit that 45 minute cut off and
5:13just have an error. And I'll show you an
5:15example of what this looked like. So, I
5:17actually had um some automations built
5:20up here where um it would, you know, go
5:24through a couple Google Doc steps here
5:25and then it would have a four-step
5:28process on one chapter and then it would
5:31go to the next chapter and have a
5:32four-step process there and then the
5:34next chapter and then the next chapter.
5:36Uh and by the way, I found a better way
5:38of doing this in NAN. I was not able to
5:40figure out how to do it in in make, but
5:43I couldn't have done it anyway because I
5:45limit this to five chapters, actually
5:47six in this case cuz it includes a pre
5:49uh prologue. And the reason I limited it
5:51to five chapters is because any more
5:54than that and I risked going over that
5:5545minut limit. And with N8N there is no
5:58such limit. And that was a really big
6:01deal for me. Even though it's a small
6:02thing for my particular use case, it's
6:04it's a big deal. And I'll show you that
6:06this one here in NADN, it's the same
6:09process, uh, but adapted so that it
6:12could actually write all 40 chapters of
6:15a novel novel if I wanted to, assuming
6:17it's a 40 chapter novel, uh, which in my
6:19case it is. And here are the same four
6:22steps that we're repeating here, here,
6:25here, and here. Um, get done here. But
6:29in this case, I was able to figure out a
6:30way for it to loop around and then go to
6:32the next chapter and then continue to do
6:35that over and over and over again. As a
6:37result, I was also able to get this to
6:39be able to read the previous chapter,
6:41which is something that, you know, I
6:43could do something like that here in
6:45make, but it would be a little bit
6:47messier and not quite as efficient as
6:49what I have here. And sure, if I'm only
6:53doing five chapters at a time with this
6:55automation, it won't take too long. But
6:57if I wanted to do say 40 chapters all at
7:00once, write the whole book all at once
7:02and then go in and edit it rather than
7:04editing as I go, then this automation
7:07would definitely take longer than 45
7:09minutes. And so I like having that
7:11additional flexibility in being able to
7:14do these longer, more complicated uh and
7:17and lengthy automations. Last but not
7:20least, and when it comes to make, when I
7:22was working with it, I found that I was
7:24just running into a lot of errors. Now,
7:26uh, uh, admittedly, this is more of an
7:28anecdotal thing. Uh, I can't say for
7:30sure if make is going to be less buggy
7:32than NAN or not, but for me personally,
7:35and I was running quite a bit of
7:37automations in make before I switched to
7:39NAND, I found that it was quite buggy
7:41and that very often it would just have a
7:44random error, and I wasn't able to get
7:45really a good answer on why there was an
7:48error. Often, I think it was just
7:50because one of the steps just went on
7:53for too long. uh and it might have been
7:55doing like a deep research. In fact,
7:56yes. Um I was doing in one case I was
8:00trying to do a perplexity deep research
8:02and the deep research was just taking
8:03too long. So it just sort of timed out
8:06and that's not okay with me and I I
8:08don't get that kind that level of bugs
8:11in NN. I I still get a few timeouts and
8:15things that happen in NAN but the
8:17frequency that they happen for me at
8:18least seemed to be far fewer than what I
8:21was getting in make. And I was trying to
8:23make make work uh for me because I had
8:26heard that it was so much simpler and
8:27easier to use and I I like simple. I
8:30like making things easy for the authors
8:32that I share things with to be able to
8:34follow along with me. And so for that
8:36reason, I kind of put off get digging
8:38into NAD because I really wanted to make
8:41make work. And it just over the over
8:45time I was just having some issues with
8:46it. Not really able to do exactly
8:48everything that I wanted to do. And so
8:49from there, I decided to really give N8N
8:52a shot. And so let's now shift and talk
8:55about some of the reasons why you might
8:57prefer N8N over Make. Firstly, N8N also
9:01has a high level of integrations. Not as
9:03many as Make does, but when you count
9:05the ability to connect to pretty much
9:07any place that has an API, it's still
9:10pretty high. um the fact that they don't
9:12have a built-in integration for an
9:14something that you might have to connect
9:16to for an API might make it a little bit
9:18more difficult to connect to that thing,
9:20but you could still do it. That said,
9:22most of the integrations that you would
9:24need as an author, pretty much all of
9:26the ones that I need in my process and
9:28in the automations that I make for
9:30people, all of those integrations are
9:32fully accessible inside of NAND. It has
9:35access to uh all of the different AI
9:37providers and things like that,
9:38including Open Router. uh has access to
9:41the entire Google suite of tools and
9:43features. So pretty much anything that
9:46you would need for the most part unless
9:48you're in a really niche case are
9:50accessible inside of NAN. Now the second
9:53big reason why you might prefer NAND is
9:55the pricing structure. So it is quite a
9:57bit different from the way make operates
10:00their platform. Make is a credit system
10:03and they're reasonably affordable. like
10:05you can get a actually there's even a
10:07free plan where you can get a fair
10:08amount of of usage although the free
10:11plan that 45minut cut off that I
10:14mentioned is actually only I think 10
10:16minutes if you're on the free plan and
10:19uh and you get far fewer free credits
10:21that way. Uh with NADN there's three
10:24different ways you can go about it. The
10:26first is to go with the public cloud
10:29version. And that one I don't actually
10:31really recommend because it's $60 a
10:33month which makes it way more pricier
10:35than make. Uh if if that's if that were
10:39the only comparison, I'd say yeah,
10:41definitely go with make. It's more bang
10:42for your buck. However, NAN is an open-
10:45source tool. And so you can actually
10:48install it locally on your computer
10:50absolutely free. No no strings attached.
10:53The only issue I've seen with that is
10:55there's a little it requires a little
10:56bit of technical figuring out to get it
10:59installed on your local device. There's
11:01a third option that I like which is kind
11:03of the best of both worlds where you're
11:05able to self-host it on your own cloud
11:07server and that's usually quite a bit
11:10cheaper than going with the $60 a month.
11:14And uh there are service out services
11:16out there like Hostinger. Uh this isn't
11:18sponsored or anything. Hosting Air has a
11:20one-click install option for NADN, and
11:23that's actually what I use personally,
11:24and I'll do a separate video about all
11:26of the different ways to install NAND
11:28and to get it started with it. But
11:29suffice it to say, you can have NAND
11:32entirely for free loc uh if it is hosted
11:35on your on your computer. And that also
11:38means that there's no credit system.
11:39You're not charged per credit. Now, if
11:41you're still using AI in the platform,
11:43you'll still have to connect your API to
11:45the system and the AI usage will still
11:48be charged through the whatever service
11:51you're using. Like I use Open Router a
11:53lot, so I get charged through Open
11:54Router for a lot of my AI usage, but
11:57that's different. And like I'd still
11:58have to have that in make and in NAN.
12:01And so when we're just talking about
12:03like the overhead costs of using N8N as
12:06a platform versus make, I would choose
12:08N8N because you can install it locally
12:09on your computer and have all of the
12:11same functionality. And the third option
12:13I've kind of already mentioned is that I
12:15found that there are far fewer glitches
12:17and timeout problems in NAN. Uh again,
12:20you still get some, but that's been a
12:22far uh far less demanding issue as when
12:26I was working with make. The fourth
12:28reason is that NAN is actually better if
12:31you are concerned with like having full
12:33data control and full security and
12:35privacy of your data. If if that's
12:38important to you, then uh NAN is great
12:40because since you can install it on your
12:44local computer and not have it on the
12:45cloud, you have a greater degree of uh
12:49ownership and security with all of that
12:52in that model. And because there's no
12:5445minute limit or or anything like that,
12:56it's actually far easier to scale your
12:58projects in NN. Um, not that that's
13:02really a concern for most authors, but
13:03that is something to keep in mind. Like
13:05if I were a business looking to really
13:08scale the AI usage in my company and
13:12automate a lot of the plat the
13:14processes, I would definitely be
13:16choosing NAND and getting a specialist
13:18to to do all of that for me and scale it
13:21because NAN's definitely better for a
13:24more robust infrastructure. And last but
13:27not least, NN because it's an open
13:29source thing, there's a really strong
13:31open source community out there. If you
13:33need uh help getting things um installed
13:36or figuring out how to do XYZ, uh the
13:40community is pretty open and uh you
13:42know, you can find many places online to
13:44ask those questions. The only real
13:46downside of NADN is that it is a little
13:50bit more complicated to use. However, uh
13:55you know that was something that put me
13:57off from it for a little while. I had
13:59heard that it was more complex and that
14:01make was so much easier. And for that
14:03reason, I started with Make and only got
14:06into NAN when I grew a little bit
14:08dissatisfied with some of the features
14:09in Make. And once I got into NAND and
14:11got past the first couple of hurdles,
14:13like getting it set up, that's probably
14:15the biggest one. I found it to be just
14:17as intuitive as Make. Now, maybe I'm an
14:20exception. uh maybe some people would
14:22find make easier, but I found that the
14:24actual setting up of an automation and
14:26putting all the little pieces together
14:28and running your prompts and all of that
14:30was absolutely no harder in NAN than it
14:35was in Make. The only thing I would say
14:37that was a little bit harder with NAND
14:39was the initial setup, but once you've
14:41got that down, it's really easy to get
14:44started and really intuitive. And
14:46honestly, I'm going to have a couple of
14:47videos that will walk you through some
14:49of those initial startup tasks and
14:52hopefully that will make it a lot easier
14:54for you to get started in N8N. And the
14:57fact that it's actually kind of about
14:59the same amount of effort to get a
15:01automation set up in make versus N8N.
15:04Given that N8N has way more potential of
15:08what it can do, I would absolutely go
15:11with NAND over make. If you can master
15:13N8N, then if you wanted to switch over
15:15to make, like maybe you get a client
15:17that wants something built in make and
15:19you uh you can build it for them. If you
15:23know how to use NAND, make will just
15:25come very easily to you. And so I do
15:28recommend the authors start with NAND.
15:31You don't have to be a coder to do
15:33things in NAND. Although you if you are
15:35a coder, that can increase the amount of
15:38things that you can do in NAN. I've even
15:40done a tiny bit of vibe coding inside of
15:44N8N to do certain steps and I was able
15:47to figure that out very easily with
15:48absolutely no background in code myself.
15:51I do not know JavaScript, PHP, uh,
15:54Python, any of any of the major
15:56languages. The only one I know a smidgen
15:58about is HTML and CSS, but those are
16:01like not even really proper coding
16:04languages. Um, so if I can figure it
16:07out, I know anybody can. And especially
16:09if if you come in and start with the
16:12things that I've created and work with
16:15those to get started and maybe you'll
16:17just work on figuring out how to tweak
16:18things for your own purposes. It's not
16:21going to be very difficult at all for
16:23you to figure out how to use it. So, if
16:25you haven't been able to tell by now,
16:27you can tell that I really love NAN. And
16:30as far as I'm concerned, I will only be
16:32teaching NADN from here on out. So, all
16:35of the future automation videos that I
16:37make are all going to be centered around
16:39N8N as the primary platform of choice.
16:43There are others, even besides make,
16:45there are others out there, but if you
16:47can master N8N, the rest should come
16:50pretty easily to you. So, I highly
16:52recommend that you also follow along in
16:54this journey and figure out how to use
16:57Nadn for your own purposes. And that's
16:59also what I'm doing inside of the story
17:01hacker group, which you can find a link
17:02down below where I am sharing all of my
17:04automations that I create. Those all go
17:07in that group, as well as a number of
17:09instruction and things beyond what you
17:11just see here. Plus, you've get access
17:13to the community. So, if you have a
17:15problem, if you're stuck on something,
17:16you can probably find somebody in the
17:18community to help you out. I'm in there
17:19answering questions every day as well.
17:21So, check that out if you're interested.
17:23That pretty much wraps it up for this
17:24video. As you can tell, we're just going
17:26to be going ahead with NADN in the
17:28future. And in the next video, we're
17:30going to actually talk about how to get
17:32it installed, how to uh how to get it
17:35going for you and your business. So stay
17:37tuned for that. and I will see you in
17:38the next