Cloud is much more than hardware though, it’s security APIs and object storage, networking etc; things that used to require a human to configure. Cloud offers about 80% of what physical hardware provided before, but it was enough that nearly all people moved to the cloud.
Anyway, I think you miss my point, a lot of dev work is either creative in its core (web design) or gluing APIs together.
No-code can get you a really long way in theory, how would that not be a commodity at that point?
It only needs to be 80% useful and it will be ubiquitous.
I'm a coder who's built a lot of platforms for startups from the ground up. I have some scorn for people who are too good to learn the technical; when I hire a roofer, he's looking at me like I'm an idiot, so same goes. But. I always thought, hey, I've got the coding skill so why shouldn't I have it all - and be fabulously successful from some good idea I cobbled together with nothing but 1's and 0's in my free time? So, I built six or seven apps on my own recognizance, none of which took off, but which were all (I think) revolutionary concepts at the time, and each of which took a year or so to be fully functional platforms.
The other apps I built in the interim, to make money. These I built as a freelancer for people who were very competent in their particular businesses and were using web/mobile apps to basically streamline things and improve efficiency; I would reject anyone coming to me with a job like "I have this idea for an app..." because I have a dozen ideas for apps that are probably better than theirs, and I know how to actually build them, but what I emphatically have never had are a functioning existing business, business model, growth strategy or investment for those things. Ideas are just ideas. Even if you take years to develop the code skills to build them swiftly. In fact, if you do, it's even more painful to watch them vanish after a little while.
1's and 0's are just electrons on silicon, after all. We're writing in sand.
I've come to ...well, not exactly love or respect the fast-talking business guys, but... appreciate that maybe focusing on just how to build things (whether you code them or not) is really not the key. It's a necessary ingredient, for sure. But like I said, I've got a dozen good ideas for apps. Non-technical people have like hundreds of bad ideas for apps, and most of them want to catapult themselves to success with neither a good idea nor a functional business model nor existing investment -- they just want to entice someone like me to work for them on faith/credit, and it's always been this way, for 30 years now.
Sorry to rant. But what does work is when people have a successful business already, or the skills to set one up, which I don't have. When they have domain knowledge in their field, whether that's the food industry or hotels or manufacturing. Then, finding someone to build the software creatively or learning to do it yourself can add a whole layer of value, because you already understand the business logic inside and out.
There will be a few unicorn dream startups that exist only 1's and 0's, in the 2020s... and as sad as I am about how all those things turned out since the 1990s, I'm not totally closed to the possibility of being at the front of one. But that's not where the bulk of wealth will be created. It'll be in building usable software that acts as a multiplier for the capacity of ordinary businesses... and not necessarily in the "take a 2.5% cut" way.
You’re overestimating the importance of business people. Viewing products from a business point of view inevitably leads to inferior products, as the focus is on purely making money. It’s not about making money or about the synergy of business and development. It’s all about creating useful shit for other people to use. You need domain expertise, good taste, and the ability to create value by building things. Focus on money and you’ll be cutting corners, trying to sell useless shit to people who buy it for others to never use.
I think we're talking about different things. Take three scenarios.
1. Mediocre idea for an app with a lot of funding, raised a competent team of coders, business school grads with no experience in the lead. This will fail. Those guys will take the money and buy jeeps, degrade the product with bad ego-driven ideas, etc.
2. Great idea for an app with no funding, founded by coders. These people will work their ass off all weekend for free. Business people can't see the value; will never see the value; if they touched it they'd ruin it, but they won't touch it. The app launches, gets to the front page of HN, a week later it's dead because there's no business model. Here you could make a case that some business people should have been involved.
3. The ideal case. This is not an idea for an app. This is not really even an app. Someone has a glass factory. Or a small chain of hotels. Their employees are constantly doing things on a whiteboard in the back of the office. They want to expand to more locations and scale up, but the system they have just gets more confusing and complicated the bigger they get, it's really hard to train new employees to use this whiteboard and pass these paper slips around the factory. They come and ask you how you can make it better and you build them an app that they pay for out of pocket, which puts the system up on screens all over the facility. One-off deal. Then they grow and they ask for more apps to manage their customers, their inventory, their billing. Multi-one-off-deals, but now you have an ecosystem you maintain for them.
Those are the business people I'm talking about, not the ones who approach you with an elevator pitch where "it's like Uber but for dogs" or something.
Forget technical details; no code systems let you hire cheap replaceable folk instead of highly paid technical folk. …or, just a few technical folk to clean up all the rubbish the others make.
..and who’s to judge?
Yes, you end up with a pile of crap.
…but, if it lets you bootstrap your business without spending half of your capital on expensive programmers, isn’t that a net win?
Let’s be honest; in most situations, cloning an existing successful business is a pretty easy, well defined task. That’s why you get things like “here’s my free clone of strava” or whatever.
Once the problem is solved, building it is easy. Defining the solution, building a customer base, handling support… those are the things which are (usually) hard and expensive.
So.. you know, don’t trash talk no code without understanding what it’s for.
The exception for this is when you’re trying to do something technically difficult; but really, most businesses probably aren’t.
If your business/startup is, then don’t pick no code; you’ll fight the tools trying to bend them into your shape and end up with expensive [no code system here] experts, defeating the entire point.
That’s it. In all other circumstances, you probably want some combination of no code stuff / generic off the shelf stuff to let you get past all the tedious stuff you don’t care about.
That's what is promised under the term "no code" or "low code", but it's not how it actually works.
Low code and no code already exist and they are already successful where they work. For instance, you can now create a website using wordpress with quite a bit of functionality without coding at all. That _is_ "no code". And you can use Excel to create quite complicated logic and calculations with very little coding. And that _is_ "low code". Even using an IDE and a high-level programming language is in some sense "low code".
Pretty much every other usage of this term by startups or others is purely marketing and promising something it can't do. I wish it were different because then I could build things even faster with less effort, but from what I can see and what the last decade shows, progress towards this goal is made in different ways.
> Let’s be honest; in most situations, cloning an existing successful business is a pretty easy, well defined task. That’s why you get things like “here’s my free clone of strava” or whatever.
> Once the problem is solved, building it is easy.
Yeah, but the mistake is to assume that solving the problem doesn't require technical folks. Even if something is not technically difficult, it still requires a lot of work and skills to translate vague thoughts in human language into computer-executable logic that behaves as one wants. This will be solved once we have some intelligent machine that can do that. Once we have that, the world will turn upside down.
> no code systems let you hire cheap replaceable folk instead of highly paid technical folk
Until they don't, then you get to this point:
> just a few technical folk to clean up all the rubbish the others make
Which gets exponentially harder and more expensive as the size of the "no-code" codebase grows.
Going no-code for a startup is crazy. Your failure state isn't "it ends up being x% more expensive", it's that you hit a brick wall and you either rewrite from scratch (if you're small and agile enough), or your startup dies instantly (if you have enough users and competition that not delivering any features ever again isn't a viable strategy).
If I'm an investor and you're pitching a startup that's using no-code for anything other than a throwaway version for the purposes of a sales pitch, that's an instant dealbreaker. I don't care how good the idea is or how smart your team is, I'll happily bet the house on it failing.
> Forget technical details; no code systems let you hire cheap replaceable folk instead of highly paid technical folk.
Honestly, from what I've seen, they're not that replacable. They're less of them so straight away it becomes harder to replace then. They cost roughly the same per hour as other devs in that area.
no code systems let you hire cheap replaceable folk instead of highly paid technical folk.
Does it? Anecdotally and based on observation, people who know their way around, for example, PowerBI and can use it solve problems earn the same as people who know their way around Python or JavaScript. And people who really understand PowerBI and actually use it to its full potential can charge every bit as much as talented programmers.
People who “really understand PowerBI and actually use it to it's full potential” are, on top of other skills, talented programmers proficient in a niche functional language, so it's not really surprising that they can charge every bit as much as talented programmers.
(Of course, not sure what this says about “no code” platforms except that the label is usually a lie.)
I understand the marketing of "NoCode" but doesn't it really mean app with a gui? Is Mailchimp is NoCode? Is Squarespace is NoCode? Or does it only apply to an app that can create an app that does something else?
Great. This will be VB6 all over again: tightly coupled apps (coupled frontend/backend, coupled SW/OS/HW) depending on a runtime that will go unsupported or killed within 5 years, and will soon require a VM to run at all.
Vendor lock-in, technical debt, no scalability, terrible code efficiency. Great way to save a few bucks...
I don't have to read. Is it in anyway bespoke, then code, always.
Is a shared excel sheet OK? Then go with that.
For technical people doing side projects doing no code is a miserable experience. Put that effort into learning a framework, or learning your existing framework better.
The exception is no code sprinklings, like using IFTTT for some automations would make more sense. Or a google sheet for some static data maybe. Or Google calendar as a ical source. Stuff like that can be smart.
The dirty secret is when using a nocode tool you are actually coding with an obnoxious interface that made the first 5% easier and the next 95% harder to impossible and way more expensive and slow.
I may be wrong in the future, and I hope so!, but the tech isn’t there yet.
The hard part of problem solving with software is not writing the code, it's understanding and solving the problem.
You have to do that whether you use nocode or code. It just so happens that the people who are highly skilled at solving new or abstract problems have built a whole ecosystem of highly effective tools for executing the solutions to new or abstract problems, and they're very efficient at using them.
If you have the problem solving skills but not the coding ability and you need to rapidly prototype an idea and you don't have time to learn how to code, then a nocode platform could be a great short term choice.
You'll have to throw it in the bin if it turns out to be a success though!
I think it helps to view no-code as the latest iteration of a pattern that started back with COBOL and has included things like Excel, Wordpress, Squarespace, or Shopify wherein some common subset software tasks become massively deskilled (e.g. being able to set up an online store was once a massively profitable skillset and is not anymore).
While understanding the problem to be solved is, of course, hard, implementing a solution is much easier if you can be freed from the most common concerns of translating your business solution to code.
> You'll have to throw it in the bin if it turns out to be a success though!
I think, again, excel shows this is unlikely to be true. There are billion dollar businesses running on a giant mess of excel spreadsheets (I've seen it). They're good enough and I expect a few of the current no-code contenders will go the same route.
None of this is to say the end of software is here (and I don't think serious -people in the no-code space expect that), just that a new class of software may be going through a process we've seen other types of programs go through.
> If you have the problem solving skills but not the coding ability and you need to rapidly prototype an idea and you don't have time to learn how to code,
IMO it is perfectly possible for a person to be as good as a programmer at problem solving, and thinking abstractly about user-facing solutions and workflows (maybe even some non-user-facing) and yet have no interest at all in learning a text-based programming language.
The whole premise of coding originated a million years ago, when it was mandatory to think in terms of a processor executing an instruction sequence, and operating systems and compilers needed to be written. Is it really the best way to make an app? I think plenty of apps are amenable to thinking about in terms of visual design elements, triggers, events, actions that act on other things that expose endpoints, etc.
Like all things, I don't know whether anyone has done a great job devising a no-code system yet. It's not my bag. But I maintain it's possible to do a great job in principle.
I used to work on niche low-code systems that still required some coding in a very simplistic language. We had hundreds of salespeople who were absolutely not programmers and had zero knowledge of CS, but were all out in the field rigging up all manner of customized UIs and integrations for the customers. I don't recall them complaining or saying it was too hard.
My boss a few years and a few jobs back wanted me to try a no code solution for our integrations with our ERP. The idea being when I was gone a Business Analyst would be able to modify it. I did what he asked but after I left I found out the first thing they did was replace it because the BA couldn't figure it out.
In some areas I think it is great and I use it. My main issues with it.
Nocode still needs the same development process and change management which is a large part of software development. Even though nocode is sold as you can just flip switches.
Nocode always comes with limitations. Code does not.
Fixing things like bad data structures never seem to happen in nocode.
Nocode has been around since at least early 80s. Sold under different names. There are plenty of obsolete workflow systems out there.
Nocode comes with the risk of having too much configuration. Highly configurable systems are known to be expensive to, guess what, configure.
Nocode solutions always end up looking like Brasilia. A grandiose idea with much of the necessary stuff in the outskirts outside of the main plan(e).
I think there will be a good deal of useful software created using NoCode and LowCode tools and almost all of it will be created by software developers. Very few non-technical people will learn enough to be useful software. Some non-technical people will have an idea and create enough to get the rest built out by software developers one way or another.
No/Low Code has a future in the hands of developers.
25 comments
[ 4.4 ms ] story [ 63.5 ms ] threadLots of vendor lock-in and paying extra for things like access to a proper database etc.
For the price of the license fees for a few 100’s of users you could pay a dev to build something better bespoke.
Lots of vendor lock-in, lots of overpaying for simple things.
Why is this any different than that?
Anyway, I think you miss my point, a lot of dev work is either creative in its core (web design) or gluing APIs together.
No-code can get you a really long way in theory, how would that not be a commodity at that point?
It only needs to be 80% useful and it will be ubiquitous.
I'm a coder who's built a lot of platforms for startups from the ground up. I have some scorn for people who are too good to learn the technical; when I hire a roofer, he's looking at me like I'm an idiot, so same goes. But. I always thought, hey, I've got the coding skill so why shouldn't I have it all - and be fabulously successful from some good idea I cobbled together with nothing but 1's and 0's in my free time? So, I built six or seven apps on my own recognizance, none of which took off, but which were all (I think) revolutionary concepts at the time, and each of which took a year or so to be fully functional platforms.
The other apps I built in the interim, to make money. These I built as a freelancer for people who were very competent in their particular businesses and were using web/mobile apps to basically streamline things and improve efficiency; I would reject anyone coming to me with a job like "I have this idea for an app..." because I have a dozen ideas for apps that are probably better than theirs, and I know how to actually build them, but what I emphatically have never had are a functioning existing business, business model, growth strategy or investment for those things. Ideas are just ideas. Even if you take years to develop the code skills to build them swiftly. In fact, if you do, it's even more painful to watch them vanish after a little while.
1's and 0's are just electrons on silicon, after all. We're writing in sand.
I've come to ...well, not exactly love or respect the fast-talking business guys, but... appreciate that maybe focusing on just how to build things (whether you code them or not) is really not the key. It's a necessary ingredient, for sure. But like I said, I've got a dozen good ideas for apps. Non-technical people have like hundreds of bad ideas for apps, and most of them want to catapult themselves to success with neither a good idea nor a functional business model nor existing investment -- they just want to entice someone like me to work for them on faith/credit, and it's always been this way, for 30 years now.
Sorry to rant. But what does work is when people have a successful business already, or the skills to set one up, which I don't have. When they have domain knowledge in their field, whether that's the food industry or hotels or manufacturing. Then, finding someone to build the software creatively or learning to do it yourself can add a whole layer of value, because you already understand the business logic inside and out.
There will be a few unicorn dream startups that exist only 1's and 0's, in the 2020s... and as sad as I am about how all those things turned out since the 1990s, I'm not totally closed to the possibility of being at the front of one. But that's not where the bulk of wealth will be created. It'll be in building usable software that acts as a multiplier for the capacity of ordinary businesses... and not necessarily in the "take a 2.5% cut" way.
j/k
1. Mediocre idea for an app with a lot of funding, raised a competent team of coders, business school grads with no experience in the lead. This will fail. Those guys will take the money and buy jeeps, degrade the product with bad ego-driven ideas, etc.
2. Great idea for an app with no funding, founded by coders. These people will work their ass off all weekend for free. Business people can't see the value; will never see the value; if they touched it they'd ruin it, but they won't touch it. The app launches, gets to the front page of HN, a week later it's dead because there's no business model. Here you could make a case that some business people should have been involved.
3. The ideal case. This is not an idea for an app. This is not really even an app. Someone has a glass factory. Or a small chain of hotels. Their employees are constantly doing things on a whiteboard in the back of the office. They want to expand to more locations and scale up, but the system they have just gets more confusing and complicated the bigger they get, it's really hard to train new employees to use this whiteboard and pass these paper slips around the factory. They come and ask you how you can make it better and you build them an app that they pay for out of pocket, which puts the system up on screens all over the facility. One-off deal. Then they grow and they ask for more apps to manage their customers, their inventory, their billing. Multi-one-off-deals, but now you have an ecosystem you maintain for them.
Those are the business people I'm talking about, not the ones who approach you with an elevator pitch where "it's like Uber but for dogs" or something.
Forget technical details; no code systems let you hire cheap replaceable folk instead of highly paid technical folk. …or, just a few technical folk to clean up all the rubbish the others make.
..and who’s to judge?
Yes, you end up with a pile of crap.
…but, if it lets you bootstrap your business without spending half of your capital on expensive programmers, isn’t that a net win?
Let’s be honest; in most situations, cloning an existing successful business is a pretty easy, well defined task. That’s why you get things like “here’s my free clone of strava” or whatever.
Once the problem is solved, building it is easy. Defining the solution, building a customer base, handling support… those are the things which are (usually) hard and expensive.
So.. you know, don’t trash talk no code without understanding what it’s for.
The exception for this is when you’re trying to do something technically difficult; but really, most businesses probably aren’t.
If your business/startup is, then don’t pick no code; you’ll fight the tools trying to bend them into your shape and end up with expensive [no code system here] experts, defeating the entire point.
That’s it. In all other circumstances, you probably want some combination of no code stuff / generic off the shelf stuff to let you get past all the tedious stuff you don’t care about.
Low code and no code already exist and they are already successful where they work. For instance, you can now create a website using wordpress with quite a bit of functionality without coding at all. That _is_ "no code". And you can use Excel to create quite complicated logic and calculations with very little coding. And that _is_ "low code". Even using an IDE and a high-level programming language is in some sense "low code".
Pretty much every other usage of this term by startups or others is purely marketing and promising something it can't do. I wish it were different because then I could build things even faster with less effort, but from what I can see and what the last decade shows, progress towards this goal is made in different ways.
> Let’s be honest; in most situations, cloning an existing successful business is a pretty easy, well defined task. That’s why you get things like “here’s my free clone of strava” or whatever.
> Once the problem is solved, building it is easy.
Yeah, but the mistake is to assume that solving the problem doesn't require technical folks. Even if something is not technically difficult, it still requires a lot of work and skills to translate vague thoughts in human language into computer-executable logic that behaves as one wants. This will be solved once we have some intelligent machine that can do that. Once we have that, the world will turn upside down.
Until they don't, then you get to this point:
> just a few technical folk to clean up all the rubbish the others make
Which gets exponentially harder and more expensive as the size of the "no-code" codebase grows.
Going no-code for a startup is crazy. Your failure state isn't "it ends up being x% more expensive", it's that you hit a brick wall and you either rewrite from scratch (if you're small and agile enough), or your startup dies instantly (if you have enough users and competition that not delivering any features ever again isn't a viable strategy).
If I'm an investor and you're pitching a startup that's using no-code for anything other than a throwaway version for the purposes of a sales pitch, that's an instant dealbreaker. I don't care how good the idea is or how smart your team is, I'll happily bet the house on it failing.
Honestly, from what I've seen, they're not that replacable. They're less of them so straight away it becomes harder to replace then. They cost roughly the same per hour as other devs in that area.
Does it? Anecdotally and based on observation, people who know their way around, for example, PowerBI and can use it solve problems earn the same as people who know their way around Python or JavaScript. And people who really understand PowerBI and actually use it to its full potential can charge every bit as much as talented programmers.
(Of course, not sure what this says about “no code” platforms except that the label is usually a lie.)
Vendor lock-in, technical debt, no scalability, terrible code efficiency. Great way to save a few bucks...
Is a shared excel sheet OK? Then go with that.
For technical people doing side projects doing no code is a miserable experience. Put that effort into learning a framework, or learning your existing framework better.
The exception is no code sprinklings, like using IFTTT for some automations would make more sense. Or a google sheet for some static data maybe. Or Google calendar as a ical source. Stuff like that can be smart.
The dirty secret is when using a nocode tool you are actually coding with an obnoxious interface that made the first 5% easier and the next 95% harder to impossible and way more expensive and slow.
I may be wrong in the future, and I hope so!, but the tech isn’t there yet.
The hard part of problem solving with software is not writing the code, it's understanding and solving the problem.
You have to do that whether you use nocode or code. It just so happens that the people who are highly skilled at solving new or abstract problems have built a whole ecosystem of highly effective tools for executing the solutions to new or abstract problems, and they're very efficient at using them.
If you have the problem solving skills but not the coding ability and you need to rapidly prototype an idea and you don't have time to learn how to code, then a nocode platform could be a great short term choice.
You'll have to throw it in the bin if it turns out to be a success though!
I think it helps to view no-code as the latest iteration of a pattern that started back with COBOL and has included things like Excel, Wordpress, Squarespace, or Shopify wherein some common subset software tasks become massively deskilled (e.g. being able to set up an online store was once a massively profitable skillset and is not anymore).
While understanding the problem to be solved is, of course, hard, implementing a solution is much easier if you can be freed from the most common concerns of translating your business solution to code.
> You'll have to throw it in the bin if it turns out to be a success though!
I think, again, excel shows this is unlikely to be true. There are billion dollar businesses running on a giant mess of excel spreadsheets (I've seen it). They're good enough and I expect a few of the current no-code contenders will go the same route.
None of this is to say the end of software is here (and I don't think serious -people in the no-code space expect that), just that a new class of software may be going through a process we've seen other types of programs go through.
isn't this, like, most humans on earth ?
The whole premise of coding originated a million years ago, when it was mandatory to think in terms of a processor executing an instruction sequence, and operating systems and compilers needed to be written. Is it really the best way to make an app? I think plenty of apps are amenable to thinking about in terms of visual design elements, triggers, events, actions that act on other things that expose endpoints, etc.
Like all things, I don't know whether anyone has done a great job devising a no-code system yet. It's not my bag. But I maintain it's possible to do a great job in principle.
I used to work on niche low-code systems that still required some coding in a very simplistic language. We had hundreds of salespeople who were absolutely not programmers and had zero knowledge of CS, but were all out in the field rigging up all manner of customized UIs and integrations for the customers. I don't recall them complaining or saying it was too hard.
Nocode still needs the same development process and change management which is a large part of software development. Even though nocode is sold as you can just flip switches.
Nocode always comes with limitations. Code does not.
Fixing things like bad data structures never seem to happen in nocode.
Nocode has been around since at least early 80s. Sold under different names. There are plenty of obsolete workflow systems out there.
Nocode comes with the risk of having too much configuration. Highly configurable systems are known to be expensive to, guess what, configure.
Nocode solutions always end up looking like Brasilia. A grandiose idea with much of the necessary stuff in the outskirts outside of the main plan(e).
No/Low Code has a future in the hands of developers.