Low Code Scares Me
Вставка
- Опубліковано 5 лип 2024
- The "low code" and "no code" promises have been around for the better part of a decade. I still haven't seen them pay off. For every Rippling and Webflow, there are dozens of...less great tools. And don't get me started on the lockin and specialization necessary.
LINK TO ARTICLE: nick.scialli.me/blog/why-im-s...
Check out my Twitch, Twitter, Discord more at t3.gg
S/O Ph4se0n3 for the awesome edit 🙏 - Наука та технологія
So many of these tools spend their time convincing people that code is the hard part. Then down the road you learn the hard way that the code was never the hard part.
What are the hard parts ???
@@abdelazizlaissaoui9079 in my experience the design. Collecting requirements from clients is difficult since they often don’t know what they actually need.
@@abdelazizlaissaoui9079 Those tools that were meant to simplify!
@@abdelazizlaissaoui9079 managing complexity, mostly. But my other gripes are most of the low code solutions I’ve dealt with had no way to integrate with source control (because no code), no good way to collaborate, no good way to document (basically a bunch of screenshots of their diagrams with explanations next to them in a google doc. No good way to do “code” review, no good way to use code when you reach that 20%, no portability because there’s no form to export the diagrams to. Most of these are problems that devs solved decades ago, and now these people building the platforms are solving them all over again.
@@abdelazizlaissaoui9079 I actually used a platform that didn’t have an “undo” button, if you can believe that, and they advertised all over SF for a while. Just imagine changing one thing that’s repeated a few times in some code. Ctrl+d a few times maybe? Ctrl+f and replace all? No such thing in low code, you have to go into every single node, check if it’s called, scroll to the dropdown that called it, change it, and hope you didn’t miss one.
non-technical management always wants a mock up as soon as possible, then say "let's use the mockup for production". lol
Im dealing with this now. So the code is spaghetti and a house of cards because it was made as a proof of concept not as an expansible maintainable and performant tool
Never make a prototype that looks like it can be used as is. Made that mistake once, they released it company wide in multiple countries. Fun times.
@@yannick5099 oh boy
I had a friend who was doing a landing page for a non technical CEO, and he used the design doc as an img of the landing page and created the website using just an tag and he demo’d with it successfully lol
This hits too close to home. I can put 10 callouts warning why this code is only a PoC and it would somehow still end up in production.
it all comes down to the fact that you can remove coding from engineering, but you can never remove engineering from engineering
This!
and coding is the easy part unless you are doing aracane compiler optimizations
Exactly. And with modern tools, coding is actually really easy as long as you use well-tested standards.
I started in no-code and ended up in all-code. From day one I had these "if only the tool allowed me to do Z on top of ABC...XY"-moments.
I heard a rule of thumb a while back: The first 80% of your app takes 80% of the time. The remaining 20% also takes 80% of the time.
Almost right, the 80-20 rule is that the last 20% of any project takes 80% of the time. Naturally that means the first 80% takes 20% of the time since it has to add up to 100%
@@Greedygoblingames /whoosh
And yet there's no mention of WordPress, the final boss at the root of all low-code evil. "Shenanigans under the hood" doesn't even begin to describe what some of these plugins do, and the only reason you're involved is because they're "90% done" and just need you to do one tiny thing for customization...
Totally agree with your comment, but as a developer who has used WordPress since its earlys I can share that it was not created as a low code tool but as a lightweight and well-organized system that gave you enormous flexibility to build websites, by programming a custom theme. And it is so true that it does not come from a low code vision that its slogan was "code is poetry". Afterwards, wp evolved towards low code, more as an inevitable destiny driven by its community of users than as a decision by Automattic, and not in a good way, but through the use of heavy and inefficient builders as well as the abuse of plugins, many of them poorly programmed. The contrast with a truly nocode/lowcode tool like Webflow (which I recently started working with) is huge.
@@rodrigomejiaarmijo I was there when WordPress was ONLY a blogging platform, & users were begging to have a customizable landing page. Perhaps I should not have walked away from WP for having such bad 'scope creep'?
WordPress actually is exactly what you build it to be (in a good and a bad way). If someone says that their site is built with WordPress, it doesn’t really say anything. It can be a site built with Elementor + 30 plugins or it can be a full React application using no plugins at all. No one really shouldn’t build large applications with wordpress but it is possible.
How I like to approach WP is purely by using it as a PHP/React framework and allow users only to edit or add content. So Gutenberg is only used as an CMS. I like how I can still do anything I want without hacking the system and offer familiar UX for the customer.
WordPress core is moving to a great direction. It is just getting a lot of hate because of low-code page builders and people forget that wordpress is so much more.
@@heikkivihersalo That's correct!
Reading this hurts. Especially considering how a couple years ago I set up a WordPress because I just wanted to publish a few things and not care about /how/. Fast forward to now, and the amount of time I've spent fixing it when it spontaneously breaks itself, and some things I literally can't even fix unless I'm willing to become a "WordPress expert" - fuck that.
I should never have used it, and now it's hard to move away.
Excel gurus.
I’ve met sooo many brilliant people waste sooo much time building complex stuff using excel and VBA instead of a proper programming language and a real database…
Those people drive me up the wall. Just because you can doesn't mean you should. It's like trying to build a house with sand and clay because it's the first material you came across and is easy to understand the basics of, instead of spending a little bit of time walking to the forest on the horizon to build the house out of something solid like rock and trees.
Hehe look up stdVBA ;) Also look up my article on "Why do people use VBA?"
> @Sammysapphira Just because you can doesn't mean you should
You're talking as if people have a choice... The truth is, if you're given the task to dig a hole, and you're given the choice between a shovel and a spoon, you're going to choose the shovel. The fact that some people elsewhere in the world can use a digger is totally besides the point. If that's not an option for you because corporate IT won't let you use it, tough luck. This is why 1/3 of VBA users use VBA. Honest to god, I'd adore a nice modern language like Rust, I've campaigned in the organisation for this too. But we've been told it's either spoon or shovel. We chose shovel.
Why? Sometimes in big corpo all you have is excel. I remember waiting for approval to install docker for 3 months, as a machine learning engineer lmao; waiting for like half a year for proper kerberos setup, etc (top5 biggest businesses in Europe).
It's all fun and smiles until you have to work in the real world (not SV bubble).
@@Sammysapphira That assumes you have the choice. If you check my article on why people use VBA, you'll see that 1/3 of respondants have no other choice. I'd love a modern language like Rust or TypeScript, but the simple reality is that engineers are given no modern tools by the IT department. IT usually can access these. Our IT department on the other hand outsource all the dev work to india and retain 0 inhouse knowledge of the systems they build.
@@boccobadz what is this "real world" you speak of? your indian sweatshop call center?
webflow is the gateway drug to full stack
This is the best comment 😂
100% true. We onboard a low-code tool for automation/alert. The perks are the out of the box integration with different apps/tools like TEAMS. It works like a charm. But when we want to customize/transform data, it couldn't handle it. Either it takes MINUTES or timeout. Something that is easily accomplish with Python or PowerShell within seconds. :(
Almost as if UNIX was onto something with it is shell scripting
I used to be skeptical, but right tool for the right job. We've got about 4 no-code production applications that we've had no problems with.
It's not so much the development time that's most enticing; rather they come out of the box with most things corporate environments need (AD integration with a Auth provider, claims based security out of the box, did well in a pen-test, included auditing, logging, in-product tech-debt monitoring etc). But where they shine are simple CRUD applications.
We learned quickly for anything more complex, we call a web-service we've coded and underlying back-end, and let the no-code do the UI.
Very rarely we'll have to do something extra fancy on the UI, but since the tool generates React under the cover we haven't had too much difficulty extending it. That said, having the tool not break things during upgrade has been an issue, and requires full regression testing. But that holds true of upgrading any UI framework up a major version, so isn't so much a no-code issue.
We need more videos out there critical of low-code! Thanks for covering it! Our company is convinced Power Platform is the way forward. It kills me inside.
A client I work with has went full low-code in some of their mission critical apps and nearly all feature requests to that team now fall in that 20% that is near impossible to implement. As a result every feature that needs a change from that team takes 3 quarters to get live due to their huge backlog. Half of their changes also have bugs due to all the hacks they had to do in the low code platform to get their requirements working. Despite this they keep expanding its use due to the sunken cost fallacy around licensing costs for the tool. It's an impressive train wreck.
I’m a fan of Googles Appsheet. It's useful for building small tools to solve problems at work or make our processes more efficient. I like that it connects easily to a variety of databases including Google sheets out of the box and has enough customisability for a work app. It also makes it easy to set up automations when data is added or changed.
A big gotcha though is that if you want to build in logic then you have to learn how to use their "expressions" which is like it's own language
This is a great point. There is a place for low code tools but if you think you can do everything with low code then its just a recipe for disaster.
For example, Appsheet is really good at abstracting away the Google Sheet complexity but if you want to build a product on top of it then its going to be impossible.
@@wlockuz4467 agreed, I wouldn't want to build anything I was selling on the top of it unless it was a simple use case.
To be fair I think appsheet is aimed more at internal business use case than anything else and it works well for that unless you want to do more complex logic.
Thanks bro, didn't know about Refine until now! Gonna be a HUGE timesaver! Coming from PHP and jQuery, this new web dev shit is so awesome!
This depends on how the low code tool is implemented. If it's implemented in a code first way then extending the code with a good set of code extension apis is incredibly easy. If the low code abstracts away all the complexity and is more complicated configuration, then this absolutely applies.
Low-code tool with a code-first experience is an oxymoron.
@@wlockuz4467 yeah, he's almost defining what a framework or library is
any examples of such a tool? I've only ever seen 1 - enso - and it's still far easier to write javascript than it is to use enso.
@@SancarnI might be wrong, but I think Unity and Unreal have this. Their graphical programming (technically, not a code!) is just a different representation of function blocks. I think LabView is something like this too, but I only used the graphical part there.
ScriptRunner Connect is also one of such tools. It's built for integrations and automations between business apps and takes care connecting these apps and makes working with APIs easier, but for the actual business logic you're required to write plain old JavaScript/TypeScript.
I had a job that used Salesforce heavily for the backend, it was written in what was legacy code about 5 years ago and was so large that it couldn’t be migrated using the tools Salesforce had created. That had so many custom fields that the naming conventions were not very intuitive. So, I definitely had the same experience with a tool that was sold as a drop-in service, but ended up needing to be heavily modified to the company’s use case.
As a side note, Salesforce has a nice language, Apex, for their service, but the weird thing about it when I used it: it is strongly typed and only built-in functions/methods had generics and the user couldn’t define them. This made for a much larger code base, since there was multiple copies of the same utility code but with different input types.
Very true about finding developers to maintain a low code system , like Out systems for example. It's astonishingly hard. Many enterprise companies are feeling this.
What a joy if you rely on a low-code solution that gets suddenly unsupported. Not that they went out of business, but just didn't bother to create an update tool to convert from old no-longer-supported version to new. Not a tiny company I am referring here.
I have seen an extreme case of this where a company was using Appsheets with 100s of hacks and terrible database design to make it work
Low code means just having an UI for tweaking paramterers of your js components. If business didn't aim to lock you down for milking, there would be source access to add your own components.
Every low code client I've worked on has asked for custom behaviour
It's fine while there are devs to maintain it but the domain specific knowledge and general complexity to the solutions makes me uncomfortable for the future
low-code tool usage still comes down to these tools being passive traps and whether users are capable of: #1 algorithmic thinking then #2 translating that thought process into a custom PROPRIETARY abstraction.
The right use cases are companies that have done ,and continue to do, the work into documenting processes in an agnostic format to describe ongoing automation needs ; this can get the easiest benefits while softening vendor lock-in.
But almost every other use case makes a severely under funded mess in a niche ecosystem they have no control over.
I have never seen one of these tools help it's customers document their process needs in a platform agnostic way.
So the trap is set because the pain of getting out get's higher and higher because there's no actual source of truth.
I am in the position of having picked up a low-code side project while being a fullstack engineer by day.
I hate that side project with a passion, and constantly fantasize about rewriting it with a 'full' programming language. Great video thank you.
rewrite it using a different no-code suite lol.
i learned to code because of shopify actually.
i found building websites with purely ui very frustrating and i ran a business where at the time i had to upload and 100 pictures a day with some processing. and shopify only let me add items one at a time. this led me on the dev journey im on today.
thank you no code for making me want to code
I have used one of those no/low code website builders just because i couldn't be bothered to write all the underlying boilerplate code. I then took a copy of said completed website and deployed it on my own.
imagine voluntarily turning a green field project into a hacky mess
I think the Webflow argument definitely is valid from a developer perspective. But my wife is a UI/UX designer and she’s tried to dabble in software development and it just wasn’t sticking for her. So tools like webflow have made it easy for her to extend her freelancing services.
I come from all code to webflow. But sometimes i migrate the webflow website and customize it, for me i only look at the specific use case for each client
Meaning sometimes i will all code the website
@@Applecitylightkiwiyeah, that is the best of both worlds.
Webflow is not a website builder akin to Wix or Squarespace, it’s, basically, a visual code editor.
I think using Webflow as an example was a terrible choice in this video - something like, again, Wix, Squarespace, heck, even Framer, makes way more sense.
Used Retool to validate an usecase for a client (took 2 months). Usecase is valid, we're now rebuilding it in React as fast as possible to run away from retool and the shenanigans we had to do to make it work for the required custom logic.
Non engineers just don’t want to do engineering. They also hate engineers because they can’t run their business without them.
I can't remember what a company I was working at looked at 25 years ago, but we came to basically the same conclusions. It allowed us to do the easy stuff with a little less effort. However, it made the hard stuff really difficult unless you could recast the problem into concepts used by the low code / no code vendor.
Updates were problematic for several reasons. However one of the reasons we saw was that if a vendor had a tiered product, some vital feature would migrate from a less expensive tier to a more expensive tier. All of a sudden the infrastructure costs skyrocketed due to the vendor's marketing decisions. Since you couldn't get support on out of date versions, the company was trapped unless everything was rewritten. Welcome to expensive vendor lock-in.
No. Just no.
I have to build a Microsoft Power App for work and I'm dreading it
PowerApps is the devil. I am an Angular and .NET dev and had to create a fairly complex app in PowerApps and I hated every step of that journey.
Oof. It'll be a terrible experience.
Best of luck!
Our IT Strategy department has deemed all our orgs apps going forwards will be built with power apps... Part of me wonders if anyone in that department has any dev experience...
@@Sancarn I think the issue with Powerapps, is that people believe it can do magic. I use it personally only when I want to manual data from other departments efficiently, I make sure to put the constraints to they don't make the regular mistakes they usually make when they're doing their manual process, but for full fledge software solutions? That will be a huge error.
@@joanacosta5385 I think it's also a success of marketing and consumer ignorance. I recently found out that we out-source all our dev to overseas. We have no in-house dev. When I said "Some of these controls are really complicated, I don't think you will be able to build X in PowerApps" the BA responded with "It depends how creative the dev is". I was astonished:
1. The dev shouldn't need to get 'creative'
2. 'Creative' solutions will end up unmaintainable, and it's the BA's job to choose a solution which prevents something being an unmaintainable clusterf*ck!
At this point we might as well build our system in a turing machine...
Don't get me wrong, PA is great for simple CRUD apps, I use it myself for that too. But some people really need to just get informed... It still blows me away that a BA, with no dev experience, can choose the technology a dev will use to build some software 🤯
Low code tools are actually how I got started in development. I learned Robotic Process Automation tool called UiPath. I would say that article is pretty spot on with regards to what is sold to upper management and what can actually be done by these tool. I think the issue really is where the origination of decision of a low code solution comes from. If its upper management then they are very likely to put the proverbial 10 pound weight in the 5 pound bag. But if it comes from developers it has a good chance of working out since low code really shines as a stand in for lack of technical skills in an area (such as Im a backend engineer, and dont want to create a bespoke front end, so I use an open source library to generate the front end for me).
What generating frontend open source library do you use then?
Yeh same for me I also started as low code uipath developer and did some vba before. This video is spot on about upper management. They really have no idea what this thing is like and what the company actually really needs. But they are obsessed with the idea of how uipath gives them power to screw their developers, dispresct them and relegate them something like "support staff" .. of course the results are abismal and the developers are being blamed for that .. uipath and Rpa is utter trash 🗑️
@@paulholsters7932 python library called dash. Without knowing anything about front end development I was able to throw together a website
No/Low-Code is like wanting a custom painting but then using an image from the internet and printing it using your office printer.
At my wife's work, they use excel for EVERYTHING. Linking Excel sheets together to make summaries of data is terrifying. They even click without looking when an error pops up. It's the devil
And yet they do not change this way of working do they? Couldn’t that be a sign it kinda works? They surely don’t bother clearing budget to replace this with a custom application right? What does that tell you? It tells you custom development is way too expensive. And that is exactly why low code exists.
Sure, quality software costs a buck. But building a low-code solution and throwing it in the bin 2 years after it's born because no-one can't maintain it isn't the solution either. Better to not build a software solution because it isn't worth the investment than throwing away your money with low/no-code imo.@@paulholsters7932
@@paulholsters7932 fun fact. They are changing. They're using a new system I've created in my spare time and they're loving it. It's fun to be useful
@@paulholsters7932It's why Python exists...
I worked with a program called Creatio. It is basically C# backend, JS frontend using EXT.js. Half the time the client wanted features, I ended up breaking out of their abstraction. This made it harder to maintain. Not to mention that every time you added a new C# class, you had to compile the entire application which takes 10 minutes. Literally the biggest PITA to develop on. I ended up writing C# in Vscode and then copying it into the application so I didn't waste my life compiling forever.
There is another negative thing about no code/low code tools that is not mentioned in this article: the collaboration is almost impossible. These tools are not suitable for large teams. The basic concept of these tools is not only to write less code (or "no code"), but to do it with small teams as well. To be more specific, with only one "developer".
Also, the lack of testing and verifying your business logic is a huge no no.
These code averse folks are trying their best to solve problems that devs solved decades ago.
Webflow huh? Just signed up! Looks great! Thanks for the suggestion!
Same experience or rather experiences since it's happened many times. It's nearly always better to build your own stuff that exactly fits your business. The only "but" is that you need to have people who know what they're doing else you can end up with something that does not work at all. Don't introduce code that solves problems that you do not have. Always have the bare minimum. That way it's always maintainable, readable and performant. And to have that bare minimum it's always starting from scratch.
The fatal mistake was having an exec that DOESN'T know how to code choose the tech for the people (that DO know how to code).
I think working in RPA, visual representation of logic is definitely a keeper and yeah, as much as the work I did, 90% of it could be handled with the standard in box tools but then you would need to work with something so unique, that you were breaking out a custom code block (C#) to do it or delve into the insanity of trying to do it with standard in the box tools (I managed this once or twice but the performance was god awful and the solution insanely complex)
Yeh Rpa is utter garbage really at the end of the day
I've worked for a company with its own low code platform to build chat-bots. It's really flexible and supports lots of things out-of-the-box (scripts with JS, http requests, logging...) and extensibility, but BRO that thing is slow after the "really small" threshold is crossed
I would love to hear your thoughts on Wappler, it keeps your code entirely open source and is a lot more of an actual "medium code" experience. People on their forms preach learning about database design and bootstrap before beginning to use the tool. The only unfortunate caveat is that you are forced into bootstrap.
Yeah. I'd love to hear what he has to say about Wappler. It's definitely one of the most professional 'low-code' tools around and there's no proprietary code lock-ins to keep you a slave to the tool but, everything has its cost.
My issue with Wappler is it's really aimed at people who know what they're doing. It's not for dev newbies.
While the docs are substantial in themselves, they're meaningless to people who don't know when and where to use each instruction. People who have asked for tutorials on how to build a facebook or twitter clone, a blog or admin portal from scratch -- as a way of learning where to start -- are ridiculed by some in the community, so they leave.
engineer opinions are very often dismissed by management
Every CIO or CTO in the world should watch this video. They got scammed by low code marketers.
I was tasked with making an admin panel for a startup during one of my summer internships. I spent more time wrestling and figuring out how to get stuff to work with retool that I believe I would have spent just writing a nextjs app.
Not to mention we had another intern design the whole admin panel, only for them to find out that 90% ofwhat they designed would be impossible to implement in retool.
Classic "I can make it in illustrator, why can't you just plug it into your program" moment?
We could not agree more!
What scares me most is the name "low code" ... when your boss thinks that you are incompetent struggling with that because it is supposed to be "low code" ! ... yeah whatever.
Low code is interesting but it needs to have some kind of "full code button switch" when needed !
I like the sweet spot between ui building tools and generated code in react or flutter . as long as the generated code is reasonable of course
Webflow now has devlink which somehow has not yet become popular - which can be exported into next js project perfectly.. add supabase and you have a strong techstack.
Another low code implementation that is powerful is relume for pre built webflow components + webflow + wized + xano
after 4 years of no-code/low-code i consider the number higher than 20%. Poor documentation, high learning curve, over-complicated integration, unforeseeable costs. Its more a 60/40 to me.
I used to be a SharePoint engineer. Yeah that was fun..
LOW-CODE/ NO-CODE requires some kind of design knowledge, and selling website templates that do a bit more than squarespce or wix is pretty profitable. I do both, coding and no code tools, like flutter and flutterflow or html and webflow/framer
This, each case is different its good to know both depending on the use case and client needs. Sometimes you have to code and use nocode
The mark of a good tool that leverages heavy abstractions is that the abstractions degrade gracefully when you need to get underneath the hood. They also need to fit into a standard CI/CD pipeline with testing. This last one seems like a hurdle that many many tools just completely bomb at.
I had a client who wanted a dashboard for their fans to leave game submission ideas (they are a games company)
I thought about building it the traditional way, which I would have charged about $5k for. Then I realized that I could just create a Notion database, add a form builder integration, and ta-da I had built a solution that was far more fully featured than what was originally planned, since Notion is so customizable. It took about a day to set up.
Can you explain what you mean by data-processing that is critical @ 6:17?
I am a low code specialist. It pays very well. I have a lot of free time and virtually no competition. I mostly write JavaScript to create custom applications and fix bugs, its not even that hard.
I work with webflow daily, and you're on point - Low code was a mistake for the most part.(or atleast the way it was marketed )
if the customizations to your low code app starts reminding you of rube golberg machines its time to ditch the low code solution.
I was once in an interview trying to land a job, The PO(higher up person) was there, HR & the Team Lead / PM was there. The Lead/PM tried to "explain" me how good Low Code is and what they are actually using for it. The "explaination" felt more like a *pitch* for the higher ups. I felt so confused over the entirety of the call, that I asked why LowCode? And the guy told me again, that its the facto standart nowadays in the businesses. I didnt like it when it got popular and I still dont like LowCode. It gives the false sense of security, maintainability & easy of use....
Absolutely agree about "making it so that non-engineers can engineer" point.
Hands up if you remember the pain of Lotus Notes! Some things haven't changed.
12 seconds in the video and they talk about the "low maintenance cost" of low code 😂
Low code is simply unmaintainable. You just throw it away and start over. So, technically, they're right ... technically.
Completely agree with you on this..
It's one thing to assemble Ikea furniture. Its another to combine Ikea furnitures together build a Frankenstein's monster of a furniture piece. At that point, just learn carpentry well.
“There are no solutions, only tradeoffs”
The full thing is:
At the pareto frontier, there is no such thing as best. Only better at the expense of something else.
At work we use low code to manage a bunch of bots that send notifications and automate templating documents in Google docs - decent enough
One of my old workplaces was obsessed with one of these. It was sold to them as a way to digitize all their paper forms into digital forms.
The reality was each form and its real life processes were just not good 1-1 with these forms and that last 20% as we as the specialized engineering team who were these low code experts used thousands of hours every year.
Oh - we use Decisions and it's a total nightmare. The documentation is terrible. It breaks on the JSON schemas I built from zod. It took a day to figure out how to do a simple Map[String, Int] lookup. The training videos leave you with more questions than answers, and it's thousands of dollars per year per server. Wild.
Hi Theo, how do you invest in platforms like Refine?
really wonder if there is any long lasting software that was built with low code tools (simple eshops blogs etc. etc. don't count)
As a founder, this exact scenario has happened to me. Built in webflow but having to rely on external modules which are x% of my income skimming off my hard work.
This is why I like NodeRed. It's lowcode until you want to get into the details, where you can execute arbitrary Javascript natively in the engine, or pass info with RabbitMQ, MQTT, OS pipes etc into other binaries.
I use it for prototyping all the time, and it even runs a lot of my home automation systems.
This hits too close to home. All of these points are things that I regularly mention when arguing for having the possibility of custom solutions. I regularly get to hear from execs that we don't need the cloud because we have this or that low code tool that can handle things just fine.
The other thing that adds to this nightmare is that powerful platforms like cloud platforms are often closely guarded and monitored. I had to go through a 3 month process just to get an AWS account for a project. Low code platforms are often available to every single person in the company. I get brought in on a weekly basis to teams that have been pushing some low code solution to its absolute limits to the point where nobody understands it, nothing is maintainable and now they need someone to help them. I regularly see people do things that we wouldn't have been allowed to do if we did things properly, like things that are totally not okay regarding data privacy and compliance. But the companies don't have any idea people are abusing these tools because they're just blasting these low code tools out there without any sort of governance whatsoever.
My main issue with low/no code is that basically every video on the subject end up being a sales pitch. Including this one (open source or not).
I had some training in a Low code solution called Pega a few years ago, absolutely hated it, it felt very niche after the more broader training learning OOP, backend and frontend development and I didn't want to go into a job that would down the road just be "more Pega" I could not imagine a less appealing developer role if I'm honest.
Most programming languages and frameworks make sense, but the way everything worked in Pega was just utter proprietary bs, it seemed like in a lot of cases it actively made development of simple features harder
We were pushed to low code at my work, by non technical people. It's not a great solution for an enterprise scale application.
FileMaker, Squarespace, Shopify, ugh. Loathe entirely.
THANK YOU! a few months ago i joined a company which has one piece of software, made with filemaker. i'm by no means a senior dev but even i could see from the start that this shit is just horrendous. Also that you freaking have to pay for licences to even use this piece of crap is astonishing.
A few software engineer friends are getting paid to do low code (or at least they were when I last saw them). One guy even managed to reverse engineer the proprietary format (I think XML under the hood), hacked together a quick script to convert it to some prolog statements, and now is writing prolog static analysis rules 😆
Crazy right. Deep down he must know that is proper bullshit.
Well at least we could get a job to maintain that mess or God forbid get rid of it.
Watching this after just having a client with a large-scale multi-function cms switch to a no-code using marketing firm. I'm just gonna let them cook 😂😂😂
Most low-code platforms I used are crap, and the things you mentioned are spot on.
Can you make a video about how your investments work. I am not looking for a finance channel video, but more, if one us (your viewers) would like to invest in a project, what would that process look like?
- How do you approach a project you like?
- How to estimate the amount of money that you want to invest, and how do you know its not too low and disrespectful?
- How do we know if a project is seeking investors?
Etc.
Thanks
Low code is like the stricly software versio. of Arduino. While Arduino convinces you that real hardware is too complicated, low code convinces you that real programming is too complicated. Both tools create dependency.
Did you know that the 20/80 split is a common occurrence that even in your house there is a 20% of your floor that has 80% of all traffic, that even applies to each room. It is always good to find the 80% that takes 20% effort to have the biggest impact, because after that you’ll have 20% left that is 80% effort.
Also double feelings about low-code. Vendor lock-in and ok having code from graphs looks great when small but spaghetti allso when big. I could imagine you could the simple staightforwrd ui with low-code and other parts with real code if it integrates well.
this type of shit has made me realize that companies just dont want to hire developers and if they do they bring foreigners in on work visas that write bad code that needs to be fixed that they can pay less. I honestly think the bad economy is being used as a convenient excuse to not hire devs especially for entry level roles and low code alternatives like web flow or wix is a cheap alternative that they'd rather use even if it means a half baked project.
I recently spent a couple of months working with no code or low code platforms. The majority of my work is with startups, quick platform deployments and sites to evaluate the market for ideas. There is really nothing better than tried and true wordpress. With the addition of htmx / custom plugins there really is not much that it cannot do, readily available, supported and hits 90% of most any startup project.
Low code… high tech debt
As a no code developer that I encountered these exact issues. I mostly customize things with chatgpt.
Next project will be a full code solution.
2:25 it's the subtle stuff in a tuber that wins ya over! < 3
4:50 My current job is working support for a proprietary no code rules engine company. This is the exact pitch our sales team uses. I have yet to see a single non-tech person fill out a support ticket. They ALWAYS just delegate building things out in the platform to their tech team anyways.
Low code.. is fine for small projects and e-commerce ( Shopify). I’ve used them all. I mostly develop on many open source platforms , even some open source can break your code.. depending on the massive refactoring they might do in a considerable update.
Coding can be a nightmare sometimes when it comes to management requests.. best is when a decision they make requires development.. but they want now.
currently pulling my hair out with a combination of slow-as-hell retool *and* salesforce
point I think you missed is trying to do that last 20% is multiplied in difficulty in a low code environment, because the parts you need to touch to make those things happen are roped off. In fact they're often what would have otherwise been very easy things.
It is also easy to lump low/no-code and managed solutions in the same bucket. There are mny cases where the no-coding part is the least thing u should be thankful for - it’s the systems design, insfrastructure, and domain-specific stuff packaged with it that are worth much more than getting someone who ‘can code’. I find ‘low/no-code’ a bad term as it tends to limit the value prop of innovative solutions, as if they would be achieved anyway if one just knew how to code.
The low code tool at a certain company made me strike that company off my list after one internship. Boy, that thing is a UA-cam fire.
Got to disagree with the Shopify characterisation. You can quite comfortably be a full stack Dev and primarily work with shopify, building apps, extensions etc and frontend is definitely not a nocode type environment. Official support for Headless etc. It can be quite a fulfilling role and is a good specialisation.
Yeah 'Shopify Engineer' here lol. We have a custom stack even on liquid builds and the app building is pretty open ended. I've had to work with all the popular stacks/frameworks at some point. I guess there are a few theme-only devs which he may be referring to but even then liquid + html/js/css isn't really a far departure from fundamental web dev.
I think he is talking about shopify devs that just install a theme, some apps and that's it, we got some people like this where I work.
@@gustavocampos1593 Unfortunately I have to do a fair bit of this at times :(
@@StanLey-dt4lf Thats it. Its no code for users but lots of people think Shopify is like Squarespace/ Webflow but its actually quite a rich ecosystem for Engineers.
Probably the only man in the world who can run a popular coding channel and also lazer flip.
In the past I've had to fix low-code generated codebase that used HTML IDs like classes and classes like IDs among a myriad of other issues.
💯 true. I’m experiencing this problem irl.