Access maybe never took off, but wasn't Visual Basic super popular at one point? I think you could arguably call it low-code given its use of a WISYWIG UI builder and eliminating a lot of the boilerplate of creating a windows application. I definitely think you could call Excel a low-code tool too. Probably SAP and Salesforce as well.
I'd argue that it was Cloud and improved monetization opportunities rather than ChatGPT which made the current iteration of low/no-code more popular, although ChatGPT certainly helps (I'd argue that past a certain level of complexity/code sized, LLMs also stop working properly). Most low/no-code tools include hosting, which is a must for anything meant to be served on the Internet, and before public cloud took off, this was a lot more complicated to offer for free or a low-cost. There's also a much bigger market for SaaS and it's easy to take payments over the Internet now, which incentivizes semi-technical to build software to be served over the Internet rather than their desktop (and also makes it easier to build the low/no-code tools themselves).
> until you REALLY need a programmer
This is the thing I was getting at, there are a lot of people who know how to program but are not super-skilled at it, and I think a lot of software engineers don't get this. Think back to when you were just learning to program and how frustrating it probably was to do "basic" things like set up your dev environment or ship a website for the first time. For the people using low-code, it's not that they can't do these things, it's that they'd have to spend a lot of time looking things up and learning how do it, which is a major time investment (and frankly require a lot of persistence/patience which I think professional SWEs self-select for, because the learning SWE learning curve is brutal for most people) that they'd rather not make. You can argue that past a certain degree of complexity that the investment is worth it, but in many cases it probably isn't, and in others it's possible they'd not have made it to the point of outgrowing low-code at all if they had to do everything from scratch.
It's not much different from how we as programmers use things like Cloudflare or AWS Lambda. Could we run and manage our own DNS or scaling? If we really need to learn it, we probably could if we set our minds to it. Is it possible we could outgrow those tools if we start on them? Definitely. But even when we know how to implement our own scaling or DNS we might still reach for these tools just because they're so convenient to get something working quickly, and most of the time we don't outgrow them anyway. Low-code is pretty much the same thing except it has a lower technical barrier to entry (eg you don't need to know how building/deploying software works) and often is easier to outgrow.
I'd argue that it was Cloud and improved monetization opportunities rather than ChatGPT which made the current iteration of low/no-code more popular, although ChatGPT certainly helps (I'd argue that past a certain level of complexity/code sized, LLMs also stop working properly). Most low/no-code tools include hosting, which is a must for anything meant to be served on the Internet, and before public cloud took off, this was a lot more complicated to offer for free or a low-cost. There's also a much bigger market for SaaS and it's easy to take payments over the Internet now, which incentivizes semi-technical to build software to be served over the Internet rather than their desktop (and also makes it easier to build the low/no-code tools themselves).
> until you REALLY need a programmer
This is the thing I was getting at, there are a lot of people who know how to program but are not super-skilled at it, and I think a lot of software engineers don't get this. Think back to when you were just learning to program and how frustrating it probably was to do "basic" things like set up your dev environment or ship a website for the first time. For the people using low-code, it's not that they can't do these things, it's that they'd have to spend a lot of time looking things up and learning how do it, which is a major time investment (and frankly require a lot of persistence/patience which I think professional SWEs self-select for, because the learning SWE learning curve is brutal for most people) that they'd rather not make. You can argue that past a certain degree of complexity that the investment is worth it, but in many cases it probably isn't, and in others it's possible they'd not have made it to the point of outgrowing low-code at all if they had to do everything from scratch.
It's not much different from how we as programmers use things like Cloudflare or AWS Lambda. Could we run and manage our own DNS or scaling? If we really need to learn it, we probably could if we set our minds to it. Is it possible we could outgrow those tools if we start on them? Definitely. But even when we know how to implement our own scaling or DNS we might still reach for these tools just because they're so convenient to get something working quickly, and most of the time we don't outgrow them anyway. Low-code is pretty much the same thing except it has a lower technical barrier to entry (eg you don't need to know how building/deploying software works) and often is easier to outgrow.