I think low code can really help while doing broader level coding, e.g. service, API, component definitions, and also mapping out communications/ service calls. When doing detailed data manipulations, data mapping can help, but when digging deep, the code is easier. If we can switch and use mixed approches between code and low code, it can improve the productivity in my opinion.
There are a lot of things which can be achieved by No-Code / low-code platforms, which otherwise used to take a lot of effort if you start from scratch using coding. Examples could be Customer Portals, Internal Apps for companies, Workflow applications, dashboard etc, employee/vendor onboarding etc. So given the need of such use cases, it is no brainer that these platforms are there to live long and thrive.
Only if you are selling low-code / no-code tools and get a lot of enterprise clients locked in on your tools.
Most tend to be too limited and do not last, but some like spreadsheets and visual databases are staple business software.
Hot take: no, everyone hates UML. Visual programming paradigms are either too application specific to be generalized or too abstract to be useful.
Microsoft has PowerApps, Google has Appsheet, AWS has Honeycode, Salesforce has Lightning- all low code development tools , and Gartner Forecasts Worldwide Low-Code Development Technologies Market to Grow 23% in 2021 to $13.8 billion
There is a future for specific use cases. For simple CRUD apps, workflows, admin panels there is a future. For SaaS applications, I am not as sure due to its feature-evolving / bespoke nature.