If you are facing minimum of £27k tuition debt in England which AFAIK gathers interest from day one I would probably skip it. And it's even worse in the US. As far as webdev goes, you could do a bootcamp and be earning money within 6 months to a year. That's money that's hopefully going into your pension fund and earning interest while your peers are still in class. Or you could learn to code using cheap and/or free…
My concern would be whether people who take the bootcamp or self taught approach would have the grounding in engineering and CS theory that someone with a degree has. I tend to feel that under-emphasizing these things are why the software industry has so much of a problem with poor designs, low code quality, and missed deadlines. Maybe it would be possible to create a two year degree that would fill in those gaps.
I think there is too much weight put on CS theory but it is valuable if you already have it through a degree.
Outside of FAANG and similar unicorn growth kind of companies nearly nobody else needs a new X or a different design of Y. I'd even go as far to say most of the time we don't even need to understand the internal of the tool itself.
Gut feel says upwards of 99% of companies just need existing solutions combined. I'd say 20% of the industry itself thinks they're actually apart of the 1% that are truly facing a problem nobody else has.
I've seen companies suffer from too much CS theory. They get caught up in the technical problems and the most ideal solutions instead of focusing on providing customer value.
We hear this through the many talks about premature optimization and reinventing the wheel but in the real world I hear a lot of individuals calling the work a hack when it isn't the most optimized solution.