Drupal 8 is not a gentrified neighborhood, it's more like the ghost cities in China: clean, modern, well-designed -- and empty.
I'm digging this analogy.
It still has random errors, our module stopped working because of a cache corruption. How does that even make sense?
Clear cache is now needed to get twig changes to propagate. Man. That is so slow and so pointless.
You're not going to get an accurate picture of the CMS market here on HN. Most folks here are developers...PHP CMS's are not what they are interested in. The prevailing opinion on HN is something like "CMS's suck, you should build it yourself with node/React" or whatever.
If you need to build a content website, and the information architecture is complex, Drupal is a very strong choice. I haven't use D8 yet, and I would not recommend building a high-profile project on it yet, unless you really want to be an early adopter. But D7 works well and will be supported for a long time yet to come.
If your info architecture is fairly simple, take a look at Wordpress. You'll have an easier time finding low-cost vendors than with Drupal. Drupal is powerful but expensive.
- It's easy to find case studies to help explain and plan your project.
- It's easy to find vendors who can build or support the technology. That makes it easier to run competitive RFPs, and fire/replace vendors without having to rebuild from scratch.
- You benefit from a security program that has a lot of inputs--a lot of bad guys trying to get in, and a lot of good guys trying to keep them out. The code is battle-tested and frequently patched.
- You benefit from a community module system that has a lot of inputs. Chances are if you want to do something, there is a module to do it, or at least get close.
- You can more easily empower "tech savvy" employees to do pseudo-development tasks like build filtered lists and manage content workflows. It's easier and cheaper to scale tech-savvy employees than full developers.
- There are enterprise-savvy support options available.
- It's easy to find hosting.
Developers, particularly those who want to innovate and build new tools and products, tend not to value these sorts of things as much.
I mean, there are downsides to SharePoint: it's commercial software and there's more friction than with Drupal. But I'm not even sure how relevant that is anymore in today's IaaS hosted world -- I can go on Azure and deploy a Sharepoint server in 5 minutes from an image and the licensing costs are baked in.
But you're right about developers not valuing these -- CMSes are for established businesses that need a technology platform to manage their messaging across multiple channels. If you consider yourself a product hacker, then something like a CMS isn't likely to be interesting to you.
Other than that, people have moved on to all sorts of CMS's:
* If you're looking for something in the spirit of Drupal (most configuration done via dashboard interface, very flexible content modeling), try ProcessWire (http://processwire.com/) (free/open-source)
* If you're looking for something more designer-friendly and have a budget for the project, Craft CMS (https://craftcms.com/) is very popular these days.
* Also designer-friendly and less complex than Craft (and less expensive), so it's better for simpler sites: Perch CMS (https://grabaperch.com/).
* For the most easy-to-use editing interface, I recommend Concrete5 (http://www.concrete5.org/) -- my personal favorite (free/open-source).
* If this is a super basic site though and you just want to use an off-the-shelf theme and let the client manage it as much as possible, you're probably best off with a hosted solution such as Weebly (http://www.weebly.com/) or Squarespace (http://squarespace.com/). These have monthly fees (but comparable to decent web hosting).
If the site is not very big, consider Wordpress. For one reason: the community and with that mainly the pulgins available. For example there are WP plugins that do the whole conversion for mobile for you or do SEO for you. It really depends on what the site should do. But Wordpress can be easily the best solution.
Now, all that is moot, to a level, because all these CMS have nice importerts/exporters, so you cannot really make a big mistake that will sink everything.
If however, the site in question is bigger a site than just "a website", consider Typo3. It's a beast. It's cool. It may not worth the effort for a small shop, but if the site has enough size, try Typo3. Beast!
If I was building a new site that made sense to use a CMS and not build something from scratch, I'd still probably go Drupal 7.
After ditching our custom CMS software for Drupal, we find we spend more time making Drupal do what we need than just starting a site from scratch.
Any company with a decently sized content (producing) team is going to need to use a CMS.
A lot of folks are also using Hubspot and Salesforce to create their sites...Gotta love vendor lock in.
Things like Jekyll or Hexo are straightforward and easy to work with, but because static sites are really just tools to transform markdown into HTML, something like Metalsmith could probably even handle online shops.
Asking people whose job isn't "building websites" to edit text files, run some sort of compiler on the command line, connect to a server via FTP, or deal with version control is a total failure mode.
But if a developer (or an especially adventurous designer) is going to be editing the content of the site as time goes on, then static files or a static site generator can be a great option.
To avoid this bias I was doing a little research for statistics. I checked the usage of Drupal core  and I found it somewhat misleading (e.g. there is a significant jump around the time of the release, so the statistics on such short term must be skewed due to test installs). Instead I looked at ctools, which is a dependency for several contributed modules and is the top used module . This shows significant growth since the start of 2016. So it is growing, people are using it. But most importantly, the numbers for Drupal 6 stay the same. So it seems that people who have a stable website are happy with it, and new users are likely to jump on the Drupal 8 train.
Also let us not forget the same voices were heard when Drupal 7 came out, coupled with the unexpected lag of contributed module availability. With a lot smaller API change, people found they have to wait 6-12 months after D7 was released to start building Drupal sites. Contribs, e.g. Views has released an RC for Drupal 7 on June 17th, 2011 , when D7 came out on the 5th of Jan, 2011 . Views is just one of the important contrib modules (now in core), and people had to wait 6 months to use it!
The contrib upgrade curve seems similar, if a little bit faster for Drupal 8. But having most of the important contrib modules in core, this impacts new builders less.
About the state of the spaghetti: frankly I don't think that customers care too much about it (after all, Wordpress does own a significant share of the market ;)). Developers do. Customers care as well, but mainly to the extent of maintainability and cost implications. And I think with personal bias aside, Drupal 8 is much better on this front, than Drupal 7 (although there is still work to be done, procedural code to kill), it does deliver a framework, which allows a wider range of developers to start working with it, using familiar patterns and frameworks (PSR-4, Symfony).
Long story short, I don't hear the swan song. But I'm happy to keep my ears open :)