"Also if this were to happen [updating the core of WikiTree], it definitely wouldn't be based on MediaWiki."
My personal opinion is that Jamie is spot-on with this observation. I think it's time to retire MediaWiki as the underpinning core application, and to very likely design one that's custom written for the purpose.
I'm not privy to any of WikiTree's operations, and I know that The Team are all well aware of the situation; so I'm not going to say anything new, and some of it may be incorrect (too, it's going to get awfully "jargony" and super boring for all you who aren't web developers). And we all know that keeping WikiTree free for all is a massive constraint to new development. But in my opinion MediaWiki was a handy, open-source option around which to build WikiTree initially, and in truth it's served the purpose admirably. However, after the heavy customization in WikitTree's first several iterations, I believe it became essentially impossible to update MediaWiki as that open-source software continued its own unique growth curve and refinements.
With that, I truly don't believe any substantial modification of the current MediaWiki core is practical; it wouldn't be terribly effective and certainly not efficient. One important factor there is that the principal operational interface--technically a hypertext preprocessor--between the code and the MySQL database(s) that store all the data is something called PHP. First invented in 1995, PHP is used in roughly 80% of all websites globally. That "engine" is also open-source software and managed by the PHP Group, which issues frequent minor releases and security patch updates to add features and improve resilience. In 2008, one of my own websites was hacked via a PHP vulnerability and rendered unrecoverable; the solution was to wipe it clean and install a previous backup.
WikiTree runs on state-of-the-art server and infrastructure systems at Amazon Web Services (this not including the fantastic work done by Aleš on WikiTree+, which processing I believe is done on servers in Slovenia). The hardware, networking, and server software (as contrasted with the core application software) is robust and maintained by AWS. That's not an issue.
The WikiTree core is MediaWiki v1.11, which was released 10 Sep 2007. The current MediaWiki production version is v1.39, released 30 Nov 2022. The last version of MediWiki to support PHP v7.1.x was MediaWiki v1.33, released 2 Jul 2019.
WikiTree uses PHP v7.1.33, the last minor release in the v7.1 branch. The major v7.1.0 release was made 30 May 2019, and v7.1.33 was released 24 Oct 2019. The PHP version in use at WikiTree reached end-of-life in November 2021, meaning that no more patches or security updates have been issued since that time. The most recent production version of PHP is v8.2.0, released 8 Dec 2022.
Given that we've been operating 14 months on a deprecated version of PHP leads me to believe that it simply cannot be updated to a more recent version without significant modification to the core MediaWiki code at WikiTree. Germane to the original question is that as of v1.36 of MediaWiki, released May 2021, it has required installation and use of the PHP internationalization extension (also referred to as just Intl, ext-intl, or php-intl).
WikiTree is also using MySQL v5.7.38. This is the database management engine that, like PHP, sits as an operational layer underneath the application code (MediaWiki and other) and above the webserver operating system. The major v5.7.x branch of MySQL was released 21 Oct 2015. It is still actively supported, but reaches end-of-life this October. MySQL major release v8.0 was published 19 Apr 2018; the latest version is v8.0.31, released 11 Oct 2022. I once created a MediaWiki v1.11 sandbox just to experiment with the core version in use by WikiTree, but long since deleted it. However--and with no level of certainty--I don't believe moving to a MySQL version later this year that is still supported would bring with it the same level of complexity as trying to move to a new version of PHP.
To be clear, there are no alarm bells going off about PHP v7.1.33 vulnerabilities, though some specific ones have been identified, mostly minor. That isn't the issue so much as the fact that trying to do significant modifications to core code that relies on that older PHP version would be a waste of time and money. And I believe, since WikiTree was built by directly editing MediaWiki code rather than with an additive/modifying module approach, that crafting it to enable use of current and future MediaWiki versions may not be in the cards--from a planning and design standpoint you'd need to, effectively, do everything twice: plan and design what operations/functions we need, and then plan and design it again in order to build it as add-on modules to MediaWiki that won't break with future MediaWiki releases. (Many types of content management systems--which is what MediaWiki is--like Joomla and Drupal moved years ago to fairly simple ways that PHP and CSS scripting overrides can be applied via server directories that remain unaffected by upgrades to the CMS itself, and MediaWiki can do the same...I just don't think it would be practical given the extent of the modifications necessary for WikiTree.)
Bottom line is I believe--uninformed personal opinion only--that operating forever on deprecated scripting interpreter engines (i.e., PHP and MediaWiki, possibly MySQL eventually) isn't tenable. Trying to shoehorn desirable and future-editable WikiTree functions and look-and-feel into a MediWiki core is also probably not tenable: basically double the work and MediaWiki would continue to be the least-common-denominator constraining future development.
The press for internationalization of websites is not new. There are various tools and services available to help accomplish this, at least to varying degrees. Sweeping language internationalization likely may never be possible, but of the more than 7,000 languages spoken in the world, over half the global population uses 23 of them; here's an interesting post from Berlitz a few days ago breaking down the numbers for the top 10. The distribution will differ considerably, though, among countries where genealogy is an active hobby and those where it's really not practiced (as we know it) at all.
But I'll bet heavily that engaging any of the popular internationalization tools or services is going to require a website that is operating on supported versions of its core infrastructure and application software. Bringing WikiTree up to date on these core softwares will have to be done eventually, no matter what. On the websites I have that are running on basic shared servers--unlike the virtual private server (VPS) that WikiTree uses)--they haven't permitted use of PHP v7.1.33 since November 2021; the hosting companies force the PHP change themselves as part of server management; if your website can't run on a newer version, it simply no longer runs.
The Lean Enterprise voice in my head tells me that you determine the root cause and address that, that you don't spend time and money fixing or responding to single-issue requirements or pain-points when the larger picture is what really needs to change. WikiTree faces resource constraints, but if planning for, say, a five-year lookahead window, those resources might be better applied to that big picture rather than firefighting incremental fixes and workarounds.