1) HTTPS. While tl.net doesn't have any storefront or anything else requiring other information than login information, and it is a good idea to have a different password for every site anyway, not having SSL still seems a bit outmoded, especially when you can get certificates for free nowadays. Chrome and Firefox also classify non HTTPS website as "unsafe" since last year october.
What technical issues would stand in the way of adding an SSL certificate to the website? If it's a case of a lot of hardcoded http links a staging and testing environment to test on would likely be a good idea... Anyway, an update to HTTPS should not be something that can be put off for too long.
2) Outdated profile information. The profile still lists the old style of b.net names per region, where not more than 4 numbers are accepted in the code. AFAIK this changed years ago: Seems odd to not update this little bit. As the profiles were regenerated from scratch at the moment of the update i don't see why you can't simply dump those tables and have everyone fill them in again who wants to.
3) Stream management page still lists 'justin.tv', which basically no longer exists. Ustream also has been repurposed by IBM since 2016 for business use, and Livestream no longer offers free services outside of a trial. I'm not sure how much of the code in there (chat embedded) is now pointless. At least replace the justin.tv reference with twitch.tv, and put youtube and afreeca on the roadmap (as far as possible).
4) This is just a guess, but when i open any page on the website, half the http requests are for stat or ad scripts. I'm not sure whether this is intended, but it seems like some optimization could be done...
The above points make me wonder how often the website is updated - on a technical level - and how many people are currently either working on this, or are available to do so. I also wonder if the infrastructure to do it safely (think DTAP) is present (i can imagine that with tl.net's size no one really wants to screw something up).