IRC wasn't targeted at children. IRC required effort and understanding to access. IRC wasn't in your pocket all the time. IRC didn't include video and picture taking capabilities built in (one would have to work at it to send content like that). IRC had moderation (admins were often responsive to issues), decentralization (each server was it's own small social network), and local communities (one could figure out the difference between a legit and sketch server); worse come to worse, people could just go somewhere else (most modern social media is tied to your phone number - something difficult for children to change - and the app your friends are on is a limiting factor, there is no running away). IRC was easily anonymous, there wasn't reverse image searches, geotagging, or having friends in your social network leak information inadvertently.
> But yeah, you can see porn on there and child predators can also access it.. so uh, child abuse or something. Lets just ban everything.
That's a straw-man of my position. Nowhere did I advocate banning anything.
I clearly stated that my issue was the harm caused for the sake of profit. Google could hire 30,000 moderators. Tumblr could (have) take(n) the issue seriously and redesigned the moderation tools it has (and it's reporting system). Kik could have improved the privacy and anonymity they offer their users (and maybe work to do something about communities that actively - basically publicly - exploit children). The operators of the fly by night chat rooms should probably be investigated, have a strongly worded meeting with an officer and prosecutor (along the lines of "your service is being actively used for extensive criminal activity; here is a search warrant"), and be arrested if they don't shut down their offerings (or radically pivot away) now that they are aware of the criminal activity they are facilitating (or if that investigation turned up their knowing facilitation of criminal activity).
My preferred alternative of federation would fix the problem by simply being better in many ways.
Your description is also a straw man of what is happening. We are talking about coordinated grooming of children. People who make many accounts (backed by using proxies, buying untraceable phone numbers, etc.), host servers and develop custom applications (e.g. targeted apps like chat rooms, or to scrape social networks for victims), and then use these tools for targeted attacks on victims (armed with personal details, secrets gained from their close friends through social engineering attacks, pretending to be multiple people, and so on). Let alone some of the recent stuff like realtime deepfakes, which I can only imagine how that is going to be used.
The victims here are children. Do you realize how regularly trained adults fail simple social engineering tests? With alarming regularity. The adversary model here is a motivated persistent personal threat. The win scenario is compromising a single child, and they can easily choose their targets. The information available includes full access to a child's social media and friend network and basically an unlimited number of tries (phrasing it like that makes me realize it can effectively be brute forced). We cannot expect children to survive that threat model unharmed.
On IRC you could log out (or log into a different server!), walk away, and play with your friends. On kik if you log out you are alone, if you walk away your life is over, all of your friends are there.