My problem with new gTLDs is that it is a limited namespace, and when it's gone, it's gone. The gTLD space is also handy for classifications. If I see a ccTLD,

My problem with new gTLDs is that it is a limited namespace, and when it's gone, it's gone. The gTLD space is also handy for classifications. If I see a ccTLD, I know that it's delegated to a country, and isn't some phishing attempt at gTLD level. Since I'm familiar with the .uk space, if I see a .gov.uk, I know that it's delegated to the UK Government, and that it isn't some phishing attempt at .uk level. This chain requires a "brand" presence at each level. Permitting a free-for-all at gTLD level would dilute this. The point of [hierarchical] namespacing is that we can subdivide them. We once had a single free-for-all area, which we called .com. People say that it's full, so we need more space. But instead of growing down the hierarchy, where there's unlimited space, they want to grow up, where there is just one namespace left. But you can only grow up once. If we didn't have a .com, and we just let people register in the gTLD space directly, then where would we be now? If we want to think commercially, "like a startup", then companies should be able to register a .com and sell subdomains to third parties. They've certainly tried. Why can't they achieve notable success? Because the .com space is diluted free-for-all space where a .com name cannot achieve the level of notability needed, just as what will happen to the gTLD space if ICANN continue on this path. The rationale for adding arbitrary gTLDs is self-deprecating. This route can only be a victim of its own success, just as the namespace problem inside .coms has already demonstrated. And then that will be it.