And then USA will have a disadvantage compared to rest of the world with cheaper LLMs and american AI companies will have tougher time surviving on domestic spend alone.
I wonder how's model improvement/dollar invested ratio going. If gains are made by simply spending increasingly vast amounts of cash that's not going to last.
No it's not a lie. It doesn't have ads or a subscription. You do need to login to continue playing more of the same type + difficulty for puzzles. There is an eventual paywall after you reach 25 puzzles + difficulty (so like after you've done 25 easy sudoku puzzles and you want to do more easy ones) for a couple bucks to help offset server costs. But it's a one-time thing. No monthly subscription at all. In all transparency, I've made $0 so far. No one has even reached the 25 free puzzle limit for any of the puzzles on there.
I think the messaging is just a bit confusing. You've said this had no ads and no subscription, and then people see you want them to create an account. If there's no money being paid, and no ads then the next conclusion is either you'll sell data or rug-pull later on. Clearly that isn't actually your intention.
Maybe try changing how you talk about the price a bit on the page. No one's going to be put off by knowing there's a lot of free content, and then later on you have a one-off fee to continue playing. But they will be put off if they don't understand how any of the pricing works, and if they feel like there's a catch you're not telling them about.
I get what you're trying to do, you want to offer something on the cheap and that's great. Just be open about when the payment is needed, and what that payment is. You'll likely get more sign ups from being open about it up front.
CGNAT doesn't have to draw a random ip:port combo out of a hat for every new connection. Nothing is stopping an ISP from implementing it by taking one ip and assigning ports 1-10.000 to customer A, 10.001 - 20.000 to customer B, and so on. Similarly, nothing is stopping an ISP from adding long-lived mappings to an otherwise-random pool which outlive the initial connection.
Some ISPs offer CGNAT traversal by letting you request a fixed ip:port combo via their self-service website. It's pretty much the same as regular NAT traversal, except that you can't freely pick the outside port. And because the number of people who actually care about it is so small: some ISPs even let you request to be exempted from CGNAT altogether! They'll already have a pool of legacy non-CGNAT customers and a pool of new CGNATted ones, so assigning a handful of nerds to the legacy pool to prevent them from complaining isn't a big deal.
> Nothing is stopping an ISP from implementing it by taking one ip and assigning ports 1-10.000 to customer A, 10.001 - 20.000 to customer B, and so on. Similarly, nothing is stopping an ISP from adding long-lived mappings to an otherwise-random pool which outlive the initial connection.
I’m pretty sure that the scarcity of Legacy IP addresses and port numbers (!) is exactly what stops providers from doing that, at least by default. I’ve seen NAT running out of ports way too many times, and shortening of connection tracking lifetime comes with a whole set of hard to spot bugs.
Can we at least stop claiming that IPv6 is "more complicated" than IPv4 when we're making it a requirement for IPv4 ISPs is to create CGNAT management portals.
EDIT: Ok, didn't notice the 8th level because of the UI. This one I couldn't trick in 5 minutes.
reply