If I had to take an educated guess OpenChat comes to mind. They create 1 canister per user/chat.
He could burn many in bulk and sell them on a dex.
Respectfully disagree. The premium is there because the companies have the luxury to charge it as they dominate in their niche, as a dev when I see the 99$/year fee for Apple dev license I don’t think “Man so premium”, but “that’s bullshit”, but it’s Apple and it owns a huge piece of the mobile market so i either play by their rules or don’t play at all.
The IC on the other hand is the laughing stock of crypto at the moment, I doubt adding more arbitrary walls will give outsiders a “premium feeling”.
They can buy/sell cycles on the market and I doubt 5$ has ever stopped anyone from doing evil, I know guys who’d spend thousands just to DDOS Minecraft servers. It will stop many from taking the IC in consideration tho.
This is my opinion, if you believe it’s a good idea make a proposal.
The premium comes from the fact that iPhone owners are also more likely than Android users to pay money for apps. It is a really good example of a ‘luxury market’ with a monopoly .
as a dev when I see the 99$/year fee for Apple dev license I don’t think “Man so premium”, but “that’s bullshit”,
You may feel that way but after the sunk cost you are more likely to want to produce value with it.
Last time I tried AWS they made me put in a card while signing up… ETH even requires that end-users pay per transaction which is insane. Using time to learn something or spending money creates stickiness…
I know of 100,000 ‘hello worlds’ that could have been 100,000 burned ICP, which could help increase the price, which could help increase demand, if we do in fact make a luxury good.
This would also make marketing efforts like Hackathons more valuable. 5000 devs try IC and burn at least 5000 ICP doing it. It shows real demand.
For no (3), issue is that pricing in ICP may be predictable in the crypto world, but enterprises who account in fiat will find this cost very unpredictable. At USD 1000 per ICP, the price per transaction will be $0.1 USD which is quite expensive in enterprise terms. To attract enterprise applications we need to keep it predictable in fiat terms.
With these 2 caveats the only downsides of this change that comes to mind are: breaking some blackholed canisters as previously stated and possibly introducing a scenario where long running transactions are rejected cause the fee initially approved is different from the up to date one, so devs will have to account for this edge case.
In my ideal world, you would have a stable fee for a 1-week period, with 24-48 hour notice for the new price. There would also be a grace window where the old fee is still accepted after the new fee has been officially adopted
Thanks for following the thread! so there is a hope that the proposal will not be discarded? We would have to wait for someone to explain the current model as it was thought, which I estimate went through several people before deciding how much a transfer would cost. We still need to know ( if it is possible), the research that was done to determine the price Thank you!
Maybe in a few days @diegop or @kpeacock could ask whoever did it or knows if it is possible to at least share a little bit of the research to find out
My understanding is that we basically set it as low as we figured we could manage without hitting Denial of Service constraints. The cost of each transaction is essentially negligible to the network - the only constraint is the maximum throughput.
One of my first questions when I came onboard was “why can’t we make it free?”
being updated vis nns is not a good thing for a frozened canister. My main point is that the fee of a transaction may be designed as a dynamic parameter in the API…
.
Thank you! 
