Hacker News Reader

Comment by htormey | Hacker News Reader

“You can see the telegram team pulling this off because you aren't in the crypto space. They are claiming to be able to solve open research problems that some of the best minds in the world have been studying for years, and don't have an answer to.”

Do you want to be specific? I suppose one question is do they need to pull off everything they claim in the whitepaper in order to be successful? Do you feel the things the Telegram team promise are an all or nothing deal.

If they just had hundreds of millions of people using Telegram for payments and perhaps filestorage/bots backed by crypto could they do that without all the things promised in the white paper? Would the market consider that a success and would that be reflected in the currencies price?

View on HN

>Do you want to be specific? I suppose one question is do they need to pull off everything they claim in the whitepaper in order to be successful?

Not OP but the whitepaper claims that it wants to produce a decentralised blockchain product with the same efficiency and scale of credit cards or similar payment methods.

Now I'm not terribly knowledgeable as far as crypto currency is concerned, but as far as I'm aware that's a yet unsolved problem because there's a payoff between security, performance, and decentralization.

The article mentions sharding but sharding usually reduces security, you can increase the block size but that will lead to centralisation, and so forth.

The scale / decentralisation trade-off is inherent to almost all complex systems. Overcoming it is a pretty big thing to promise.

The question is how centralized does a network like this need to be?

Not sure if I am reading the white paper correctly but I think Telegram appear to be going with a masternode type solution initially for their proof of stake solution

https://en.m.wikipedia.org/wiki/Proof-of-stake:

“Another form of staking is running a masternode,[9] a form of decentralized server. The main disadvantage of operating a masternode is the relatively high barrier to entry as opposed to staking alone. In order to secure the network, those willing to run a masternode are required to purchase a certain number of coins as collateral at current market price.”

Will a solution like this, which is more centralized than something like btc be good enough? Will people be happy with this level of centralization or refuse to use the service on principle? I don’t know. I do think that at least the earlier parts of their roadmap seem plausible (2018). The 2019 section however does seems rather ambitious and ambiguous.

Which is weird, because when you have a master node there's no need for proof-of-anything.

>The article mentions sharding but sharding usually reduces security

I think if you're careful about it, sharding can be done without impacting security much. Suppose you would split the Bitcoin network exactly in half. Each half has only has as much hashpower, so you can attack it for half the cost. But each side also contains only half as much money, so the potential payout went down by the same factor as the price of attack.

Of course it's more complicated than that, because a realistic attack isn't about stealing money from everyone but from a handfull of large targets. Also, TON is proof of stake and not proof of work. And you have to scale sharding not with money present, but with transaction rate. And the shards (presumable sidechains?) will transfer money between each other, and you have to scale that as well. It's a complicated problem, but I wouldn't rule sharding out as a viable approach.

I highly doubt they can do the promised "millions of transactions per second" in any realistic scenario, that would be a major breakthrough. VISA can only do 57 thousand transactions per second [1]. But I also don't think they need remotely that much. Paypal has an average transaction rate of ~200 transactions/second, VISA of ~1700 transactions/second. At least paypal's average rate would be achivable even with Bitcoin's codebase if you were to take some liberties with block time and size (which isn't entirely unreasonable, the internet got faster in the last 8 years)

1: https://mybroadband.co.za/news/security/190348-visanet-handl...

> But each side also contains only half as much money...

I don't think this is correct, if it were split in half each side would start at the same point with the exact same amount of money. Unless you were talking about the amount of money each half controlled maybe?

But I thought the point of sharding was to enable double-spending which would only really concern itself with the amounts the attacker(s) controlled. Well, I also recall reading about using sharding to destroy faith in the network as a whole but that doesn't seem to apply here.

Well, they could just deploy something more or less centralized and claim they've solved the problem in their marketing, just as they have done with their encryption claims.

Reply on news.ycombinator.com