Segregated Witness and the technology it builds on are overhyped

Like all good lies, they have a kernel of truth behind them. However, like all good lies, they omit or distort some key specific details. And like all good lies, they are repeated so often that they may eventually seem like the truth. Let's look at some of them.

Segregated Witness solves the scalability problem

Segregated Witness only fixes the malleability issue for transactions that only use Segregated Witness UTXOs. All other transactions remain malleable. The fix is ​​enough to enable technologies like the Lightning Network. This comes at the cost of reduced fungibility, since only some coins can be locked in the Lightning Network.

Segregated Witness solves quadratic hashing

Segregated Witness solves quadratic hashing for some specific transactions, just as it solves malleability for some specific transactions. However, quadratic hashing is an attack vector, and solving it is as useless as birth control in some specific cases. Segregated Witness does absolutely nothing to solve anything related to quadratic hashing.

Segregated Witness allows block sizes up to 4MB

Technically, this is true. However, while block sizes can go up to 4MB, people don't need larger blocks. To illustrate this, imagine your friend has a kid and tells you that he needs a bigger car than his 2-seater. How do you think he would react if you suggested that he buy a car that was twice as big but still only had 2 seats? This is exactly what some people try to exaggerate.

Taking into account current usage, we can estimate that Segregated Witness will increase transaction volume by 1.7x to 2x transaction processing capacity. Segregated Witness allows block sizes up to 4MB, but only increases capacity by 2x, which is not something to be proud of. It adds a new attack vector because people may start mining overly large blocks, which ultimately reduces long-term scaling capabilities because the effect is twice as bad.

People don’t want big blocks just for the sake of getting bigger blocks, but because of the value that big blocks bring to users, just as people don’t want big cars just for the sake of getting bigger cars, but because big cars can carry more people or objects.

Lightning Network can reach world-class scale

The Lightning Network, or LN for short, is a technology that allows users to lock funds into a channel and then transact offline, the ultimate scalability solution for blockchains. It’s pretty good until you get to the cryptic text that’s always omitted for some reason.

The first is a change in the security model. With LN, the person you are transacting with can steal your funds at any time, and you have a given time to publish proof of the theft. This is a disadvantage in itself, except that people can work with the benefits of LN. However, with the chain running at full capacity, it may be very expensive or even impossible for you to publish proof in the required time frame. In the world of LN, a full block equals stealing. (Translator's note: If the block is full, you can't broadcast the proof of theft, or it will be very expensive to broadcast it, so it is equivalent to being stolen)

The second problem is that it requires an awful lot of capital to implement. If you want to pay $100 in fees using LN, you need to lock $100 or more into a channel. If you want to be able to receive $100, you need to lock that amount or more on the other side of the channel. If we imagine an extremely centralized network with a single hub in the middle that handles payments for 10 million users (this is the current order of magnitude of Bitcoin users) and allows those users to send $100 payments, all users would have to lock at least $1 billion in channels, and the hub itself would have to lock $1 billion in all user channels, and then that’s just to complete the $100 payment, what can $100 do. This is just equivalent to the Lightning Network at the scale of the Bitcoin network, which is a billion dollars in payment circulation. If we grow the Lightning Network to a billion users and an average of $1,000 in fees per household, we need the hub to handle a trillion dollars.

But we assume that we won't have one giant hub, but rather several interconnected hubs. Let's imagine that we can find routes for Alice to send her payment to Bob that are routed in an average of 3 hops, and then we need to either send Alice's fee to Bob or lock more funds in each channel. The inevitable conclusion is that LN strongly incentivizes centralization, because longer routes require operators to lock up more funds.

Schnorr signatures will provide additional capacity

Schnorr signatures are very cool, and indeed cheaper and more concise to verify than the currently used ECDSA, so you can fit more signatures in a block. However, the signature goes into the witness part of the block, and in current usage, if we keep the same block size limit, the expected witness data for Schnorr signatures is about 1MB, while the Segregated Witness block could have up to 3MB of witness data. Therefore, Schnorr signatures cannot provide any additional capacity because it is optimizing the wrong thing. (Translator's note: Because SW technology allows witness data not to be counted in the 1M transaction capacity range, even if the 3M witness data is reduced to 1M, it will not affect the transaction capacity.)

in conclusion

When someone keeps lying and exaggerating about their skills, rely on censorship to make sure they have ulterior motives. Don't believe them or you will be scammed.