In May 2017, at a hotel in New York, representatives from 58 companies signed an agreement to change Bitcoin. Among them were miners who accounted for more than 80% of the network's computing power and exchanges responsible for most of the world's trading volume. The agreement completely failed within six months. Understanding why is to understand where the real power lies.
First, the technical distinction that organizes everything. A rule change can tighten or loosen. Tightening means prohibiting something that was previously allowed: the old nodes continue to accept the new blocks because everything valid under the new rule was also valid under the old one. This is called a soft fork, and it doesn't require anyone to update anything.
Loosening is the opposite: allowing something that was previously prohibited, such as blocks larger than the limit. The old nodes reject these blocks because, to them, they are invalid. If a group insists, the network splits in two, each with its own history up to the point of separation and its own life from there on. This is a hard fork, and it requires everyone to update — or accept the division.
The dispute that tested this lasted from 2015 to 2017 and became known as the block size war. The one-megabyte limit per block was set by Satoshi in 2010 as a defense against abuse, almost improvised. As usage grew, this limit began to tighten, and fees increased. One side wanted to increase it: cheap payments on the blockchain itself, available immediately, without relying on layers that didn't yet exist. The other side wanted to keep it small: large blocks make running a node expensive, and a network that only large companies can verify ceases to be the network described in the previous lesson.
It's worth noting that the first argument was neither foolish nor ill-intentioned. High fees do indeed drive away small users, who are precisely the ones who need it most, and the promised alternative — the layers — took years to truly function. The discussion was legitimate on both sides, which is why it was so bitter.
The technical solution that prevailed was SegWit, an ingenious soft fork: instead of increasing the limit, it moved the signatures out of the transaction body into a separate area, which was counted differently. The practical effect was more capacity without loosening any rules. Additionally, it fixed an old flaw that allowed altering a transaction's identifier without invalidating it — a flaw that prevented the Lightning Network from existing.
However, activation depended on signaling from the miners, and they weren't signaling. The deadlock was broken by a user movement: the BIP-148 proposal declared that, from August 1, 2017, nodes adopting it would reject blocks that didn't signal SegWit. It wasn't a request. It was a warning that those miners' blocks would be ignored by those receiving payments. Weeks before the deadline, the miners signaled.
The other side went its own way, as was its right. On August 1, 2017, Bitcoin Cash was born, a hard fork with eight-megabyte blocks; everyone who had bitcoin also had the new currency. And the second part of the hotel agreement, a hard fork scheduled for November that would double the block size, was canceled days before due to insufficient support among the nodes. In November 2018, Bitcoin Cash itself split again.
The lesson runs throughout the entire course. Miners order transactions, companies move volume, developers write code — and none of the three changes the rule alone. Those who validate are those who decide because the money you accept is the money your program considers valid.
The blockchain is public, and this publicity was the price of not needing to trust anyone. It's time to measure this price. In the next lesson, what exactly it reveals about you.