Quick aside: Always irks me when someone says “Blockchain” singular, like it’s a proper noun entity cargo cult god. “A blockchain” is a particular type of data structure akin to a decentralized linked list, or in the context of one particular cryptocurrency you’d say “the blockchain”. I know that the crypto VCs use the word in that singular proper noun way, but they’re doing that intentionally to cultivate a misplaced reverence and we shouldn’t ape them.
Anyway bitcoin core v30, which just released, expands the maximum data size of a non-monetary transaction parameter called the “OP_RETURN” from 80B to 100kB, meaning that normal bitcoin transactions really can now store plain old jpgs with basically zero extra encoding. It was a wildly controversial change within the dev community, some people probably being paid under the table to push it through etc, but yeah it can do that now. I assume that some blackhat has already put CSAM onto the blockchain in protest.
However at this time you only have to pay < 0.2 sats/Byte to almost guarantee your transaction goes into the next block, so absent miners showing discretion and refusing to mint big OP_RETURNs, that 100kB transaction would actually only cost about 20,000 satoshi, or ~22 USD.
It was a wildly controversial change within the dev community
Sounds like it would be, isn’t the reason to keep storage expensive that everything included in a transaction needs to be stored forever by every single network participant running a full node?
Yeah that’s the dominant perspective, and the way I see it. The core devs seem to have the perspective that if they don’t allow storage of arbitrary data, they’ll only encourage the abuse of other transaction parameters to work around it, or encourage anticompetitive side channels where users pay miners directly to include their zany data. And that by opening the floodgates, they allow market competition to decide how much that data is actually worth.
The blockspace has not exploded to capacity with OP_RETURN data in the past week, so the whole controversy could just end up being what politicians these days call a “nothingburger”, without even the demand to reveal who was right now that the NFT craze has mostly ended. And if that’s the case, the core devs technically made the right choice. We’ll see.
What I was talking about is the storage requirement for holding a blockchain containing every NFT ever minted on that chain. That’s a lot of storage if it’s the actual images.
Oh yea you’d be trying to store essentially a whole image hosting site, it’d be absurd and only data centers would do it. In the case of bitcoin the block size limit is roughly 4MB and they rarely if ever hit that. Whole chain going back to 2009 is less than 800GB.
I’ve heard Bitcoin on-chain CSAM allegations for years but could never verify it. It would be interesting if a pruned copy was safer than an archival copy, forcing users onto Tor if they’re unwilling to trust anyone.
Quick aside: Always irks me when someone says “Blockchain” singular, like it’s a proper noun entity cargo cult god. “A blockchain” is a particular type of data structure akin to a decentralized linked list, or in the context of one particular cryptocurrency you’d say “the blockchain”. I know that the crypto VCs use the word in that singular proper noun way, but they’re doing that intentionally to cultivate a misplaced reverence and we shouldn’t ape them.
Anyway bitcoin core v30, which just released, expands the maximum data size of a non-monetary transaction parameter called the “OP_RETURN” from 80B to 100kB, meaning that normal bitcoin transactions really can now store plain old jpgs with basically zero extra encoding. It was a wildly controversial change within the dev community, some people probably being paid under the table to push it through etc, but yeah it can do that now. I assume that some blackhat has already put CSAM onto the blockchain in protest.
However at this time you only have to pay < 0.2 sats/Byte to almost guarantee your transaction goes into the next block, so absent miners showing discretion and refusing to mint big OP_RETURNs, that 100kB transaction would actually only cost about 20,000 satoshi, or ~22 USD.
Sounds like it would be, isn’t the reason to keep storage expensive that everything included in a transaction needs to be stored forever by every single network participant running a full node?
Yeah that’s the dominant perspective, and the way I see it. The core devs seem to have the perspective that if they don’t allow storage of arbitrary data, they’ll only encourage the abuse of other transaction parameters to work around it, or encourage anticompetitive side channels where users pay miners directly to include their zany data. And that by opening the floodgates, they allow market competition to decide how much that data is actually worth.
The blockspace has not exploded to capacity with OP_RETURN data in the past week, so the whole controversy could just end up being what politicians these days call a “nothingburger”, without even the demand to reveal who was right now that the NFT craze has mostly ended. And if that’s the case, the core devs technically made the right choice. We’ll see.
What I was talking about is the storage requirement for holding a blockchain containing every NFT ever minted on that chain. That’s a lot of storage if it’s the actual images.
Oh yea you’d be trying to store essentially a whole image hosting site, it’d be absurd and only data centers would do it. In the case of bitcoin the block size limit is roughly 4MB and they rarely if ever hit that. Whole chain going back to 2009 is less than 800GB.
I’ve heard Bitcoin on-chain CSAM allegations for years but could never verify it. It would be interesting if a pruned copy was safer than an archival copy, forcing users onto Tor if they’re unwilling to trust anyone.
Pruned copy would indeed be safe, it omits OP_RETURN data completely.