I don't cry for frustrated attempts to abuse bitcoin to store unrelated data.  But Bitcoin's resistance to censorship for financial data generally also makes it fundamentally impossible to block the abuse, at most the cost can be modulated up or down a bit (e.g. by forcing them to disguise it better and use more of the block capacity to do so) but when the originator(s) of the data is willing to spend hundreds of millions of dollars in transaction fees to do it then their usage can displace that much in other transactions no matter what.

It's particularly important because the vast majority of people only care about the non-financial data to the extent that it disrupts their own usage.  People with data that care about how costly it is to store are not the people shoving data in Bitcoin, even with the data in a witness and at 1/svb it is *outrageously* expensive.  Instead you see people who are motivated by seigniorage (also the underlying motivation of most altcoins):  They want to mint something that is rare and they hope to sell it to a greater fool for more than it cost them to make.  Making data embedding more expensive doesn't hurt their economics, in fact it may help them (more rare!).  The disruption they cause to users is not a product of the number of bytes they send, it's a product of the money they're willing to spend on fees and that isn't likely to change much e.g. by forcing them to make their data pretend to be genuine financial data.

If it were the case that people looked at amazon ec2 prices or IPFS and said "oh look it's cheaper or similar to stuff our data in bitcoin!" then sure details around the cost would matter.   But if you imagine that the cost to store in amazon s3 forever is the cost of an investment that will return enough to pay the fee forever,  you end up concluding that amazon costs 0.0000066 satoshis per byte for forever storage.  Less than a hundred thousandth the cheapest fees in Bitcoin, and amazon's prices are nowhere near the cheapest prices for internet accessible storage.  So don't fall into the trap of thinking anyone storing data in Bitcoin cares for even an instant about small modulations in the cost per byte.  Some of these things seem to even intentionally bloat their data, like stuffing in whole json blobs instead of an efficient serialization.  They could just make their jpegs smaller or use a better compressor too if they cared. They don't.

But even if we thought that making it more expensive for them would be useful it has to be balanced against what it costs.   Having some developer going around playing wack-a-mole blocking stuff that isn't perfectly immitating ordinary transactions is a huge liability.  They'll inevitably block stuff that *is* financial data either by mistake or because come on-- what kind of person whats that job? It's always going to end up being done by someone who feels like they have the right to pass judgements about what other people are doing.   Now, you can fairly argue that because the blocking is not absolute that it's not a censorship concern, but then you've just built operation chokepoint for Bitcoin.  It turns out that censorship isn't a binary thing, there are degrees.  And while joe-blow-wants-to-donate-to-some-fringe-activitism doesn't have the skills or resources to bypass Mr. Transaction Dictator, the data flooder people absolutely do.   Even if some reduction in abuse could be achieved (far from clear to me), I doubt it's worth the harms of creating a quasi-censorship apparatus, I doubt it's worth the harms of muddling Bitcoins' freedom to transact value proposition with the idea that there is some person or group of persons trying to filter out 'bad' transactions.  I doubt it's worth causing authorities to ask why it's not filtering out the transactions they don't like and why when it does it's not really that effectual.

And finally, to the extent that something can be done and is worth the costs, it has to be done at mining not at random relay nodes in the network. Lets not fall into the trap of "Something must be done!" "This is something!" "Then it must be done!".


You do your own side a disservice though a poor argument.  Satoshi spoke out specifically and fairly vigorously against stuffing data in transactions, on a number of occasions  (e.g. https://bitcointalk.org/index.php?topic=1790.msg28696#msg28696 ) and quite ruthlessly limited transactions for relay or mining down to a very few kinds of precisely specified templates.  In private email, he had to be talked into the idea of allowing even just an arbitrary 32 bytes as a hash to discourage shoving data in wholesale for commitment.   If you're looking for endorsement for shoving external data into Bitcoin you won't find it in Satoshi.  Saying Bitcoin should support any kind of transaction doesn't imply he meant pretextual transactions that just serve to shove data in, particularly since he spoke out against doing so.

But it's also a poor argument because anything Satoshi said would have been a decade ago and said without the practical experience of actually seeing Bitcoin used, and that he was saying it as part of a discussion and not some contract or constitution.  We have forum posts from him, not stone tablets.  Many people here have an understanding of Bitcoin today that would have been impossible for Satoshi to have.  At most we can say his comments shed light on how people thought about this kind of activity early on, and you can see that it was at best controversial and never 'officially endorsed'.