"im not owned! im not owned!!", i continue to insist as i slowly shrink and transform into a corn cob
an NFT post in the year of our lorde 2k22.......... đ«
NFTs are weird. And also are probably post- the point of their flash-in-the-pan cultural relevance, what with the whole âbeing resold for 0.000583% of their worthâ thing. Which is to say I shouldâve written this when I first had the idea a few months ago. Alas!
Two narratives (or, narrative + counter-narrative) around NFTs seem to pop up most frequently when I read about them: that they are a technological breakthrough to enable ownership in the digital world, vs âlol no theyâre a link to a jpeg and I can just right-click and save itâ
My natural crypto-skepticism would lean me towards the latter! Not least in the aftermath of various stories that highlight how hastily-made crypto communities and DAOs can wildly misunderstand the nature of legal ownership and copyright.
And I donât think NFTs will be the great savior for funding creative works that boosters chalk them up to be (again, see the valuation cratering example linked above!)
But I do think that conceptually, itâs possible to reverse engineer from the design of NFTs a decently coherent philosophical theory of ownership (even if, spoiler, itâs not one I totally agree with.) Which I dunno, is maybe kinda interesting! Itâs a free substack, whatâre you gonna do about it!
Before that, letâs talk about digital files. And before that, talk about physical stuff.Â
A fairly reasonable model to abstractly define a lot of creative commerce is âIâm going to pay the artist money for access to the artwork.â If you then did the whole rational-self-interested-actor take to that model, youâd assume that if someone wanted the artwork, and could get access to it without paying the money, and wouldnât suffer any consequences for that, then they would do it! And go spend the money on like, hamburgers or something. Get the art and the hamburgers for the price of just the hamburgers! I am writing this part of this email during lunch.
In the real world where access to the art comes through actual physical things (whether thatâs a printed copy of book, or a theater performance, etc) itâs relatively easy to prevent unauthorized access to the work (they would have to get ahold of the actual book, or come in through the locked doors of the performance venue.) And copying, while possible, is pretty difficult for an unauthorized person to do. There is nothing technically stopping you from taking a novel and reprinting it word-for-word, but itâs incredibly tedious. Nothing stopping you from memorizing the play in your seat and reciting it to your friends at home, but it will be an imperfect copy of the real experience.Â
So unauthorized access to the art (either through accessing the original physical thing, or copying it) is difficult enough that it can be treated as a marginal threat, rather than an existential one.Â
That all changes in a digital world, where now you can make perfect copies of a file for free in seconds.Â
If you buy into the model above (only paying the artist for access to the artwork, therefore if you can get the artwork for free youâll do it!) then this is a catastrophe, because the physical safeguards of cost, time, or physical reality that would prevent infinite copying and distribution are all gone!
When you hear NFT boosters talking about how NFTs enable âownership of digital filesâ or âbring back scarcityâ, theyâre responding at heart to this crisis: how can digital content have value and make money in a world where it can be copied, transmitted, and therefore accessed infinitely and immediately for free?Â
Before talking about how NFTs try to solve this, letâs look at a couple other attempts over time. Caveat that this is based on vibes, not detailed research, so take this more as thought exercise than a detailed history of online attempts at copyright enforcement.Â
One approach that seemed to loom at least culturally large in the late 90s and early 2000s was: weâll just sue the shit out of everyone. Sure, weâll sue the places like Napster where people are illegally sharing protected works, but weâll also sue the individual people who are participating in this illegal sharing. And weâll couple this with scare tactics Ă la âyou wouldnât download a carâ ads and the FBI anti-piracy warning in front of every VHS.Â
In looking at this approach, it seems to operate from the premise of âWe will try to stop, or at the very least severely punish, the act of making, distributing, or accessing illegally copied work.â While there were technical approaches to aid in trying to identify and catch media pirates (like trackers on popular torrents, or enlisting ISPs as enforcers at the behest of music and movie studios), there wasnât really anything to technically prevent the actions, just a promise of punishment after-the-fact.Â
I think you can reverse-engineer this into seeing what, at their core, do the rights holders find valuable: in this case itâs possession of the actual file on your device. The things that are being punished are things that allow yourself or others to personally have a file that you didnât pay for. Whether or not you actually viewed the content thatâs stored on the file yourself is irrelevant, you would be sued just the same.
So in this case, you could make a case that ownership = physical possession.Â
Now there are downsides with this approach, namely: itâs completely impossible to effectively sue everyone. If the file can be copied and distributed instantly, infinitely, for free, then uhhhhhh not enough lawyers for that!Â
Since itâs impossible to sue everyone, the approach seemed to generally be âwell letâs be as ruinous as possible to the few we can catch and sue, and that will hopefully scare people even if the chance of them personally getting caught is low.â And per a half-remembered research paper talked about on The Weeds a few times, that totally doesnât work: severity of punishment has less effect on deterrence than certainty. A random but severe punishment will get a lot of people thinking âyeah but it wonât happen to meâ, while a milder but guaranteed punishment will get people to think twice and at least some of them to change their minds.Â
Letâs compare that with another approach: Digital Rights Management, or DRM. Thereâs a lot of flavors of DRM, but the one that Iâve seen the most is encryption and authorization based. If the âsue everyoneâ approach tried (and failed) to stop people copying, distributing, and possessing copied files, the DRM approach actually lets the copying happen without issue. You can generally copy and download a DRM-protected mp3! Photoshop is available as a free download!
However, the content on the file is often encrypted, in a way thatâs computationally expensive to decryptâŠunless you have the key that unlocks it. And that key is something that is unique, tied to you, and tied to you being someone who purchased the file (or authorization to the file) in the first place. Itâs why if you would buy a song off iTunes, then add it to a thumb drive, then a friend copied it to their computer and tried to load it, youâd get a big âAuthorization requiredâ pop up asking you to enter an iTunes Store account and password. Feel free to take the file, but if you want to actually listen to the song, youâve got to pay or enter the password of someone who paid.Â
So if the âsue everyoneâ approach says the valuable thing is possession of the file, the DRM approach says the valuable thing, the thing it means to âownâ the work, is being able to access the content.Â
In general a better approach, since the copy protection is as infinitely reproducible as the files itâs protecting. And itâs something thatâs still around, though perhaps less zealously than the heydays of the early 2000s (in part because the approach, for music, movies, and TV, has migrated away from ownership of discrete files to unlock towards streaming subscriptions to a large catalog.) But itâs also not without its drawbacks, namely that itâs a technical solution and therefore can be circumvented, even if itâs hard. And if you break DRM once, then that cracked version can be distributed infinitely and youâre back to square one. And DRM also has annoyances to perfectly legitimate user behaviors: it sucks to pay for a movie but have it say you need to upgrade to a better monitor because it canât work with their DRM system. It sucks that I can sell the physical copy of Normal People in my bookshelf, but not the kindle one I bought from Amazon. And so DRM-free is more of a selling point now for some services and platforms, but still, itâs maintained a reasonable hold as a method to fight the infinite reproducibility crisis.
Finally then, this brings us to NFTs. If âsue everyoneâ implicitly values possession of the file on your device, and DRM implicitly values access to the content, how do NFTs work? And what do they implicitly value?
NFT stands for non-fungible token:
Non-fungible: canât be funged. (Aka fungible = one thing is fully replaceable with another, so non-fungible is another way to say âuniqueâ)
Token: a traceable cryptographic thing on a blockchain
Other examples of cryptographic tokens on blockchains would be things like bitcoin, or ethereum or dogecoin. But in contrast to something like bitcoin, which is a fungible token (1 bitcoin = 1 bitcoin = 1 bitcoin and you can trade them all for each other and doesnât matter which specific bitcoin you have), each NFT is unique and not replaceable with another.Â
Blockchains, which to steal my favorite laymanâs description from someone who I canât remember currently but is probably Matt Levine, are essentially spreadsheets of transactions where you can only add new rows, you canât ever delete or edit them. And theyâre designed such that thereâs consensus between everyone participating as to what the ârealâ version of the spreadsheet is. In that world, since you have a record of transactions that everyone agrees on and no one can modify, you can reliably say what transactions happened in what order and therefore who owns what.Â
If we go back, another flavor of the infinite reproducibility crisis is that if the same file can be copied exactly, then how can you prove who is the ârealâ owner of it? There is zero technical difference between the copy of âInside Outâ by Eve 6 that I bought, vs the copy you surreptitiously made while I was in the bathroom and you copied over to your computer. So who âownsâ the thing now?
Blockchains enable, by use of the magic append-only spreadsheet, everyone to agree that certain tokens belong to certain people. And so if artwork could be encoded into a token on a blockchain, there could be this digital record of ownership. And in particular, youâd want every token to represent a specific piece of artwork; while it doesnât really matter which nickel I own as long as I own one and itâs worth 5Âą, it does matter if I own a Picasso or a Monet, and which piece specifically. Hence the non-fungibility!Â
In practice, because of technical quirks about the blockchains used for NFTs (and potentially blockchains in general, I dunno!) a token can only store a very small amount of data, much less than would be possible to encode even a jpeg of a painting, let alone a longer or more data-intensive work like a song or a movie. So most NFTs tend to encode a url that points to a piece of media.Â
This is where the critique of âYou donât even own your ape picture, you own a link to your ape pictureâ comes from, and itâs not wrong per se. But, unsurprising Iâd assume given how Iâve framed this whole piece, it really is more reflective of different definitions of what it means to âownâ something.
NFTs, as designed, treat the valuable thing, the thing that is protected by the technical design and paid for by people trading them, as having social proof of ownership. When I buy an NFT, I am primarily paying for everyone to nod and agree âYep, he sure owns that thing, huh.â Iâm paying for the link, and thatâs the point.Â
Vs. the critique1 is arguing for a different definition of ownership. Especially when the critique is âlook, I just right clicked and saved your jpeg, whole lot of good your fancy NFT did youâ thatâs implicitly arguing for more of the DRM model (where ownership = access to the content) or even the âsue everyoneâ model (where ownership = possession of the file on device.) And yeah, NFTs do absolutely nothing for you if those are your models of ownership!
But similarly, those mechanisms are less effective, Iâd argue, for the NFT model: DRM lets you access data, but doesnât give you social proof of ownership, same with lawsuits.2 In fact, one could imagine an NFT-like thing actually fitting into either of those other models! Imagine if there was some mechanism for a blockchain to be an identity list that maintains who is authorized to access certain content, and a DRM system was built on top of that such that if I sold my NFT, Iâd no longer be able to decrypt my DRM-protected mp3 file, but the person I sold it to could. Does that sound kinda dystopian, yes, but itâs conceptually interesting at least.3
I think then, there are two main, intellectually honest avenues of NFT critique that I actually mostly agree with!4 Cards on the table, I think NFTs are interesting but more of a misapplied novelty than an actual game-changer!
One avenue is to be clear-eyed about how the NFT model of ownership is different than other models, and then say that âyeah, and the NFT model is bad and other models are better.â
Should ownership actually mean âsocial proofâ? Is it not more inherently valuable for it to mean possession or ability to access content? I personally find the âaccess contentâ version most compelling, because, you know, hopefully thatâs why I care enough about the artwork to buy it!
Similarly, what externalities come from defining ownership as âsocial proofâ? It seems like the only real things it enables are resale and speculation (if you can prove you own a thing, you can prove youâre the one who can sell it) or in-group/out-group community signaling (Iâm part of the BAYC club, look at me!) Both of which I can see being fine when kept to reasonable levels, but would be very easy to take to gross extremes (especially the combination of the two, where you have community-fueled speculation and speculation-fueled communities!) Arguably, thatâs what we did see during the height of NFT mania last year!
Also, a lot of people seem to incorrectly think NFTs do work with those other models (ownership/possession), or try to huckster other people into buying into them under the presumption that thatâs the case. And no! Just because you own a Bored Ape NFT doesnât mean you can prevent someone else from looking at it or making it their twitter avatar! It does allow you to yell that them and probably get other people to look at the blockchain and agree and say âTut-tut, not very sporting, thatâ but if you actually want to lock it down, youâre better off trying DRM or a lawyer.Â
The other line of critique would be to say that NFTs arenât even that good at their model. And this I also think I agree with! Even among those who understand that an NFT encodes a link to content rather than content, almost all the discussion Iâve seen tends to operate as if owning the link to the content and owning the content the link is pointing to is the same. Or at least that by having an auditable record of ownership of the link, that transitorily should confer social understanding of ownership of the thing the link points to.
But you can make an NFT out of anything! If two people both mint NFTs of the same jpeg, you can rightly say that each person owns their respective NFT, plausibly say they own (by the âsocial proofâ model) the link encoded in the NFT, but who âownsâ the underlying jpeg? Both? Neither? The one who minted it first? What if a fan minted the first one, and the original artist minted the second?Â
And even if you assume the person minting the NFT is able to confer ownership of the original file, and thereâs only one NFT of said file, the NFT -> link -> content ownership chain is true for as long as the content the link is pointing to stays the sameâŠâŠ but thereâs nothing in the most popular NFT specs to ensure that that will stay true! Case in point, the Super-Fungible Token (link currently very NSFW, but who knows if it is by the time you click!) a project where an NFT was created that had a link pointing to an image, but where anyone could change what image was served at that link, at any time.
If one did want to make explicit that ownership of the link = ownership of the thing the link points to, NFT specs could include say both a link and a hash of the content served at the link, so at least if the content changed it could be easily proven. And Iâm sure thereâs other smart ways around it! But itâs not how they (mostly) work currently!
So overall I think NFTs are conceptually and technically interesting, but a pretty flawed approach, and certainly not the great savior of creative funding on the internet. I like that they explore and expose a model of what ownership even means, even if I think that model is less valuable and a lot of NFT discourse obscures what the real model is. It does make me curious about what other models may be out there and may come along in the future, either as entirely new models on the back of new technologies, or new technologies to refine and better execute any or all of the models explored above.
And again, we may be past the flash-in-the-pan moment on the whole NFT thing already, in which case I wrote and you read 3k words about four months too late to be worth it. This would be the ideal place in this essay to plug something like my own NFT line, huh. Alas.Â
In part, weâll get to more critiques in a sec đ
Maybe in a fun anti- way though, getting cred via being sued by Metallica or whatnot
Good time for a sidebar where I say one of my weirdly radical beliefs is I think hmmmmm maybe copyright shouldnât exist? But itâs also a belief more grounded in fury and vibes than actual study so I donât feel comfortable arguing it for real. But just wanted to throw this defensive sidebar in before it seems like I would endorse blockchain DRM as anything other than a thought exercise!
I guess the environmental bit of blockchains, especially proof-of-work blockchains, being inherently ruinous to the climate is another one that I think is mostly intellectually honest, but I have more complicated feelings about, and also doesnât really have any bearing on the ownership model topic at hand so all that to say hey, hereâs a footnote