Thread by @nishitaARK

1/ Another long thread on web3... this time, what does web3 hope to solve?

Below, a thread on 1. the importance of decentralization and 2. a simple mental model to quantify the value it provides.

And, in here somewhere, a few of my musings on... what we mean by ownership...

2/ Before jumping in, I think it's important to clarify that I strongly believe that crypto != web3.

Crypto is a broad term that describes the many ecosystems & innovations developing on public blockchains. IMO, this is broad on purpose.

3/ Specifically, web3 is a movement developing within, focused on a user-owned web. Based on my understanding, its four core promises for ownership are the following:

- Interoperability
- Transferability
- Scarcity
- Immutability

This framing is important, because...

4/ ...I've noticed that people often describe/explain web3 by the features it enables—rather than these underlying properties.

For example: yes, web3 can provide a seamless, single sign-on that allows users to access many web applications.

5/ And: yes, web3 can let users wield a digital sword across games or showcase a cool digital hat in multiple virtual worlds.

Importantly, however, web3 should enable these features through a truly interoperable, transferable, scarce, and immutable foundation.

6/ Laying out these promises suggests that web3 can only develop on top of public blockchains and crypto networks. To me, it also emphasizes that (a true) web3 probably necessitates very high levels of decentralization.

7/ Imagine, for a second, that some centralized entities decide to create their version of a new web. They choose to afford relatively high levels of openness and composability in the data layer (maybe even through a transparent blockchain).

8/ However, these entities also maintain some unique level of control (maybe a majority, maybe a plurality) over this new web's foundation and, inevitably, stake in the underlying economics.

9/ Is this web3?

If you choose to describe web3 by its features... then maybe—from this frame of reference, I think it is actually quite difficult to articulate why centralized entities can't create a highly convenient, seamless web... one with an illusion of user ownership.

10/ For example, at the extreme: with their reach and network effects, Meta and/or Apple (etc) could probably create a pretty compelling (centrally controlled) unified single sign-on or set of inter-game assets.

11/ Some would also argue that a somewhat centralized new web is already playing out on some crypto networks.

Which begs the question: do you truly own your (really convenient) single sign-on or digital sword if it is directly or indirectly controlled by a centralized entity?

12/ The problem is that such versions of the web appear to embody the promises of web3 to a degree. In these webs, users’ digital assets are reasonably interoperable, transferable, scarce, and immutable... as long as centralized entities (and regulators) want/allow them to be.

13/ While the ramifications may not be observable immediately, ownership comes with embedded risk... some probability that, at some point, your data/asset/identity may be manipulated by that centralized entity. Ownership in such a system can only exist with trust.

14/ A leap into the future (likely not a large one): if our lives are increasingly spent online and digital assets quickly migrate to web3, our very identities (comprised of our data & interactions/relationships with this data) will live on the networks that power this new web.

15/ So, let me rephrase a question I posed above: do you really own your digital identity if there is some non-negligible (or meaningful) probability that it could be altered, controlled, or censored by some vector for centralization?

16/ A mouthful I know... and to be fair, I know that centralization is a spectrum—in fact, I don’t believe you can ever really achieve full decentralization (there will always be some non-uniform resource distribution... more on that another time).

17/ To me, it seems like web3 should be built on the foundation that is least centralized—or, in other words, most decentralized.

18/ Interestingly, to achieve higher levels of convenience (through lower fees or higher throughput, or both), public blockchains often end up increasing levels of centralization. (This appears to be a clear trade-off.)

19/ And so, something we’ve been thinking about at ARK... if decentralization serves as an ‘insurance policy’ of sorts, how much should we be willing to pay for it (in 'inconvenience')?

https://twitter.com/wintonARK/status/1462637905159938052?s=20

20/ This is obviously a very complex problem... but I’ve been mulling it over with this very simple framework.

expected value of asset > network cost

In the above formula, network cost can include many things (gas fees, throughput constraints, general inconvenience, etc)...

21/ ... and my assumption (as expressed above) is that greater centralization leads to greater convenience (hence lower direct 'network costs')

22/ And, the expected value of an asset:

p = probability that your tx (or asset) gets censored in some way

a = value of tx (or asset) given it does get censored

b = value of tx (or asset) given it does not get censored

expected value = p * a + (1 - p) * b

23/ I find this framing useful bc it highlights that... as p increases (w more centralization), acceptable network costs decrease (for the tx to be, on net, valuable). It also highlights the opposite: as p decreases, we should be willing to pay more for our insurance policy.

24/ Importantly, the lower bound for the value of a censored tx isn’t necessarily zero. The loss of a digital asset may very well translate into a tangible, major economic loss. Decentralization becomes even more valuable as transaction value/significance grows...

25/ I know this is an oversimplification—we should actually worry about this calculus for all future time intervals (some sort of an NPV calculation)... also the probabilities likely change over time... as well as the values of the assets...

26/ But I think the main point remains: we should be willing to pay (potentially significantly) more for the insurance policy that more decentralized networks afford.

27/ With all of that said, I totally believe that some scaling is required on blockchain networks that aim to serve web3 (again, all about trade-offs)... but I think those scaling efforts (ex: through L2s) should probably still aim to share a highly decentralized foundation.

28/ Also, to be clear, I am very excited for the promises of web3! I’m just a bit skeptical of kinda-decentralized, kinda-centralized attempts at web3. This thread serves as my working hypothesis on why decentralization matters—more to come soon!

Feedback much appreciated!