Open-ended question: Is web3… anything?

I think web3 is actually a bunch of things, some of which are more likely to succeed than others. E.g. the idea of using private keys for authentication/authorization is obviously better if you can get over the UX hump (go go SSH!). If we remove all the hyperbole from it, the best explanation I've heard so far is that web3 is a protocol/architecture for near-trustless* commitments. That doesn't sound like much, but commitments underlie virtually all of modern society, which I think is why the rhetoric around the space gets so crazy. Some examples: 1. How does someone know you are who you say you are? The government issues you a number of some kind, it makes a commitment that you are unique and this number identifies you. 2. How do you prove you own a house? The government issues you a deed declaring that some parcel of land and whatever is on it is yours. The government reifies this commitment as a copy of the deed stored in public records. 3. How do you prove you have the money to buy something? You either physically possess fiat currency, or you have a bank account, with a number that is unique to you and contains your balance. The commitment of that balance is upheld only by the bank itself. In each of these scenarios the commitments are upheld only by authority. I am the government and you trust me, so you'll believe that this number is this person or this house belongs to them. I am the bank and you will believe that you have $100, because I say so. With something like the ethereum blockchain, you can accomplish the same kinds of commitments, but without needing to trust some third party to continue to uphold their side of the deal. Removing the need to trust the other party allows for things that weren't previously reasonable. For example, let's say you want to fix up your local neighborhood's park. You have the time, but doing so will take a few thousand dollars in supplies. You could go to your neighbors and try to raise the money from them, but that requires them to trust you with whatever donations they give. 40 years ago in the US, that probably would've worked, but these days, trust is in pretty low supply and I suspect you would have a very difficult time convincing folks. Trust-less commitments, however, give us a different way forward. You can specify the amount needed and contractually state what is allowed to be done with it as well as contingencies for if the obligations aren't met. Once submitted, that contract will be carried out exactly as specified, it is observable to all, debuggable, and non-falsifiable*. So as a community we could come together, raise $2k, require that the person who is going to purchase the physical supplies also puts in 2k as collateral in the event that the community votes that the funds were misused and now no one needs to really trust each other. If enough money isn't raised, everyone gets their money back. If you tried to make off with the 2k, you just lost the 2k you put up yourself, so it's a wash. In this scenario, we talked about a local community because it's much more reasonable to imagine giving someone you know a donation. But if you put this on the internet, these sort of trust-less commitments could spring up for all sorts of things. From "investing" in the community of projects like Hest to large scale collective action like improving an entire city's parks. In principle a system that can make commitments without the need for authority could underly most forms of ownership and commerce. It also allows you to create systems of economic incentives that would otherwise require you to trust that individuals never change their mind or become bad - which has historically been a very bad bet. We'll have to see if it pans out, but at least in theory, the concepts behind web3 could provide a very different foundation for society, one where guarantees don't have to be backed by people with guns (aka governments), but instead a series of transactions on an immutable ledger backed by cryptographic math that can be automated, observed, and trusted. (*) nearly trustless because you have to believe that the network has not had a 51% attack and that the system and contracts are without bugs. Given the size of these networks and their history, by far the greatest risk are the contracts themselves, but just like there's standard contract legalese, I suspect standard, verified contracts would become the norm and you'd just be putting them together. I believe in practice, it is far more trustless than any other extent system.

Yeah, not all that great. Annoyed by certain financial inefficiencies, we were looking at doing stuff with tokens, in, maybe it was 2017. Once CryptoKitties appeared, we thought this was all going to go too weird for our buttoned-down tastes. Now over Thanksgiving I have my sister asking me, "So I was on TikTok, and what is an NFT?" "A scam," I said. For now, that's my honest answer to a first approximation. I'll change my mind when those gun toters recognize deeds and paperwork by consulting the blockchain. I'm most interested in boring administrative work: auditing, supply chains, compliance, logistics, paper pushing. Complaints about mining costs fall on deaf ears when we already build whole cities more or less dedicated to the same task. For me, if web3 ends up being anything, I certainly hope it ends up with more web1ish interoperability rather than the web2ish walled gardens.

@ibdknox that’s definitely one of the more approachable definitions / explanations i’ve seen of it: thanks!…looking from Europe that all seems very unwieldy and improbable but, as you note, community trust levels in the USA seem to have fallen to a very low point, so maybe i’m just being parochial (i think trust-in-state for, say, Nigeria or Venezuela, would also be very low). It does seem like an attempt to fix a politico-social problem with a technical fix tho.

💯 agreement with all that @ibdknox wrote. I see the biggest issue with today's web3 movement in overemphasizing the analogies with currencies, the use for trade (-> money), anonymity, and globalness. I suspect that a community-run blockchain (that neigbourhood park) in which most people know each others blockchain ID as well as their physical identity, could be a lot more useful than today's Bitcoin and Ethereum. In other words, a blockchain that complements existing trust relations, rather than building a completely separate parallel trust universe. For that to happen, we need people with different goals adopting and adapting Web3 technology. In the meantime, we will have to figure out how to prevent the first generation of Web3 from doing more damage.

I've seen web3 being related to different things: • identity based on public/private keys, no more signups, logins or passwords, reuse your identity across the web • user owns the data, like with identity cryptography and distributed systems allow you to hold data and decide which parts are shared with whom for which period of time • app distribution, fission.codes or holochain (haven't checked it yet) allows to distribute apps without having to manage the servers where they run • internet money protocol, having a native and decentralized way to charge for stuff on the web • user owns digital assets across places, you get a new digital art in one place and display it somewhere else, all places show a mark that validates that you own it

@Mariano Guerra that's basically what we've built, though we haven't figured out the best way to do the monetary part.

@ibdknox looking forward to it then 🙂

So, there's very rarely anything on FoC that properly engages me, but this thread does. It's exactly where I fit. I'll be watching this thread... 😊

@ibdknox just curious if you’ve been digging into Mark S. Miller’s work, that’s where I came across the line of thinking you described

I have not

you may find it worth a look, he strikes me as an Alan Kay type figure in the ‘decentralized trust’ world. in particular the E language influenced a lot of other projects. a possible starting point: http://www.erights.org/talks/pisa/paper/

I agree with @ibdknox that people talk about many different things with that name (a name I don't like very much, honestly). So far, the main things I like about blockchains are: 1. you have an identity (or identities) that you can take with you 2. all public activity for an "application" is stored in a shared database that survives the death of whatever organization(s) created it As a concrete example, an art market/community I've been part of since March recently shut down — quite suddenly and without notice — when the founder rage quit and killed the servers. Within hours the site was back up at a different URL with all user data intact. When people say "you could just use Postgres", this is the part that they're missing. OTOH, grander claims about a reorganization of society around the blockchain strike me as pure Panglossian absurdity. I have prepared two memes and a TWTR screenshot to summarize my critique of "guarantees don't have to be backed by people with guns" (a fantasy I think can only be enjoyed by people who have lived their lives insulated from the realities of force and violence).

We have has "Web3 as social technology" and "Web3 as infrastructure for interoperable apps and data" in this thread. As a complement, reflections on Web3 as an ecosystem supporting evolution: https://subconscious.substack.com/p/is-web3-a-petri-dish

Does Viznut's concept of Permacomputing fall into Web3? Or is that instead Web4, what happens after Web3 accelerates the collapse of civilization?

Permacomputing looks orthgonal to WebX issues, not being about "Web" at all. I suspect most Web3 ideas could be implemented via Permacomputing, with the obvious exception of "proof of work".

one where guarantees don't have to be backed by people with guns (aka governments), but instead a series of transactions on an immutable ledger backed by cryptographic math that can be automated, observed, and trusted.

This highlights some of the magic thinking behind crypto idealists though. There are benefits to this crypto on a cynical level. It's allowing tech companies to print their own money, it's a new path for digital artists (for now) to make money based on investment speculation, etc. It has made some subset of the population wealthy by getting in early on the right coin. But, you still need people with guns to back the "cryptographic math that can be automated, observed and trusted". I understand that like ML its a way to say "well the system said it was true" and launder data (which in this case is actual money!) There are also a lot of assumptions about trust in the US 40 years ago that are just wrong. 40 years ago was the 80s, community trust wasn't necessarily any better than it is now, and in certain communities it was significantly lower. At any given point you also have to ask who benefited from community based systems in the US. I also see community projects of various scales happen without crypto, so for those reasons I don't think your examples work very well. If crypto is part of the picture it doesn't fully solve any issues with community projects (real or perceived) and @Konrad Hinsen's approach where the is community trust still involved is necessary. If crypto makes it less work to organize, that's great! But a token can't guarantee trust in all steps in the process. This example bothered me for a lot of reasons, fundamentally the "tech will solve a perceived\imaginary social problem that I assume wasn't an issue in the past because people were better then". I guess I'm re-iterating @Jack Rusher and @wtaysom' s concerns in a more direct way. The libertarian house cat tweet was definitely on point. For libertarian systems to work, strong property rights are required. We have many abstractions (bureaucracy, laws, lawyers) and ad-hoc social systems to avoid violence in disputes, and crypto is possible another good one in certain cases (you can automate some of the bureaucracy in a decentralized way) but you need to have the underlying threat of violence to enforce rights in all situations.

fwiw, I was stating my understanding of the argument - I don't have a horse in this race. That being said, pointing at a group of people and calling them house cats doesn't strike me as a particularly useful way to have a conversation. Blanket writing off a group of people as incapable of understanding the system under proposal is a very low brow dismissal.

So as a community we could come together, raise $2k, require that the person who is going to purchase the physical supplies also puts in 2k as collateral in the event that the community votes that the funds were misused and now no one needs to really trust each other. If enough money isn't raised, everyone gets their money back. If you tried to make off with the 2k, you just lost the 2k you put up yourself, so it's a wash.

Could the community not just vote to get their money back, because they disagreed about the work that was done? You see this sort of dispute on Kickstarter quite a lot. I'm sceptical that there is significant scope to avoid classic arguments+evidence+authorities+judgements style dispute resolution. I would put Justice up there with Intelligence as one of the last things that will faithfully encoded in computers.

I saw it as a humorous way to describe a common belief, but maybe it is too spiteful. Ignore the tweet if you like, or the libertarian reference, because ignoring external systems that support specific ideals or lifestyles is wide spread, and most people I know (including myself) do it. Most of that rant is a desire for people to be more pragmatic and less idealistic about crypto

Yes, on KS they don't get the money back, but under this proposed smart contract mechanism the creator would automatically lose their deposit (if I understood it correctly). If it wasn't clear in my example, the backers would have no legitimate right to get their money back. This is the problem with systems that try to automate away the role of a judge/jury.

@ibdknox I was criticizing the belief system you were describing, not you — sorry if that wasn't clear! The point of my "low brow dismissal" is that property and contracts ultimately depend on violence coercion, and Libertarians philosophy is heavily in denial about this.

I recall that "Web 3.0" was originally supposed to refer to the Semantic Web, then Linked Data, now SOLID (social linked data), but cryptocurrencies and blockchains have largely grabbed the label and its associated meme of "after Web 2.0". However, what both have in common is that they appeal to a post-centralised or re-decentralised movement, taking back our power from Big Tech and so-on. That's the bit that's most meaningful to me, personally. Owning and controlling your own data, including your own money and your own identity without going through a controlling, regulated and centralised service. I believe we should be able to exchange messages and bigger documents, money and anything else we like, directly between us without surveillance, censorship or regulation. These inevitably serve governments and big corporations better than they do the actual mass of us on the planet. And that can be achieved without either Linked Data OR Blockchains, of course! That decentralisation movement started back with the original P2P techs at the turn of the century, like Napster, Gnutella, Freenet, etc., and lives on today in various P2P projects (including my own, of course!)

I also remember the days when Web 3.0 was a label for the Semantic Web. Which is still alive and well (Wikidata alone is a huge application), but not much discussed in public. It has become an implementation detail. So the label was free to take, and I guess the blockchain crowd could grab it only after it was itself hijacked by the gamblers that poured money into cryptocurrencies.

If the decentralization common thread ends up being realized, we should consider ourselves lucky.

Please elaborate on that, William, thx 😄

Just in response to Web 2.0 and mobile having become so centralized, decentralization could come as a nice change.

It's very hard to get traction and lift-off, because by definition all the Vast Sums of Money in the world that are needed to build, and above all promote, stuff are held rather centrally, by Big Tech et al!

Seems like Chris posted a slightly evolved version of his thoughts to Ivan's question on his blog https://www.chris-granger.com/2021/12/09/is-web3-anything/ which was posted to HN today https://news.ycombinator.com/item?id=29506914 which generated a lot of comments most of which aren't that enlightening. Having overcome my initial reaction that anything vaguely Libertarian is childish nonsense, I think I can now see potential for a "Level 2" solution (as blockchainers would say). That is, I believe the argue-it-out, backed-by-force legal system will always be the bedrock but there's a lot of potential for a system where cases essentially arrive at a court "on the blockchain" with certain facts (to an extent) mathematically guaranteed. To take the community parks example, there is no prospect to avoid disputes about whether the contract was actually fulfilled, but it should be possible to avoid disputes about, what the contact was, or who the parties were, or the chronology of certain events. These issues aren't the main point, but it's still valuable to take them out of the equation (to an extent, "they stole my private key" etc will always exist). I came to this revelation watching Tiger King season 2 😅 (where they try and suggest signatures on legal contracts were forged)

Yeah, I think a big part of what makes the discussion so wacky is that there's an inherent all or nothing assumption that all sides end up making. The HN comments assume that a foundation wouldn't have more things on top, it's somehow solely the commitment system, which is uselessly reductionist. Foundations are for building on and I have no doubt that much of what you'd build on top would be evolutions of things we have today. It's not like law or government or organizations would no longer exist. 😛 They'd just probably look different and have a different purpose if you changed that piece of the puzzle.

This article has some nice examples of blockchains being integrated into real-life processes (rather than creating a parallel universe).

Yeah that's complete bollocks

I just remembered to come recheck this thread. No insight to add, but I just wanted to say I'm glad for the art that has been entering my Twitter feed thanks to NFT hype. A sampling: • https://www.hicetnunc.xyz/objkt/501882 • https://www.fxhash.xyz/generative/380 • https://objkt.com/asset/KT1BbpLGaoFV1XzCcVMdg3eHA3LthFs5HXNZ/3 Note to self: go to Twitter more often just to scan @Jack Rusher's feed.

Here's a (random) article describing the death of hey nice tunic and how this is actually a test of Web3 principles kinda sorta. https://www.theartnewspaper.com/2021/11/16/founder-of-hic-et-nunc-pulls-the-plug-on-the-leading-digital-art-marketplacebut-its-half-a-million-nfts-live-on

An interesting take on what defines Web3: https://jaygraber.medium.com/web3-is-self-certifying-9dad77fd8d81 TL;DR: "Web3 is user-generated authority, enabled by self-certifying web protocols."

I just discovered https://jackrusher.com/journal/what-does-it-mean-to-buy-a-gif.html 💡

Your KNFT illustrates the well-known fact that most blockchain applications don't in fact need a blockchain. Public keys and/or hashes are often enough. And I agree on the behavioral data of course, but I see the article as arguing for a clean classification of Web 1/2/3, rather than for an exhaustive description, or even a list of all salient features.

So there are really two interesting parts of Moxie’s argument, in my view. The first is that “people don’t want to run servers” which is totally true (more on that in a sec). The second is that decentralization makes computing interfaces and development more — not less — complicated, and therefore more exclusive

Moxie's article reminds of a a data science workshop about ten years ago, in the early days of blockchains. IBM being a sponsor, it got a slot to present its latest technology, and one of them was a blockchain based system. Automated trust! No more authentication management! In the end I pointed out that "no trust" was obviously wrong: clients at least had to trust IBM that what they got actually was a blockchain (in proprietary software). The guy felt almost insulted and replied "IBM has proven over many decades that it can be trusted". He didn't really get my point.

True, though much of that is not too different from just regular software installed on a users computer. We don’t even really know what everyone being exposed to “mainstream traffic” would be like or how it would be structured. This is why efforts like libp2p are interesting to me

Here's an earlier talk from Moxie that goes a bit further on the decentralized/centralized question, in case anyone would like a more fleshed out version of that part of his argument:

Stuck without electronics for a few hours today, I find myself thinking about this whole thing from (likely idiotic) first principles. • It's really unfortunate that "proof of work" isn't proof of useful work. • Behavioral data is valuable. In web2, behavioral data is centralized. • Web3 is about hand wave decentralization. • What is the very first thumbs-up on Youtube of Gangnam Style worth? It seems to me the value is greater than 0. • As someone operating an open-source project, I would love to know when people use it. • People should have a stake in the governance of a product proportional to their use of it[1]. If two people disagree on a GitHub PR, I'd love to know who used it more in the past year. And no fair running a script a million times just before you open the PR. • People would be pissed if I had Mu phone home every time they ran Mu. Even though a (say) bare series of timestamps would not be worth much to anyone but me. • If Mu blows up tomorrow, why should you not get to sell proof of your early adoption to a collector? All this leads to a question: is there any way today to scalably prove that an event occurred at a specific time? Without needing to chain it with every single event that ever happened in the past? Ideally without anything leaving your computer. The certificates live in your computer, ideally take up little disk space and little computation, but giving you the ability to slice and dice them to share them with someone at a future date. [1] "One serious problem was that a lot of new people had shown up: the users. And there was no plan for how to incorporate their votes. Some of us (myself included) thought the users should not have a vote because it was not their language. It was alleged to me that the users’ investment was in using the language. That was a new concept for me and I did not immediately resonate to it, although I have come to see this as a legitimate point of view.” —Kent Pitman

"Proof of useful work" is something that Filecoin (https://filecoin.io/) claims to implement, the useful work being data storage. I haven't looked at their implementation, so I can't comment.

I'm not a fan of PoW, which seems like an ultimately self-defeating design (also ecosystem defeating at scale). Out of an excess of idealism, I'd rather see massive decentralized computational resources used for things like protein folding rather than shouting hashes into the void. Proof of Stake is better, which is one of the reasons I prefer Tezos over ETH on the rare occasions when using a blockchain makes some sort of sense. @Kartik Agaram Timestamps were the original blockchain use case, 14 years before BTC. See Haber & Stornetta, 1991: https://www.anf.es/pdf/Haber_Stornetta.pdf

I don't know much about them but it sounds like zero knowledge proofs may be useful: https://en.wikipedia.org/wiki/Zero-knowledge_proof