learner-long-life · January 14, 2022 21:49
diff --git a/gistfile1.txt b/gistfile1.txt
 In response to https://moxie.org/2022/01/07/web3-first-impressions.html

 A well-stated critique is worth its weight in gold. I'm a fan of the work on encrypted text messenging by
 anyone named Moxie Marlinspike, and acknowledge that it reflects a particular ethos, exhibited by Signal,
 a project that Moxie co-founded with Brian Acton. This ethos is strongly opinionated on behalf of users,
 makes the best choice for most people in a large variety of cryptographic
 protocols and their parameters, and offers it in a tightly integrated desktop/mobile app and centralized
 server infrastructure.

 Marlinspike raises a valid gap between the decentralized ethos of web3, and a few centralized providers of its data.
 I offer this disagreement on his main points:

 "Most people don't want to run their own server": (my paraphrase) "most" does not mean "all". Most people don't want to edit
 Wikipedia either, and yet we have a thriving knowledgebase which is the best encyclopedia by most measures, in the world.
 Speaking as someone who runs my own email server, and
 of the 10k's validator nodes currently running or staking across all blockchains, I am glad we have the choice to do so,
 and we have a part to play, however invisible, in the open ecosystem of ideas and technologies that is the internet.
 Moreover, I'm glad the technology is by default interoperable and server-agnostic. More on that later.

 "The vast majority of web3 (he assumes the EVM ecosystem here) calls are made through two centralized providers." No one can
 pull out his old man cane and wag it thusly faster than me. In 2016, I remember when there was no public Ethereum node except
 Augur offering theirs for free to the develop community. Someone thought, "this is ridiculously central, we're web3,
 let's get Consensys to plow some money and make it a proper service" so Infura started. Then around 2018 other people
 thought, "this is still ridiculously central, we're web3, let's crowdfund a startup to make general queries" and so Subgraph
 and Alchemy were started. We do need more nodes, full archival nodes, light nodes, and other kinds of nodes that haven't been
 invented yet, to achieve the full promise of data being universally accessible and mashupable independently of a particular
 client. Some would say this is the spirit of state synchronization protocols like Braid [1].

 More people are spinning up Ethereum nodes all the time, there are between 2,000 and 5,000 currently, not all of them
 public (able to serve API calls from third-party developers and clients), but in theory capable of serving the community
 of clients for which they were designed [3]. This is a security consideration, not just one limited by the web3 movement
 and its motives. There is a distribution to turn your Raspberry Pi into an Ethereum
 node by flashing a particular image. [2] Can we do a better job as Ethereum developers to direct our users clients
 to different API endpoints, or even better, take the majority vote of all of them? Of course. There is only thing
 I respect more than this eloquent critique, in which Moxie clearly spends time writing an NFT dapp to better
 understand the web3 community, is a critique which includes a proposed or implemented solution. And that more
 respected thing, is to do what Augur, the Ethereum Foundation, Consensys, and the web3 community did and is doing:
 continue to spin up nodes, and make them easily accessible to the public.

 However, a beautifully integrated web3 service is unlikely to come from Moxie given the context of his considerable
 previous work and the style of his public service.
 Signal itself famously discourages people from decentralizing their own servers, and delivers a mostly great security
 experience for the general public. Such a tight integration does save people time and labor, and is an approach that
 Steve Jobs himself favored with Apple products. If Jobs were to guest lecture in Moxie's class, he might say,
 "Signal users are busy people. They don't have time to understand PGP and keyservers. We let them get back to what
 they are good at doing."

 At the core of Moxie's disagreement with web3 is this tension between the current state of this young movement
 and its stated ideals, as well as possibly the unsustainable amount of popular attention and financial investment that
 cryptocurrencies (independently of web3) are receiving. My interpretation of this tension are the questions,
 "who is a cryptographer?", "who deserves to use cryptography?", and "what should we be using it for?"
 Most importantly, following Steve Jobs's wish for Apple users, what should we let people get back to doing, before web3?

 I'll close by countering with a complementary tension.

 If open protocols (let's use email) are run by a few, or only one, company, how should
 users feel about their vulnerability to this company and its choices? Is our only option to help Gmail avoid being
 a central point of failure, to grant Google even more data center real estate and trust them with even more of
 our information? Open source licenses don't cover blocklists or relaying between
 email servers, for example, the social and "in-flight" aspects of technological software, which often has a much
 longer life in terms of maintenance than the software development process itself.
 Who determines who is allowed to send mail? Does the IT staff of the thousands of universities
 around the world deserve our support and admiration for continuing to run a few independent internet services,
 the ones that haven't converted to the Google Business Suite yet? Or do they warrant criticism
 for not accepting standard wisdom that Gmail *is* effectively email now, and instead refocusing their efforts elsewhere?

 [1] https://github.com/invisible-college/braidjs
 [2] https://ethereum.org/en/developers/tutorials/run-node-raspberry-pi/
 [3] https://www.ethernodes.org
	In response to https://moxie.org/2022/01/07/web3-first-impressions.html

	A well-stated critique is worth its weight in gold. I'm a fan of the work on encrypted text messenging by
	anyone named Moxie Marlinspike, and acknowledge that it reflects a particular ethos, exhibited by Signal,
	a project that Moxie co-founded with Brian Acton. This ethos is strongly opinionated on behalf of users,
	makes the best choice for most people in a large variety of cryptographic
	protocols and their parameters, and offers it in a tightly integrated desktop/mobile app and centralized
	server infrastructure.

	Marlinspike raises a valid gap between the decentralized ethos of web3, and a few centralized providers of its data.
	I offer this disagreement on his main points:

	"Most people don't want to run their own server": (my paraphrase) "most" does not mean "all". Most people don't want to edit
	Wikipedia either, and yet we have a thriving knowledgebase which is the best encyclopedia by most measures, in the world.
	Speaking as someone who runs my own email server, and
	of the 10k's validator nodes currently running or staking across all blockchains, I am glad we have the choice to do so,
	and we have a part to play, however invisible, in the open ecosystem of ideas and technologies that is the internet.
	Moreover, I'm glad the technology is by default interoperable and server-agnostic. More on that later.

	"The vast majority of web3 (he assumes the EVM ecosystem here) calls are made through two centralized providers." No one can
	pull out his old man cane and wag it thusly faster than me. In 2016, I remember when there was no public Ethereum node except
	Augur offering theirs for free to the develop community. Someone thought, "this is ridiculously central, we're web3,
	let's get Consensys to plow some money and make it a proper service" so Infura started. Then around 2018 other people
	thought, "this is still ridiculously central, we're web3, let's crowdfund a startup to make general queries" and so Subgraph
	and Alchemy were started. We do need more nodes, full archival nodes, light nodes, and other kinds of nodes that haven't been
	invented yet, to achieve the full promise of data being universally accessible and mashupable independently of a particular
	client. Some would say this is the spirit of state synchronization protocols like Braid [1].

	More people are spinning up Ethereum nodes all the time, there are between 2,000 and 5,000 currently, not all of them
	public (able to serve API calls from third-party developers and clients), but in theory capable of serving the community
	of clients for which they were designed [3]. This is a security consideration, not just one limited by the web3 movement
	and its motives. There is a distribution to turn your Raspberry Pi into an Ethereum
	node by flashing a particular image. [2] Can we do a better job as Ethereum developers to direct our users clients
	to different API endpoints, or even better, take the majority vote of all of them? Of course. There is only thing
	I respect more than this eloquent critique, in which Moxie clearly spends time writing an NFT dapp to better
	understand the web3 community, is a critique which includes a proposed or implemented solution. And that more
	respected thing, is to do what Augur, the Ethereum Foundation, Consensys, and the web3 community did and is doing:
	continue to spin up nodes, and make them easily accessible to the public.

	However, a beautifully integrated web3 service is unlikely to come from Moxie given the context of his considerable
	previous work and the style of his public service.
	Signal itself famously discourages people from decentralizing their own servers, and delivers a mostly great security
	experience for the general public. Such a tight integration does save people time and labor, and is an approach that
	Steve Jobs himself favored with Apple products. If Jobs were to guest lecture in Moxie's class, he might say,
	"Signal users are busy people. They don't have time to understand PGP and keyservers. We let them get back to what
	they are good at doing."

	At the core of Moxie's disagreement with web3 is this tension between the current state of this young movement
	and its stated ideals, as well as possibly the unsustainable amount of popular attention and financial investment that
	cryptocurrencies (independently of web3) are receiving. My interpretation of this tension are the questions,
	"who is a cryptographer?", "who deserves to use cryptography?", and "what should we be using it for?"
	Most importantly, following Steve Jobs's wish for Apple users, what should we let people get back to doing, before web3?

	I'll close by countering with a complementary tension.

	If open protocols (let's use email) are run by a few, or only one, company, how should
	users feel about their vulnerability to this company and its choices? Is our only option to help Gmail avoid being
	a central point of failure, to grant Google even more data center real estate and trust them with even more of
	our information? Open source licenses don't cover blocklists or relaying between
	email servers, for example, the social and "in-flight" aspects of technological software, which often has a much
	longer life in terms of maintenance than the software development process itself.
	Who determines who is allowed to send mail? Does the IT staff of the thousands of universities
	around the world deserve our support and admiration for continuing to run a few independent internet services,
	the ones that haven't converted to the Google Business Suite yet? Or do they warrant criticism
	for not accepting standard wisdom that Gmail is effectively email now, and instead refocusing their efforts elsewhere?

	[1] https://github.com/invisible-college/braidjs
	[2] https://ethereum.org/en/developers/tutorials/run-node-raspberry-pi/
	[3] https://www.ethernodes.org