These are things we are already doing. But remember, oftentimes, if you want the most experienced developers, or people who have been working in cryptography research for years, they may take some time to become crypto natives. This is a tradeoff
10 Likes
The DFINITY Foundation is an independent contributor to the Internet Computer ecosystem. It is not the Internet Computer itself. I addressed my views in earlier posts in this thread (Transparency within the Dfinity Foundation). TLDR; our job is hard enough without getting involved in witch hunts, decentralizing the organization with committees, and so on. The fact is that the DFINITY Foundation team has achieved, and continues to achieve, far more than other organizations in the crypto space. Developing technology for a network like the Internet Computer is very hard. Period. Running multi-national R&D and crypto research operations at scale is very hard. Period. Therefore it’s no wonder things can take a long time and sometimes take longer to deliver than expected. Analysis of our GitHub repo clearly shows that we are consistently the top performer in the industry. I don’t think this line of attack is helpful
15 Likes
Dom great explanation, thanks. IC has a great future 
2 Likes
We are constantly talking to developers. We offer things like IC.Lab to help developers that need to collaborate with our developers. However, we have limited resources internally, and have to triage by focusing on the most important priorities, which change as the ecosystem develops. For example, in my earlier reply, I discussed how we are currently optimizing the cryptography relating to Update TX, as the result of bottlenecks occurring with the SNS-1 launch, so future SNS launches will run more smoothly. It’s certainly the case that the delivery of some promised frameworks has slipped, and priorities have changed according to developments on the ground, but this is not because we are neglecting anyone’s needs, only because we are resource constrained. We are constantly optimizing DFINITY and trying to focus and expand our capabilities, but we also need the community’s help as contributors.
22 Likes
I’ve seen you make this claim for a while now and imho it’s not completely true, per Dfinity’s own admission ICP uses Threshold Relay.
In Proof of work a single node can prove it executed calculations correctly and others can cheaply verify his claim WITHOUT running the calculations and comparing results, the problem is this is restricted to a specific type of calculation which doesn’t do anything useful other than provide security guarantees to the network. On ICP on the other hand all nodes must run the same calculation and reach consensus on the outcome, it is much closer to ETH 2.0 than to a Proof of useful work one, isn’t it? Main difference is how much each party contributes to reaching consensus isn’t weighted by amount of tokens owned. ZK based solutions are closer to what I’d define as Proof of useful work.
I don’t agree with the conclusion that only cause a node can be hosted on a rented server it is not decentralized, that is what happens when a network is permissionless and lets the free market decide how to run their nodes as long as they meet the network’s needs. You mentioned hertzner banning Sol nodes and while yes it cause temporary issues, the chain is back up and running again. That is cause it doesn’t rely on particular assumptions on where nodes must be hosted to run, what would happen instead if future regulations caused ICP nodes to shut down in all major jurisdictions? Would the network be able to oulive such a scenario?
KYC and NNS permission requirements aside, many are concerned about the low replication factor most subnets have: 13 is low when compared most chains and increasing it negatively impacts latency and throughput, furthermore there seems to be softcaps which makes running a 100 nodes subnet not possible at the moment. Once the network starts holding enough TVL the safety assumptions provided by deterministic decentralization with a 13 nodes subnet might not be enough in the eyes of many.
I think you understand the appeal of high repl and permissionless chains as they were both a part of the initial Dfinity whitepaper and Badlands proposal (of which I was a big fan) and it’d be great to see them implemented someday.
8 Likes
I 100% agree, @dominicwilliams. I’ve tried to defend DFINITY against those unproductive witch hunts several times in my short tenure here, and I’m a big believer in both the IC and the smart people behind it. I don’t think there is any better team in the industry. I also don’t believe in committees to decentralize power, since those inevitably just result in a forced reduction in diversity of perspectives and a tense power struggle to artificially come to a collective prioritization. Like cryptographers, I believe that this problem can mostly be solved by math. I actually want to make your life easier, not more difficult!
But how do you even know if you are optimizing priorities given your perpetual scarcity of resources to fulfill those priorities? You would first need to know how far off DFINITY’s priorities are from the community’s priorities. This implies that a collective prioritization process is in place to measure this gap on a regular basis, which is not the case. I recognize that DFINITY is an independent organization that has the right to have different priorities from the community in allocating its scarce resources. However, if you can’t see how far off you are from the community’s priorities, then you will only see more painful threads from passionate IC champions like @lei in the future.
3 Likes
@zane, I’m not an expert on this topic, but I think a healthy compromise on replication can be achieved via different canister (or subnet?) types. I think there are at least three types already: system, application and storage, which decreasing levels of replication for those three. There is also a “fiduciary” distinction, which obviously increases the replication requirements for decentralization purposes. Clearly, it is not possible for someone to host an enterprise dApp on a legion of mobile phones, so the Badlands idea could not be a generalized solution. However, for the most sensitive data or hackproof consensus requirements, that might be a great optional solution to have in the future.
My understanding of the delay here is the ledger approval process. We’ve repeatedly requested that cbor-based requests be added to the ICP Ledger App, but it seems to still only support the limited set of protobuff functions on the NNS. You can’t use it to just sign a transaction to regular canisters(and if this not the case it would be great to have some understanding of how to pull it of with the native APIs).
Firstly, Psychedelic and some others wanted DFINITY work on People Parties deprioritized since it was “application layer”. But we can now see, that while the functionality is sorely needed, no complete alternative was created by the community. We need anonymous Proof of Personhood (via People Parties) to a) prevent Sybil attacks on SNS decentralization sales, and b) boost the voting power of individual human beings inside the Network Nervous System DAO, so the network is not governed by capital alone. Both of these things are hot topics right now, and People Parties work will have to be restarted. I can tell you, unequivocally, that there are many community members who did not make so much noise during those discussions, who are now upset that the People Parties framework is not available for them to use.
Secondly, Psychedelic propounded the idea that systems like the SNS, and Internet Identity were “application layer”. I think they are more than that, and are essential parts of the decentralized compute platform that the Internet Computer provides – albeit, I think it’s great if people want to create alternatives, or extend them. There can be more than one idea. However, the SNS plays key roles in the 20 Year Roadmap I once described Announcing the Internet Computer “Mainnet” and a 20-Year Roadmap | by Dominic Williams | The Internet Computer Review | Medium, and which people invested into, and Internet Identity makes Web3 accessible to the masses and solves for things like privacy, which are previously described objectives of the project. There was loud criticism regarding e.g. the design of the SNS from this quarter. However, what people don’t see is the quiet majority from the community asking that we not be diverted because they need the SNS and are committed to the accompanying design and philosophy.
Thirdly, it should be noted that while Psychedelic was developing a large number of pieces of infrastructure, only a few have entered production. Had we listened to the arguments they made, the developer community would now be depending on their unreleased Daowabunga framework, because we would not have developed the SNS framework. That would mean no full end-to-end decentralization of Web3 and DeFi services, and no decentralization sales that the community can now use to fund their projects. Meanwhile, those who rely on the Plug Wallet (which is great, btw, I don’t want to criticize it) are now very worried that Psychedelic may not continue to maintain it. These are the reasons that the DFINITY Foundation, as a not-for-profit organization that is focused on contributing to the Internet Computer ecosystem in perpetuity, should continue contributing core infrastructure components in critical areas like identity and DAO frameworks – even though, again, we would still encourage others to extend the work we contribute or create alternatives.
In the end, Psychedelic changed their focus because they found it much easier to raise venture capital funding to develop a decentralized CDN (content distribution network) for accelerating access to Web3 content. That’s partly because what they are doing is a great idea, and I wish Harrison and his team all the luck in the world with that, and it’s also partly because traditional VCs are very heavily already invested in the Solana, Avalanche and NEAR ecosystems, for reasons of timing, and the kinds of attacks our ecosystem suffered at launch described by https://CryptoLeaks.info, and they rightly see the Internet Computer and ICP as potential disruptor of those investments.
Our answer to centralized investment cartels and other industry participants hoping to steer the evolution of the blockchain ecosystem towards themselves, and profit by creating alternatives to the Ethereum ecosystem through force of capital, control of the crypto press, and many other things, is the SNS and decentralization sales i.e the power of crypto. The SNS framework will democratize access to the Tech economy by allowing a much boarder range of Web3 and DeFi projects to raise funds from all over the world, and this approach is already drawing enormous talent into the ecosystem. I believe that in the end, decentralized systems (“algorithms and protocols”) will out-compete centralized funding cartels, and that as the Internet Computer ecosystem continues to take off – and there are already many fully on-chain Internet Computer projects around the world that I can see going to the moon – more and more traditional venture capital will find its way into our ecosystem anyway.
We need to stay focused on the fully decentralized future that truly defines our grand crypto and blockchain endeavor, and the amazing projects building fully on-chain decentralized Web3 and DeFi services, which will run under the control of their communities through DAOs, and which can only be built on the Internet Computer blockchain today. ICP alone can play the role of a World Computer, and ICP alone hosts smart contracts that can directly interoperate with other chains through chain key cryptography. This is the future that I and DFINITY are focused on delivering. Projects will come, and some will go, because this is inevitable in a large ecosystem. What is clear for anyone following events, is that the number of amazing new Internet Computer projects appearing daily greatly outweighs any losses. We are the only blockchain ecosystem experiencing rapid growth during the Crypto Winter. Network usage on other major chains has literally cratered, while ours is up 550% in the past 4 months. We must be doing something right.
59 Likes
I have my concerns with SNS and NNS but this was very moving and confidence inspiring
I will go to sleep happy tonight - WAGMI
4 Likes
Well said, @dominicwilliams. I can honestly say that I agree with every word of your last post, and not just this excerpt. In particular, I believe that the II (including Proof of Personhood) and the SNS framework are absolutely critical to the foundation of the IC. If you surveyed the community’s collective priorities right now, I think they would essentially agree too. But unfortunately we don’t have that data…yet.
@dominicwilliams let the world see more of this side of you. You can be very inspiring. This post was very digestible, I think you should continue with this vibe going forward. I think you should post this on other forums for more visibility.
17 Likes
Agree so much. That single post has changed my outlook back to bullish after becoming very demoralised. Would love to see more of Dom in other media as well, maybe even on Joe Rogans podcast kek
10 Likes
Your ideas on proof of personhood were what sold me on the Internet Computer. I was blown away by the idea of people parties as an anonymous way of claiming human verification. Thank you, and everyone at Dfinity, for this.
7 Likes
Running on big clouds makes the network infrastructure centralized as Dom said. Instead of Hertzner, imagine 3 or 4 of the top providers banning crypto nodes. That would like being the network down for days if not weeks before they can run a secure network again. I can easily see big cloud providers shut down these blockchains if they ever become a competitive threats to their business. The likelihood of Western governments banning data centers from running nodes without due process is far lower than corporations taking measures to protect their bottom line. We have already seen it happen. If all major countries ban crypto then it becomes irrelevant who can run a node or not. It would have very little use cases. Very few people would risk being criminal. No institutions would touch it. I find the idea that crypto can be untouchable from all governments, especially those from big democratic ones like the US and other western countries is unrealistic. A more realistic goal for decentralization security and censorship resistance from corporations and undemocratic countries imo. The United States by itself could probably kill crypto dream of ever being mass adopted by simply labeling everything a security without even banning a single node.
5 Likes
@lei we started a separate thread listing our SNS-1 Lessons Learned. I believe we cover all the findings you list in section 1:
1.1 User repeated deductions
Explained in Repeated ICP transfers.
1.2. Sybil Attack / Front-run Prevention
See Bots participating in the SNS decentralization sale and Use of airdrops. You point to various existing solutions to protect against bots. Indeed, we have also started discussing how we could build on existing community solutions and create a less manual and platform-specific airdrop functionality.
1.3. The Subnet node outage and congestion issues
We reported on the degraded performance of the NNS subnet during the SNS launch in a separate post mortem. This coincided with the SNS subnet got stuck bug explained in the lessons learned post. Of course SNS launches must sustain much more participants than what we have seen with SNS-1. However, I disagree with your TPS argument. If you only consider ledger transactions that were part of the SNS sale then your calculation is correct. However, transactions have a much more general meaning on the IC.
1.4. Lack of Rigorous Testing
You are right, a “public beta”, or maybe “alpha” is a good description of the SNS-1. We should have anticipated some of the limitations we hit before launching. We have identified various test scenarios that we have to run before opening up to further SNS launches.
1.5. Lock-up time
Yes, that was an oversight as explained in Neurons with randomized dissolve delays.
I hope that our explanations and transparent communication reassures the community that we learn from the SNS-1 launch and implement adequate improvements.
12 Likes
There would still be hundreds of ways to run a node VS the only one IC offers.
Not so sure about this. I’d say it’s more likely governments will want to shut down cryptos if it starts becoming a systemic issue, see Tornado cash, than all big tech turning down paying customers.
Crypto started out with the idea all of that would have been made irrelevant. A cryptocurrency like BTC might suffer in terms of price action if that were to happen, but an unstoppable and decentralized world computer? That is really valuable to a lot of people and would keep its value no matter how many roadblocks are put in place. It might not have the same efficiency and latency as your average web2 app, but the point of crypto is to be decentralized FIRST and efficient second.
The US does and has done lots of undemocratic stuff in the past decades, they are just better at marketing than others and slighly less bad than absolute dystopian countries.
The crypto dream was never supposed to be become rich buying internet money, but to create an alternative system that could survive any kind of attack.
5 Likes
Yeah but even the subnets with highest node count pale in comparison with other chains.
True, the idea is one should be able to choose based on its needs where to be on the decentralization spectrum, but at the moment Dfinity is building the network and optimizing the protocol with a specific set of assumptions: KYCed and permissioned nodes, powerful hardware, low repl count, fixed rewards, no staking/slashing, etc…
So for those who need less throughput and more decentralization there might never be an alternative on the IC, infact Badlands was kind of proposed as an alternative network, but I think it’d be nice if Dfinity kept in mind there is a need for more decentralized subnets too and didn’t corner themselves in a position that doesn’t make it possible to encompass those uses cases without forking the network.
9 Likes
In Dom and Dfinity we trust. I have the upmost faith, truely.
Having said that, I was highly disappointed by the SNS launch. This could of been a great way of showcasing the IC capabilities and I can’t help but feel, it was missed a opportunity.
Launches like this are rare occasions and I’m sure there could of been better ways of testing the technology, instead of doing it such a public and if im honest, such a disorganised manner.
4 Likes
I think Badlands is another example of a great idea that wasn’t given enough consideration before the public rejected it. The timing of the article’s publication didn’t help that.
8 Likes