Let's solve these crucial protocol weaknesses

I would like to know what happened with @lastmjs after all the fud he made on twitter with the AO thing? What are his thoughts after @PaulLiu explained in detail how everything works there. I would really love to see people complaining but also coming back to say sorry when they were wrong and not let their EGO dominate them.

1 Like

What do you mean “they were wrong”? Jordan had some great points, and so did others. That is, the concerns are real but at the same time, over time, can be overcome. Thanks to him, we learned a lot about AO, the intricacies of ICP design, and the future roadmap. Give the man the respect he deserves. He has been one of the top contributors to ICP outside Dfinity engineers for years, and doing so in the bear market. He is clearly passionate about IC and wants it to succeed.

Btw, one of his last post on Twitter.

6 Likes

[quote=“free, post:50, topic:28329”]
terms of safety / tamper resistance / trust.
[/quote

What about ownership ? One key point and huge use case for enterprises and entrepreneurs like me looking to create new business models.

To be honest I don’t see how different this idea is for how the data is managed on systems like flux, now the AO thing… can you explain to me please, in detail how IC will provide ownership, and can keep claiming to be able to manage and host data on chain with this idea?

I’ll be pushing to preserve the on chain capabilities as much as possible even though it will take some time to develop and take some research, like some in the dev community are pushing for things to be done faster and simpler instead of developing revolutionary technology and features, I really care about innovation, because I see that’s where is the key of success to the IC not on “faster off chain computation” and “cheaper off chain storage” do you really think that’s the future? Are we trying to compete with aws with the capabilities they already have? Or are we trying to bring to the cloud industry a revolutionary way to build software, run verified computation and have compute and ownership over data, even though that mean in the short mid term it will carry some limitations?

of course there will be limitations in a completely new operating system, but what are the huge advantages of using this OS? I don’t see ANY value on a platform offering off chain storage and off chain computation the add value of the IC will be lost immediately and many people will start to go somewhere else that’s a fact. Again dfinity team members don’t allow a few developers to tell you what to do and don’t hurry to make them happy affecting a huge component of the IC.

The truth is some developers care about not hitting instructions limit because they are not using the architecture how they should be using it, and the message size limits because they are not using all
The subnets to upload the data. Correct me if I’m wrong.?

I don’t have the time to give you a detailed explanation, especially since you say you don’t have a technical background.

But on the one hand I believe you basically misunderstand what is meant by payload references / hashed payloads / side channels. It’s just a way of making the block arbitrarily large and thus increasing subnet bandwidth to whatever the physical network can support. So instead of being forced to gossip the full block to all replicas on a subnet in under 1 second (which is where the current 4 MB limit comes from) you would gossip (or transfer via IPFS; or compute next to the replica) an arbitrarily large payload and take your time doing that without slowing down the subnet; then, when a majority of replicas have that payload, you would just include its hash in the block and gossip that tiny block in the 1 second window you have. But once that block has been notarized and finalized, it (logically) includes all the large blobs you have gossiped around earlier and all or most of the subnet’s replicas now have it as part of their replicated state.

On the other hand, you already have stuff like HTTP outcalls (without which the IC would be a rather limited proposition), so a canister / subnet can already virtually depend on “off-chain data”: there’s nothing stopping me for storing my ledger in a SQL database running on my laptop and having a canister on the IC that simply makes HTTP outcalls to my laptop whenever you want to make a transaction. By which I mean to say that the boundary you believe to be so clear between on-chain and off-chain is very much not so. The only way you can trust that a dapp is fully “on-chain” is to have access to its source code, understand it and satisfy yourself that all the relevant bits and pieces are where you would expect them to be. As soon as you hit a black box (e.g. a canister to whose code / configuration you do not have access to), there’s no way to be sure of its safety / tamper resistance / trust.

Same with an application you develop for your own (e.g. as a DAO / company / whatever) use: without open sourcing the code you can still make sure that your data is on-chain, replicated, tamper resistant to whatever level you desire (e.g. by putting it on a fiducial subnet, with higher replication). And you can choose to offload some of it to IPFS or to your laptop, but no one is forcing you to. It’s just tools that you have available and it’s your choice whether to use them or not.

8 Likes

I have started a new thread that focuses on AO’s new whitepaper: Let's Review the AO Whitepaper's Characterization of ICP and AO's security model

The paper has a brief section on ICP, and I think it would be interesting and useful to ensure it is accurate, and to further analyze AO’s security model.

2 Likes

I think there are many that hang on the every word that Dfinity have put out about IC without having the technical knowledge to understand the reality of what is possible vs what is being proposed.

As a software engineer, this has been a very useful thread for bettering my own understanding, albeit I will need to re-read it a few times.

It seems that IC will never be able to deliver what most have been expecting.

I couldn’t disagree more with this statement.

That’s to be expected from certain quarters I guess. Will be more than happy to be proved wrong.