Near Protocol vs Internet Computer on Ai app creator

General
1

This is just to compare and to light a fire under Dfinity’s arse.

Dominic has talked about a new feature coming to ICP that enables users to create web 3 applications by just speaking to an Ai. Just today he posted this on his X page:

…running web applications and internet services are created and updated simply by talking.

Users will create these for any purpose – to create something like a secure personal notes manager, or a personal website, or something for others, like social media, games, web3, or enterprise infrastructure, which involve online communities large and small – simply by issuing instructions in natural language.

That sounds game changing but I want to know how is this unique and if other blockchain projects are already working on the same. Here’s from Near Protocol:

NEAR.AI will build towards Open Source and User-owned AGI (artificial general intelligence).
Here is the plan:

  • Teaching machines to code: AI Developer

The NEAR.AI team, led by Alex Skidanov and me, will first build an “AI Developer.” We are attacking the core problem of how to build end-to-end applications from user intents. This includes Web3 applications, which require a high degree of precision and security given that they deal with user funds.
Source: NEAR AI

But interestingly, Near has also said this about their Ai project:

When we first looked at smart contracts, we felt they were an interesting subset of coding, but they were overshadowed by too many other challenges with blockchains at that point in time.

SO yes, I agree all of this stuff will be revolutionary, but who will get the credit? Part of that will involve showing how ICPs Ai app creator differs from Near’s. Also, it doesn’t hurt to be the first mover in this space. So far, both blockchains have not released this product despite a lot of promises.

2 Likes
2

It’s interesting. Even the latest LLM model can’t build my dapp, so I wonder what is going to be released, when, and if it is worth the hype. I just hope we are not hurting ourselves with these types of announcements.

3 Likes
3

And the Elon Musk of blockchain hits back…

2 Likes
4

None of what NEAR is doing will happen on-chain. They’ll take a hybrid approach which doesn’t address the fundamental issues Dom writes about. It costs 1 NEAR (~$4) to store 100KB of data on NEAR. Ignoring that, and looking at this from a product/adoption perspective - I still think they could garner success and attention from a hybrid model, especially if they’re seen as first movers. Dfinity needs to aggressively productize/market the technology while developing it to stay competitive and gain traction/attention in the space.

6 Likes
5

Agreed on all your points!

I’ve been getting DMs from people telling me that my bringing up Near is creating FUD. I respectfully disagree as I believe we should keep an eye on what other blockchains are doing and adjust accordingly (like getting more aggressive pushing products out!). At least we know that Dominic is aware given this response of his:

(Nonetheless, we predict that certain web3 projects that tend to imitate ICP, in desperation will use something like Vercel to create dumbed-down versions of this paradigm, then mis-sell their services to the token-buying public as being “onchain,” but beyond successfully selling more tokens for them, their schemes will ultimately fail to be competitive with mainstream users/creators worldwide.)

I never question Dfinity’s capability, but at the same time I don’t want them nor the community to become complacent thinking that other blockchains can’t steal our thunder just because we have the most advanced blockchain. Didn’t Near copy Chain Key tech from Dfinity but come out with it before we did?!

1 Like
6

Fwiw Dom already said that inference is off-chain. I hope they manage to fine-tune the model enough to create a few cool dapps, but it feels unlikely.

I wonder who is liable if the dapp doesn’t do what the user wants, contains a bug and loses funds, Dfinity? It sounds tricky.

Anyway, happy to see Dfinity going in this direction.

7

Upgrading the web 3 app is a big difference between the two crypto projects. Near is built on Google Cloud infrastructure:

NEAR, an open-source collective behind the NEAR blockchain protocol, ecosystem, and foundation, and which is building its infrastructure services on Google Cloud.

Source: NEAR building useable Web3 infrastructure on Google Cloud | Google Cloud Blog

And according to Dominic’s article, Ai created apps running on web 2 infrastructure would face these challenges:

In the new paradigm, the users/creators will want to update their running web applications and internet services in real-time, simply by telling the AI about the improvements, extensions and fixes they want.

Traditional IT has not been designed to make this possible. Anyone involved will know that upgrades are a big deal, and with production systems, typically occur on long time intervals. That’s because changes often have to be made across multiple components in a synchronized way (e.g. updating database tables, changing web server configurations…) and it’s a fiddle.

Moreover, when you change the design of web applications and services running on traditional IT, the upgrade process often involves refactoring/migrating data, which process is slow, computationally expensive and error prone – again, preventing traditional IT realizing the crucial real-time nature of the paradigm, which will involve running web applications and internet services being updated almost at the speed of chat.

8

The difference is pretty minimal. The IC is not going to win just because you can deploy an app more easily than on any other cloud. It’s not going to happen. Just ask any honest engineer at Dfinity.

2 Likes
9

1. From @dominicwilliams comment on X:

Here's how it will work:
  
1) You'll describe to AI the custom app or service you want 
2) It will return a URL to your web browser, and there it will be, ready to use!
3) You and others will use the app, causing it to accumulate content and data.
4) You'll describe improvements, extensions and fixes.
5) Then you'll just refresh the web page to see them.
6) Loop to 3, and continue iterating to realize value.

2. IMO atm you can’t effectively achieve that with current centralized clouds:

  • Note: the flow was first presented few weeks back so OG’s already knew :shushing_face:

3. Regarding demos or release of new features, I guess it will happen only after this:

  • Note: November 12, 2024 == next Tuesday
10

I recently built a Cloudflare app, because the IC disn’t support it, and the AI literally wrote all the code, gave me all the commands, and iterated over it. The only difference was that I had to copy-paste the code and run the commands.

11

Sounds cool - depends what all types of apps can be created this way - still the ICP approach seem much ahead to me :melting_face:

To stay on the topic, not really sure if it’s something what Near can do :upside_down_face:

And mainly, let’s wait for release of (demo of) the ICP feature to be able to compare more (indeed that will run fully on a chain - which is one of the most emphasized features of ICP since the beginning :crossed_fingers:)

12

It’s a capability of the latest AI models. So it should work with most languages/frameworks/clouds. Notice that it wasn’t even fine-tunned.

13

IMO this is more about the infrastructure, but as the ICP version is not out yet, it’s too soon to compare now.

Yes, AI is used by devs for writing code and others for some time already, but that’s by far not all what is planned on ICP - I think one of the main points is that even people who are not developers and can speak (and define a bit) will be able to create full and working custom apps - if that’s true and will work, I believe that it will drag to ICP huge new audience:

image
image950×1560 193 KB

1 Like
14

This is why a productization strategy will be so important for adoption and growth. AI is going to be integrated at every level of the web2 stack by big players and startups alike - we simply can’t compete in their territory or on their terms (at least for many years to come), so we need to think hard about where precisely this technology can provide the most value right now and standout to developers/consumers as the best option (by far). We need to identify and dominate the niches/narratives that web2 giants can’t touch, and slowly iterate/improve/expand from there.

15

From my understanding, but we’ll see when the feature is out:

16

The IC doesn’t have much of an advantage in this regard. It’s sad that I have to point this out instead of someone from Dfinity. Yet another Utopia-like announcement.

17

Sorry but I haven’t seen similar functionality so far (Near Protocol vs Internet Computer on Ai app creator - #15 by plsak), regarding UTOPIA I still wonder why the community reacted the way it did (because it’s IMO great and very needed option).

But as stated already, let’s wait and compare more when the feature is actually out, at least some demo is hopefully not far away:

18

It seems that Near released something. I don’t have a Near wallet to try. near.ai

IMG_2580
IMG_25801179×2314 196 KB

Source: x.com

1 Like
19

Yes, they released a product that is built and running on Vercel, which is just a layer on top of AWS.

1 Like
20

ICP AI is going to run off-chain as well, and this forum is on web2. We should be less on-chain maxis if we want to build actual products. It’s too expensive to run everything on-chain.

1 Like