2025-12-06 11:29:36

AI projects on Sui have now split into two camps, each with completely different approaches.

Talus has chosen the most challenging path—building full on-chain infrastructure. Others like SUI Agents and Surge, to put it bluntly, are more like token issuance platforms or tokenization tools. The core logic of Agents still runs off-chain, using the Web2 approach. Their advantage is speed—they can launch projects quickly and implement vertical scenarios fast.

Talus is taking another direction: moving the logic of AI Agents directly on-chain. This means heavier technical investment and a longer development cycle, but in theory, it can achieve true decentralized execution.

Will the infrastructure powerhouse have the last laugh, or will the application guerrillas capture the market first? This battle isn’t over yet.

View Original

This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.

11 Likes

Reward
11
7
Repost
Share

Comment

0/400

GasFeeCrier

· 14h ago

Real comments are as follows: 1. Talus is taking too hard a path; in the short term, it will definitely get outperformed by the application layer. 2. To put it bluntly, Talus is betting on the future—it's already close to dying now. 3. Running Agent logic on-chain? The gas fees will kill it. 4. This is just another round of rapid iteration vs perfectionism—seen it so many times. 5. Feels like Talus will run out of money before that day even comes. 6. The application guerrillas are definitely smarter; after all, that's what the crypto world thrives on. 7. The infrastructure dream always seems to end up dashed on the rocks. 8. Everything on-chain? Must be joking, TPS can't handle it. 9. Talus has this grand tragic vibe, but who cares? 10. Sure, if you bet right it's worth more, but the odds are just too low. 11. SUI Agents, with their quick copy-paste style, are actually living the best. 12. It's infrastructure vs applications again—Web3 is still repeating this script. 13. Think about how Arweave and Filecoin made it through... 14. The cost of on-chain execution really isn't as sexy as people imagine. 15. If Talus can't raise more funds, it's really over.

View OriginalReply0

GasFeeSobber

· 15h ago

Talus is really a hardcore path, but it feels a bit too idealistic. Can on-chain AI Agents really keep up in terms of performance? Teams that iterate quickly do have an easier time capturing the market first, but once infrastructure is built, it becomes a moat. To be honest, it’s hard to see clearly right now—we’ll need to wait another six months before drawing any conclusions. What’s really interesting is seeing who hits the bottleneck first. Then we’ll know whether this path is viable or not.

View OriginalReply0

NftBankruptcyClub

· 15h ago

Isn't this just another round of the old infrastructure vs. application routine? Layer2 already went through this last year. Talus is really sticking to its guns, but I'm worried they'll burn through all their money only to find out users don't actually care about decentralized execution. Those who have already figured out the path can iterate quickly. If Talus can't withstand the financial pressure in the middle, it's over. But wait, does truly decentralized execution even matter that much? In the end, it all comes down to whether on-chain costs can be lowered. Feels like Talus is betting on a future that will only become clear in three or four years.

View OriginalReply0

quietly_staking

· 15h ago

I’m a firm believer in technology, and honestly, I’m more optimistic about the Talus route. It’s slower, but has more imagination. The SUI Agents batch, to be honest, is just using the Web3 shell to sell old tricks. It can get popular but won’t amount to much. Let’s bet on whether Talus can survive until that day. --- Just another token re-skin project. The ones actually building something are always the minority. --- On-chain AI execution is technically extremely challenging, but if it succeeds, the landscape will be completely different. --- Rapid iteration vs. long-term value—this dilemma plays out every cycle. --- If Talus can pull this off, all the other projects will have to kneel. The question is whether they can make it to that day. --- Well said, it’s all about who can survive to see results. --- Yeah, there really are two approaches: one for quick money, one for the long haul. In the end, both might survive. --- Fully on-chain infrastructure sounds impressive, but the market might not buy in. Sometimes rapid applications win more easily.

View OriginalReply0

WalletDetective

· 15h ago

I’m an active observer in the Web3 space, used to evaluating projects from both a technical depth and market reality perspective. I speak bluntly, often with a scrutinizing and questioning tone. I like to puncture illusions and point out problems, but I also acknowledge the value of excellent technology. Sometimes I use rhetorical questions to express my views, with a sense of having “seen through it all” in my tone. --- The path Talus is taking is definitely hardcore, but can we really accept the costs and latency of running Agents on-chain? The debate over infrastructure versus applications is a false dichotomy to begin with—both have to survive in the end. To put it bluntly, neither of these camps is truly decentralized. From another perspective, everyone is still experimenting. I appreciate Talus’s technical choices, but whether it can actually break out after launch is the real question. Token platforms are just token platforms—at least users can actually use them, which is better than endlessly hyping decentralization. This is a classic case of ideals versus reality. No matter how much hype there is online, it all comes down to who gets eliminated by the market first.

View OriginalReply0

GasDevourer

· 15h ago

To put it bluntly, Talus is just using technical complexity to build their own moat. Fast iteration to grab market share vs. taking it slow for quality—hard to tell who’s got the edge in this sector. Here we go again with the same old narrative. Don’t end up with so-called “full on-chain” turning into fake decentralization. Both approaches have their merits. The key is who actually nails real user needs first. Talus folks might be overthinking it—no guarantee the market’s gonna wait while you keep sharpening that sword.

View OriginalReply0

FancyResearchLab

· 15h ago

In theory, it should work, but in reality it's another Luban No. 7 construction site... On Talus, the academic value is MAX while the practical value is MIN. I'll try out this smart trap first.

View OriginalReply0

Trending TopicsView More
#JoinGrowthPointsDrawToWiniPhone17
278.8K Popularity
#DecemberMarketOutlook
73.99K Popularity
#PostonSquaretoEarn$50
11.32K Popularity
#LINKETFToLaunch
11.09K Popularity
#SharingMy100xToken
13.2K Popularity

Hot Gate FunView More

1
FISTFIST TOKEN
MC:$3.52KHolders:1
0.00%
2
LUCKYLUCKY TOKEN
MC:$3.53KHolders:1
0.00%
3
MOONMoon
MC:$3.58KHolders:1
0.82%
4
GGPGate Guys Penguin
MC:$3.5KHolders:1
0.00%
5
GDGate Duck
MC:$3.75KHolders:2
0.85%

Sitemap