In the past few days while browsing X, I found many people discussing a new project @Everlyn AI. Some say it is the Web3 version of Sora, while others say it is the culmination of AI + DePIN + IP protocols.
It makes me think that if this is true, then it could be the narrative core of the next bull market? After some digging, I can only say that the AI part is indeed substantial, but the Web3 part is basically empty.
First, let's mention the good parts; the team indeed has real talent. The founder previously did AI research at Meta and has published papers. The Everlyn-1 model they are working on follows an autoregressive approach, generating videos frame by frame like GPT, rather than using diffusion models. Theoretically, it can indeed be faster, longer, and more cost-effective.
But the problem is that their own materials state xDiT (Diffusion Transformer). These are completely different technical routes. One side talks about autoregressive, while the other talks about diffusion; this logic is a bit messy.
I also specifically checked their GitHub. The result is that most of the code is leftover files from previous research projects; it's not a runnable video system, nor is there any blockchain-related code.
Now let's talk about the Web3 banner they are waving. What decentralized video AI layer, on-chain copyright verification, creators earning tokens—sounds like all future stuff.
I also took a look, there is no white paper, no testnet, no on-chain browser, and no on-chain data. The only thing that can be confirmed is that they really issued one on BSC.
Aside from that, everything remains at the narrative level. Without on-chain logic and without code implementation, don't call it a decentralized protocol. Some content shared by others is even more exaggerated than what the official says; this is basically second-hand marketing rhetoric.
There are also many posts on Twitter saying that the advisor of Everlyn AI is Turing Award winner Yann LeCun, the chief scientist of Meta.
I specifically checked LeCun's official website, LinkedIn, and X, and I didn't see him mention Everlyn. The only source is a tweet from Everlyn itself saying, 'We are honored to invite LeCun as an advisor.' It feels like typical celebrity halo marketing, pulling a name to bolster the situation.
Of course, they did receive investment from Mysten Labs (Sui team), claiming to be $15 million. However, neither the Sui Foundation nor Mysten Labs' official channels have released any related statements; it seems more like a strategic partnership, not that kind of high-profile endorsement.
Then there is their token $LYN Launched on Binance Alpha, not the main spot market. Their marketing is mostly done on the Kaito platform, earning points by posting through Yap. Many of the content offered are promotional posts with bonus tasks.
Projects like Everlyn AI are actually quite representative. The AI is real, the team has substance, but once you stuff the Web3 narrative in, it can easily become distorted.
I remind everyone to stay calm when seeing 'famous team + top investment + exaggerated marketing'. First, look for on-chain data, white papers, and other hard evidence; after all, projects that can run, be checked, and verified are what truly count as real Web3.
