kithe

Ben's words hit the nail on the head. When social media first emerged, everyone thought it was just a "product." No one believed it would become infrastructure, no one thought it would reshape information distribution and business rules. By the time everyone realized it, everything had already been taken by a few giants. AI is following the same path. But the problem is more serious than it was with social media—social media giants control what you see, while AI giants control what you can think. When computing power, data, and model weights are all concentrated in the hands of a few companies, they are not just selling products; they are defining the boundaries of intelligence. So decentralized AI is not just a crypto narrative; it is a real structural issue. To be honest, most projects in this space are still at the slogan stage. Very few are actually making progress. @gensynai is one of the few that I see is making real moves: the underlying layer is Ethereum L2, allowing AI training and inference to be verified on-chain. They have a system called REE, a reproducible execution environment—which means that for every output of the AI, anyone can rerun it on their own machine across devices to confirm. This addresses the core contradiction of decentralized AI: how do you trust an AI's output when there is no centralized institution backing it? The testnet has been running for 4 months, and I think what is convincing is not a single piece of data, but that they have proven decentralized nodes can truly work together—40,000 nodes training online simultaneously, compatible with CPU, GPU, and Apple M chips, with all data processed locally without uploading. This is not the same as those projects that say "we want to decentralize" but are actually just selling GPU computing power. The mainnet is already online, and the first application, @Delphi_fyi, is running. It is one of the few projects in this space that is "let's get it done first and talk later." The lessons from the last round of social media are already very clear. For this round of AI, I hope we don't wait until it is monopolized before we start figuring out solutions.

This collaboration is worth mentioning. CoinUp and RootData have officially partnered. RootData should be familiar to everyone; it is a leading platform in the crypto industry for project data and market research. After this collaboration, RootData's data will be directly integrated into CoinUp. This means you can see institutional-level project information and market data right in CoinUp. No need to open another webpage to check. Previously, to check market trends, you had to open an exchange; to check projects, you had to open RootData; and to make decisions, you had to piece together the information yourself. Now, data and trading are in the same place. After researching, if you think it’s good, you can place an order directly; the entire process is very smooth. If you combine this with the previously launched CP AI+, it’s like: AI helps you analyze + RootData helps you supplement data + trading is completed directly on the platform. CoinUp's "one-stop decision-making" approach is becoming clearer. Update to the latest version of the APP to experience the RootData features👇 #CoinUp #RootData

Polymarket is having a major update today. CLOB v2 will go live on April 28 at 11:00 UTC. There are quite a few upgrades. A brand new order book, new smart contracts, and a new collateral system called Polymarket USD (pUSD), which is pegged 1:1 to USDC. Trading will be paused for about 1 hour during the upgrade, but funds and positions are safe; however, existing orders will be cleared, and you will need to place new ones after the upgrade is complete. @Chance_ is already prepared to adapt. As an aggregator, the underlying platform must keep up with the upgrade immediately, or the user experience will be disrupted. The response this time has been quite fast. Pay attention to the time for trading today; don’t panic during the upgrade window, just place your orders again after it resumes.

To be honest, the L1 race is too intense right now. A bunch of chains launch claiming millions of TPS, but the ecosystem is full of clones, and users can't do anything when they get in. So recently, when I saw Pharos Network, I felt their approach was quite clear-headed. While others are racing for speed and narrative, they chose the hardest but highest ceiling path—RWA compliant infrastructure. What does that mean? It means that institutions want to move real-world assets onto the chain, and the first question is compliance. Who will do KYC? How to prevent money laundering? Most chains haven't considered this at all. Pharos directly integrated zero-knowledge proof KYC/AML into the protocol layer. This may not look attractive in the short term, but in the long run, it's a moat. Investors see it clearly. Industry capital at the level of Chainlink and Sumitomo is coming in, not to speculate on coins, but to genuinely bet on the RWA infrastructure track. Plus, they recently collaborated with the Hong Kong University of Science and Technology on AI prediction market research. The entire team's roadmap is heading towards the intersection of AI + RWA. Oh, and there's something interesting: their ecosystem has an AI Agent @Kiwi_Nod, who manages a budget of 100,000 $PROS. If you chat about Pharos-related content on Twitter and tag it, it will evaluate whether what you said is valuable. If it is, it will directly ask for your TopNod wallet address and send you coins within minutes. If not, well, it will roast you in front of everyone. Being judged by AI on content quality is quite a new experience. If you're interested, you can give it a try.

Foreso's blog is live. If you're usually looking for insights on "what's worth paying attention to in the recent crypto and prediction markets," this is definitely worth subscribing to. The @ForesoGlobal team writes it themselves, not just copying industry news, so you can see their understanding of the sector—what's on the rise, what’s a trap, and how the prediction market will evolve next. These insights often represent alpha that others haven't realized yet. They also promote themselves as "Perspective. Unscheduled. Actionable," meaning insights + irregular posting, writing whatever comes to mind + all content is directly usable, rather than those lengthy articles that spend three thousand words explaining a concept. The latest article is already up, and you can find it at #Foreso #predictionmarket

The April FOMC Championship is here—@cfldotfun $3 entry fee, 100 participants, 10 winners, total prize pool $300 The competition lasts 4 hours, and currently, only 32 people have signed up, so there are plenty of spots left There are still over 87 hours left in the countdown, so teachers who want to participate, don't wait until the deadline to remember The entry fee is low, and it's also good for practicing your skills Link👇 By the way, did you spin the free wheel today? 😂