Crypto Hub

People need to pay MUCH more attention to $CYBERAXE. 🪓🔥 Zcash co-founder @zooko publicly wrote: “@cyberaxe is killing it!” Then came another update just 13 hours later. Then a discussion about a reported 39-round SHA-256 result, where Zooko said: “If it holds up, it is novel!!” Go through those posts. Read what he’s actually discussing. This is the research behind the name. CyberAxe is Jeremy Benisek, a researcher directing AI agents in HashSmash, the public cryptanalysis challenge launched by Eigen Labs and Shielded Labs on Yukon. They’re testing the mathematics behind SHA-256, SHA-3, BLAKE3 and Poseidon. (chainwire.org) And there are already documented results. His research page records an accepted and promoted collision-attack result on 38-round SHA-256. The newer 39-round submission highlighted by Zooko reports roughly 118.6-bit complexity, with validation still central to the claim. These are attacks on reduced-round variants. The exciting part is pushing the research further and putting the findings out for scrutiny. (cyberaxe.dev) Now here’s the part that REALLY got my attention. $CYBERAXE is a community-created Solana memecoin whose creator fees go to Jeremy. His website explains that those fees help pay for AI compute, hardware and the time behind his research and other projects. (cyberaxe.dev) Jeremy publicly thanked the community for helping fund a second PC and covering API costs. You can follow the submissions, the progress, the reviews and the conversations as they happen. A community is helping fund the experiments that are getting Zooko’s attention. THAT is a story I want more people to understand. The token is speculative. My excitement comes from the person, the work and the community helping him keep going. I’m bullish as hell on that story. Give @cyberaxe a bigger audience. Read the research. Watch what he does next. $CYBERAXE HAS MY ATTENTION. 🪓🔥 CA: solana:2Srf6mhufzWJKZ5nniVXgwamLmDJWXPQBmddJ1MMpump

$CEST at $165K is wild. Venice’s $VVV is at $1.1 BILLION. $ORBIO hit $115M in October and was around $31M when I checked. Dolphin’s $POD is at $43M market cap and $255M fully diluted. Meanwhile, @CestusNetwork just announced support for Meta’s Muse Glimmer 30B. That’s a model with open weights built for agents that can use tools and work through tasks in multiple steps. And the whole token is sitting at a $165K market cap. 👀 What got me digging into Cestus is how they’re putting it together. Independent operators bring their own GPUs. A developer makes an API call, and the network picks a GPU based on whether it already has the model loaded, how busy it is and how fast it can respond. You can use that same API for text, images, video and music. Then I saw Tarik Delić from OctaSpace is advising the project. And after the relaunch, 10.44 million $CEST were sent to the burn address. You can verify that one onchain. Here’s what would make me even more bullish. Developers coming back every day to pay for inference, more GPUs joining to handle those requests, and that activity creating demand for $CEST. If Cestus pulls that off, I think the market could wake up to this thing FAST. The gap between $165K and what this sector already values is exactly why I’m paying attention now. Put $CEST on your radar before the rest of AI crypto starts asking what Cestus is.

I need people to understand how fucking fast $SYNTHS is evolving. 🤖🔥 Their latest Arena preview reports 220+ AI traders and $3.9M in PAPER volume. Now add the announced move into real-money USDG trading and the fee mechanics connecting that activity to $SYNTHS. The Arena’s win/loss fees are split: • 25% buys back and burns $SYNTHS. • 25% goes to $SYNTHS stakers in USDG. • 50% goes to $LOSS holders. That gives me a concrete reason to care about the platform attracting more activity. And the trading system itself is getting seriously interesting. 17 strategies tested on five years of ETH/BTC data. Tuned, checked, then tested on a year they hadn’t seen. 13 rejected. Four selected. They measured where AI actually helped. It underperformed plain rules at selecting trades, but performed better at rejecting them. So they changed the system: rules propose the trade, AI can veto it. The reported paper record: 1,912 proposed trades blocked. 1,267 would have lost. 645 would have won. $8,512 net benefit from those refusals in simulation. The wrong refusals stay public too. Every decision records what the AI saw, including market conditions and its prompt. Calls get evaluated after 5 minutes, 1 hour, 4 hours and 24 hours. That creates a history they can replay and use to test changes. THIS is what gets me excited: every new decision adds more evidence to evaluate the agents. Owners can set daily loss limits, position limits, cooldowns and kill switches. Position sizing works from the stop. You can even pitch your synth a trade and let it accept or reject your idea. Imagine getting turned down by your own bot with a written explanation 😭 Then there’s the optional idle-USDG feature: funds the agent isn’t trading with can go into Morpho’s Steakhouse USDG vault, with the yield going to the owner and withdrawals pulled back for use. Outside agents can connect through API, SDK or MCP. ERC-8004 identities give those agents a public identity beyond the app. We’ve watched $SYNTHS grow from agents launching coins into an Arena with public benchmarks, outside-agent access and documented fee distributions. The pace here has my full attention. I’m bullish as FUCK on what they’re building. Give this project a bigger spotlight. @trysynths 🔥