在首尔路上偶遇了Jeff，看Jeff这意思是hyperliquid to da moon了

ALM knows. The world’s richest man - with one of the most capable minds and widest networks - now sharing you have roughly 18 months for the K-curve to either lift your financial goals - or leave them behind. Retar Dio 🤡🤍

Huge congratulations to everyone who made the list! 🏆🔥 The call numbers are looking insane, and it’s amazing to see the ALM community delivering quality crypto calls and getting rewarded for it. 📈💰 Keep dropping those calls and raising the standard! 🚀👏 Big congrats #ALM

Q4 isn't just another quarter for #ALM. three years of building, learning, taking hits, figuring shit out and still showing up. now we're finally entering the part where all those little pieces start coming together. i don't know what Q4 brings, but i'm damn sure glad i'm here for it. @Crypto_Retardio we been waiting for this one. ALM coded. 🫡

Most communities ask you to bring value. @Crypto_Retardio makes a point of giving back to the community too. From engagement giveaways and content opportunities to ALM Rewards, leaderboards and trading incentives, there are plenty of ways for members to get involved. You don’t have to be the biggest trader in the room. Show up. Contribute. Learn. Trade. That’s the #ALM community. If you’re on Hyperliquid, tap in. Use LOLOKLLM when getting in #ALOTOFMONEY Coded 🤍

Sometimes the real flex is to protect your capital, learn, and let the process compound. That’s why I’m bullish on what @Crypto_Retardio is building. Trade smarter. Stay grounded. Keep building. 🤝 Want in ? Use LOLOKLLM #ALM

Yesterday's VIP trades might just be a preview of what's coming next. And if the bigger plan plays out, we're only getting started. @Crypto_Retardio goal is to supercharge the vault over the next 3 months, build more capital, help members learn to trade, and get everyone better positioned for what's ahead. By January, the vision is 24/7 live trading with the main mods and experienced traders, alongside more opportunities for the community. Add the planned token, airdrops, giveaways and the rest of the ecosystem, and you start to see the bigger picture. Of course, timing and market conditions still matter. There's no point rushing what we're building. 18 months ahead. A lot to build. A lot to look forward to. 👀 We're still early in this journey. 🤍 Want in ? Use LOLOKLLM as your code #ALOTOFMONEY

It was underground this whole time now it's legal...

Memes are not the only way to retirement - ALM coded. Hyperliquid

Born in Higgsfield. All over your feed. Introducing Higgsfield AI Influencer. Create your own AI influencer and bring them into any trend. Try now with up to 5 FREE generations on Higgsfield and in the ChatGPT extension. Powered by Genjutsu. Credits: @JeanPhilMadame

i like hyperliquid

There's a fact that should unsettle anyone excited about self improving robots One major regulatory body already answered the question How to certify a machine that keeps retraining itself Their answer was you don't. You just ban it from doing that at all ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯ According to a 2021 liability regimes analysis referencing UNECE's own working party on automated vehicles Regulators reached a preliminary consensus that self and continuous learning should not be allowed in vehicles Because it's flatly incompatible with how existing safety regimes work I could not independently confirm the exact internal document behind that consensus It sits in a working group paper and not a public summary So take it as reported by that analysis rather than something I verified firsthand But the reasoning behind it holds up on its own and it's worth sitting with ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯ Certification assumes a fixed thing to test You run it through crash scenarios, safety checks, edge cases and once it passes You certify that specific unchanging behavior A system that keeps learning after that point breaks the premise entirely Two robots of the identical model, deployed in different environments can end up behaving differently within months, sometimes weeks At that point, what exactly did the certificate certify? ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯ Legal scholars call this a foreseeability problem You can only be held liable for what you could reasonably have predicted A system that rewrites its own behavior after leaving the factory makes prediction close to impossible And regulators dealing with vehicles decided the cleanest fix was removing the variable No post deployment learning allowed ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯ Robots don't have that luxury. Continuous adaptation is the entire selling point And the regulatory response to that tension is, charitably, scattered The EU treats it as a risk category: AI systems classified high risk get required mitigation measures for post deployment learning under the AI Act Plus a separate Machinery Regulation reaching full application in January 2027 specifically for safety-critical control systems The US has no federal answer at all, just a growing pile of conflicting state laws ► Colorado's AI Act landing this past June ► California's Transparency in Frontier AI Act in January ► New York's RAISE Act still pending With more states adding their own rules every quarter ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯ ► China took the opposite route entirely Tightening its Cybersecurity Law with mandatory content labeling and state aligned risk assessments ► Japan went the other direction still An explicitly innovation first AI Promotion Act that deliberately keeps the touch light Five governments and five completely different theories of how much oversight a learning machine deserves. The EU's AI Act does at least attempt something specific for this exact problem Article 72 requires providers of high risk systems to run post market monitoring Continuously collecting real world performance data specifically so that risks from systems that continue to learn can be caught early It's a real mechanism and not just language But notice what it actually does: it watches for problems after they start happening It doesn't recertify the system's new behavior before that behavior ships Monitoring isn't the same thing as approval ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯ Medical AI is a few steps ahead here and it's worth borrowing the framing At a 2026 radiology conference, Hardian Health's Hugh Harvey put it plainly: Regulatory approval is just an authorization to sell and not proof you're done And post market surveillance has to run as a continuous lifecycle process rather than a one time pilot study you walk away from Robotics regulation is not there yet It's still mostly stuck at the approve once stage medicine is actively trying to move past On the research side, there's at least one serious attempt at a solution A 2022 paper by Bakirtzis and colleagues proposes what they call dynamic certification Iteratively testing and revising which use contexts a system is actually permitted to operate in Rather than certifying one fixed behavior forever. It's promising It's also still a research proposal and not something any regulator has adopted This actually puts @axisrobotics in an interesting spot and I want to be direct about this rather than gloss past it Their whole architecture is built around exactly the category of behavior regulators are struggling with The system is explicitly designed so policies evolve continuously through feedback rather than isolated retraining cycles It's also, near word for word the definition of the thing at least one regulatory body has already said it doesn't know how to certify Nothing in Axis documentation addresses regulatory readiness directly None of this means the approach is wrong It means the industry is racing ahead of a question nobody outside a handful of working groups has actually answered yet What does it even mean to certify something that refuses to hold still Thats all for now and see you on the next one Here's what Brian, Coinbase CEO said about Axis Robotics: x.com/buildonbase/status/210… This is my unique link: s.kaito.ai/LQnXR5C Check here for more info about Axis robotics: linktr.ee/axisrobotics