In short. Scientists from UCLA made an iron, which is a neural network. Not "iron + software". It's just that the very structure of the material thinks. Wires with a thickness of billionths of a meter themselves rebuild the connections between themselves. Just like a brain. Only without breakfasts, coffee and existential crises in the morning. You often "turn on" the connection - it holds on. If you don't touch it, it goes out. Do you know? I had the same with Spanish🤪 And the main advantage is that it consumes little energy. It's a gift for satellites, robots and all small equipment that has no sockets or a communication channel at hand. While the whole world shouts "more GPU!", someone quietly collects the iron that learns itself. I love this kind of news❤️‍🔥

Good morning, everyone ☀️ Touch some grass, enjoy the weekend & keep hunting for alpha 👀 At @PlayOnMint MINT Status comes with real perks, including weekly lossback. Stake MNTD after TGE, level up your Status, and make every session count🎮 Keep playing. Keep grinding🚀

Optimus Gen 3 may have appeared earlier than Tesla intended😳 A 3D asset found inside Tesla’s Android app allegedly shows a new version of the humanoid robot, with the current Optimus V2.5 reportedly shown alongside it. The hands and joints are what caught my attention. They look much more enclosed, with thicker forearms, slimmer wrists and almost no visible cables or mechanisms. The X Trajectory suggests some actuators could have been moved into the forearms, while the enclosed joints may provide better protection. But these are still assumptions based on a 3D model. The production reports mentioned in the video are also unconfirmed. So it is too early to connect this asset with a production ready robot. Still, if Tesla is moving toward a cleaner and more protected design, the next Optimus could look very different from what we have seen so far.

The future might show up right in your driveway. A robot could deliver your order, a driverless car could pull in, and an autonomous security system could monitor the perimeter. That raises a simple question: what would surprise an autonomous system in your yard first? A rough driveway? A blocked path? A narrow entrance? Some unexpected object? That is where @vangrid_io gets interesting. Regular smartphones can capture real world spatial conditions before the robot arrives. The more places autonomous systems understand in advance, the fewer surprises they face on site.

The most interesting part of @PlayOnMint starts after XP for me. Right now it is simple: play → earn XP → climb the leaderboard → receive a share of Season 1 $MNTD. But after TGE, that motivation disappears. That is when the real test begins. Will players stay? I see 3 questions. 1. What happens to retention after the airdrop? The $125K pool and low entry threshold are great for driving activity. But they do not prove people want to use the platform without XP. 2. Does MINT Status have real value? 7 levels, rakeback, lossback, daily rewards and raffles sound interesting. Now I want the numbers. How much actually comes back to players? How many users hold $MNTD for Status? How much does Status change the actual game economy? 3. How liquid will $MNTD be? There is no public sale, the first distribution comes through the airdrop, and around 8% of the tokenomics is allocated to liquidity. For a casino token, liquidity matters a lot. MintABear is also still a bet on future utility. 4,444 NFTs alone will not keep the loop alive. What they provide months from now will matter much more. That is what I want to watch. So far, the mechanics look stronger than the evidence behind them.

Good morning friends 🖤 I woke up today with one thought in my head: why am I even writing this?🫣 Not to look busy. Not to tick the box - "the post is out". I want there to be a trace in a month. So that I myself understand what I'm doing and can explain it to another person, looking into his eyes, and not looking away. If someone connected a forgotten wallet from my post, which was lying around with a dead load. Or threw out an empty text that just took up space - it's already good. It's not for nothing. Signs, numbers, coverage - let them be a side effect. Not a goal. Another scary thing is to become another voice in this noise, which said nothing. Maybe if each of us drags at least a drop of benefit into the world, the world will become a little kinder. That's the idea I'm sharing - together with the @NucleusCodes project 💫

PlayOnMint opened - I thought I'd come in for five minutes An hour later I'm still flipping through 😄 Slots, live casino, table games, Originals (Plinko, Mines, Crash - must try), fast play, jackpots. Top providers - Pragmatic, Evolution, Hacksaw. It takes longer to choose a game now than to play 🎰 @PlayOnMint 😎

AgiBot: from a single robot to mass deployment @AGIBOTofficial has produced its 20,000th robot - this is the total number of all robots of the company, not only humanoids. The anniversary copy is A3 Ultra, transferred to the Chinese amusement park Chimelong. Key detail: more than 300 AgiBot robots have already been deployed on the territory of one complex. Areas of application: excursions, hotels, retail, education, visitor service, sports. The focus of the industry shifts from the question "how capable of one robot" to the question "how effectively a team of hundreds of robots is coordinated in a single space". So you and I are witnessing the infrastructure of mass deployment.

Boston Dynamics just gave Atlas a real job to learn from🚀 A new training center is now operating inside Hyundai’s Georgia factory. Atlas is training on logistics and parts sequencing, with more manufacturing tasks planned over time. Hyundai says it plans to deploy 25,000 Atlas robots across its factories and build US production capacity for 30,000 robots per year. But there’s an important detail: These are future plans, not 25,000 robots already working on factory floors. What matters now is that Atlas has moved from controlled demos into a real industrial environment. The next test for humanoid robots isn’t how impressive they look. It’s how reliably they can work for 8 hours a day💁🏻‍♀️

Good morning friends 🖤 Billions are being spent to map the world. LiDARs. Mapping cars. Satellites. Expensive sensors. And yet there’s one problem none of them can solve very well: the world keeps changing. A road closes. A building appears. A sign changes. A parking lot becomes a construction site. So you send the expensive mapping car again. Or… Someone with a smartphone walks past it. That’s the interesting idea behind @vangrid_io Instead of building one massive system to constantly scan the world, it creates a distributed network of people capturing what’s happening on the ground. The funny part? The future of autonomous machines might depend less on bigger machines… and more on ordinary people carrying phones.

What if a surgical robot could build its own map while operating❓ That’s part of what researchers are exploring with a new flexible robotic endoscope. The system creates real-time mosaics of the endoscopic view, helping overcome one of the biggest problems of flexible endoscopy: limited visual context. It also combines magnetic actuation with teleoperated and semi-autonomous navigation. The technology was tested in sheep for fetoscopic laser coagulation. Still early-stage research, but the concept is pretty interesting. Could better spatial awareness become one of the biggest upgrades for surgical robots⁉️

I understand those who are afraid to shoot a street👀 "What if people get into the frame? And the car numbers?" It's a normal fear. That's why I didn't dare to shoot the city for a long time. I drew a simple rule for myself: if there may be faces or numbers in the frame, they should be closed before the video leaves the phone. Not later. I won't "figure it out on the installation". Not after loading. That's why I got @vangrid_io mechanics. It's not just a promise of privacy there. The unbled face does not leave the device. That is, protection does not take place somewhere on the server and not after the data has already been sent. It begins where this data appeared. For me, it's much more convincing than any words "we care about privacy"💁🏻‍♀️

What I find interesting about @PlayOnMint is the system itself: PLAY → XP → $MNTD → STATUS → REWARDS Players earn XP through slots, casino and sports. XP affects the size of the $MNTD airdrop. Staking the token then unlocks rakeback, lossback and other bonuses. MintABear is also built into the same Status and Rewards system. $MNTD has a fixed supply of 1B tokens on Base, with 30% allocated to community rewards. So the player’s activity is directly connected to further benefits inside the platform.

I read about how local residents block data centers for billions of dollars - and I see not just a conflict of "progress vs. Luddites". I see a question that is usually asked too late: who actually pays for the AI infrastructure?🗿 Formally, AI is chats and pictures. In reality, these are gigawatts of energy, millions of gallons of water and electricity bills, which are growing by 76% in a single state, because one data center has been built nearby. It's interesting that both sides say kind of reasonable things. Some - "we need infrastructure, otherwise there will be no growth". Others - "we are not against AI, we are against paying for it with our air and water". 843 protest groups across the country are not a marginal phenomenon. This is a signal that the speed of construction has overtaken the willingness to negotiate with people who live nearby. $32 trillion of investments by 2050 sounds impressive. The question is how much of this amount someone calculated in terms of human consequences, not only in terms of ROI 🫣 How can humanity pay for rapid progress? The question remains open...

Good morning friends 😎 Have you seen the news? @Tesla_Optimus learns on live data in Berlin💁🏻‍♀️ According to information from @teslaxander in Gigafactory Berlin the Optimus prototype is hidden behind a frosted glass marked "Work in Progress". Confirmed by the plant's management: since mid-August, employees have been participating in a global robot training program - they wear special backpacks and helmets with cameras that record movements. Conclusion: Tesla is expanding the testing and training of humanoid robots on the European platform, using data directly from real production processes, not just simulations 🏭

Something that is not on any map🫣 Maps know the roads. They don't know the steps. How many steps to the door, whether there is an elevator, where the intercom is - for a person it's a trifle, the eyes will figure it out themselves. For the robot, this is a wall, because this information is simply not in the data. If you are waiting for a courier or guests, just say: "second door, seven steps, intercom on the left". Three seconds, and wandering in the yard saves a person. @vangrid_io just teaches cars such trifles - not through a satellite, but through shooting people who are really standing at this door.