elona

La solidaridad y la empatía son opcionales. El respeto a los tratados europeos y a los datos, no. 👇🏼 Número de entradas irregulares entre 2021 y 2026 según Frontex: Por Italia: 478.600 Por los Balcanes occidentales: 340.600 Por Grecia: 259.800 Por España: 234.760 Por la frontera oriental: 48.000 No es tiempo de dividir. Es tiempo de seguir construyendo una UE fuerte y unida.

SN76 // UNDER NEW COMMAND Phylax has taken over. New mission: secure the AI software supply chain. Immediate objective: recruit exceptional miners. Pin the code. Survive independent validators. Earn your rank through detection quality.

@Phylax_Subnet @bittensor Good luck guys!

AI agents are installing skills, MCP servers, packages, and repos faster than humans can review them. Phylax executes them in secure sandboxes and produces cryptographically verifiable safety attestations. Because trust_me_bro() is not a security model. app.phyi.dev

New project for #SN76 ! $TAO

@JensenHuang @nvidia So $TAO ? @JensenHuang

For my first post, I’m sharing a letter @NVIDIA signed on why open models matter. AI will transform every industry, power every company, and be built by every country. Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty. The world needs both frontier closed models and frontier open models. images.nvidia.com/pdf/Open-Weigh…

Feels like a good time to double the block size in anticipation of more applications and use cases for $NOCK. How’s 2mb? 2mb x 4 blocks per 10 min = 8mb per 10 min

@ambient_xyz I don’t understand how could support you. Do u have a token?

Two very interesting things happened on the same day this week: a) Palantir's CEO went on television and called frontier-model pricing a wealth tax on the companies that pay it. Hours later, Together AI raised 800 million dollars, led by the venture arm of Saudi Aramco, on the plain thesis that open-weight inference is the cheaper way out. Demand and capital hit the same target at once. The premise under every AI valuation (that a few labs can meter intelligence at a premium indefinitely) is now being shorted from both sides. That premium was never really about cost. Inference gets cheaper every quarter. The premium is about position. You integrate, your switching cost climbs, and the price rises to match your dependence. Now add the calendar. b) On August 2 the EU AI Act's enforcement powers come online, and provenance stops being good practice and becomes examinable. An auditor will ask which model ran, on what data, under which rules. A terms page is not an answer. So the moat moves. When intelligence is a commodity and the compliance clock is real, the scarce asset is proof. Proof comes in two shapes. One is a trusted party vouching for you, a stamp you still have to believe. The other is a receipt anyone can re-derive without trusting you at all. The first relocates the bottleneck to a new gatekeeper but the second removes it. Ambient is built on the second kind. One open model, priced by an auction that falls as the network grows, every answer carrying a record your auditor can check without asking anyone's permission ;)

Compute Networks will eventually allow @Nockchain to orchestrate compute jobs through the chain itself. To do this right, compute jobs have to be verifiable and measurable. Luckily we’ve got just the thing.

Been a quiet 2 weeks publicly, but a busy one behind the scenes: ✅ Incorporated the company & engaged legal counsel ✅ Flew to China to meet our potential technical lead ✅ Acquired a subnet ✅ Continued making strong progress on our pre-seed Putting the right operational foundations in place for long-term success before redesigning the incentive mechanism. More to share when the time is right. We will not be driven by hype, but by the product.

Great idea from the @webuildentangle's team. I'm really interested to hear what the owners think of "Root Reborn." - - bittensor:native

Our weekly Deep Dive is back! 🚀 Mainnet Launch Nodexo 3.0 is officially launching on Subnet 106, with testnet already live and mainnet going live on June 23rd. This marks the completion of Phase 1 of the roadmap outlined in the whitepaper, bringing the core marketplace and verification systems online. ⚙️ Proof of Hardware At the heart of Nodexo is Proof of Hardware, a system designed to replace trust with cryptographic verification. Before a GPU can earn, it is fingerprinted using its physical identifiers and bound to an on-chain identity. Validators then issue continuous timed challenges that only the claimed class of hardware can complete within the required window. This ensures providers can prove what hardware they operate rather than asking users to trust marketing claims or centralized operators. 💳 Flexible Payments The launch introduces three ways to access compute. Users can pay per rental through x402 using only a wallet signature, maintain prepaid balances in TAO or USDC, or consume compute through conviction allowance credits earned from locked subnet tokens. This flexibility enables both traditional users and autonomous AI agents to access compute through the same marketplace. 🔒 Conviction Allowance One of the most unique features of the launch is the conviction allowance system. Participants who permanently lock subnet alpha tokens earn a daily allocation of compute credits proportional to their locked position. The underlying stake remains intact while continuously generating usable compute capacity. This creates direct utility for token holders beyond speculation and ties network participation to real-world compute consumption. 🧠 Looking Ahead With the marketplace, Proof of Hardware system, registry contracts, gateway contracts, and conviction ledger now live, Phase 1 is complete. The next milestone is Phase 2: specialized compute markets operating within Subnet 106, each with their own token while inheriting the same hardware verification model. The long-term vision is a market of markets, all secured by verifiable hardware and ultimately collateralized in TAO. Stay tuned, more to come!

We know you've been waiting. Save the date!

🔥Nodexo 3.0 (#SN106) just did something new in the compute AI space... Most GPU subnets work like this: → Pay → Rent GPU → Done @nodex0_ flips it with bittensor:native ’s Conviction mechanism: Lock your Alpha/TAO on the subnet → never spend it. Your tokens stay yours (just locked on-chain). In return, you earn daily compute credits forever — real GPU power proportional to your locked amount. SUBNET 106 It’s not another rental marketplace. It’s the first compute subnet that turns long-term belief into perpetual compute access. Demand for GPUs = demand to lock TAO → less circulating supply + real skin in the game. Happy to see the Mainnet very soon (June 23) _ _ _

@nodex0_ (SN106) just dropped their whitepaper, and there's one idea in it worth understanding properly. But first, what is it? Nodexo is a peer-to-peer GPU rental marketplace built on Bittensor. Operators list their hardware, renters access it on demand, and the network continuously verifies that the machines are real through cryptographic hardware proofs. Not trust, not SSH access, not a dashboard promise. The machine proves itself on-chain, tied to an immutable device fingerprint, or it doesn't earn. No account required to rent. A wallet signature is enough. Which also means AI agents can access compute the same way a human does, with no onboarding step a non-human can't complete. Now here's the idea worth understanding. Nodexo is the first compute subnet to use Bittensor's new conviction mechanism in a genuinely creative way. Lock your TAO to SN106 permanently. Your stake never gets spent. You keep earning your normal staking yield. But on top of that, you receive a daily compute allowance, GPU access proportional to your share of total locked stake across the subnet. The pool is sized as a fraction of the fleet's real listed capacity each day. If no machines are running, the pool is zero. If the fleet is full, the pool is large. Compute you don't use doesn't roll over. It resets daily. This elegantly solves something DePIN projects have always struggled with: aligning long-term capital with long-term infrastructure. Most incentive models attract mercenary stake that leaves when yields compress. Conviction mechanics make leaving costly, your stake bleeds out over a 60 day half-life if you switch to decaying mode. The people getting free compute are the people who've committed to staying. The second idea in the whitepaper is just as interesting: validators need only a CPU to run on SN106, because verifying hardware proofs is cheap by construction. Any validator can operate their own storefront over the shared fleet. No single frontend controls access to the network. SN106 launches June 2026. Worth following @nodex0_

@nodex0_ Well done guys!

nodexo 3.0 testnet is live. proven hardware. real GPUs answering timed challenges, fingerprinted, bound to on-chain identity before they earn a thing. no phantom cards, no weaker silicon wearing a bigger name. the chain knows the difference. conviction allowance is running too. lock alpha to the subnet, earn daily compute credits, your stake never spent. belief, paid back in compute. three ways in: x402 per-rental for agents, prepaid credits, or the allowance you earn by conviction. validators run on cpu, anyone can open a storefront. a protocol with many doors, not one. mainnet lands tuesday, june 23. we lit the first compute subnet on bittensor. this is the one that rewrites what compute is. 106 $TAO