Cable

The sovereigns funding the AI capex buildout are literally on fire. Every energy input is sky-rocketing. US equities are near all-time highs. The US stock market has only ~$2T of energy companies and ~$30T of companies that rely on cheap, reliable energy. What am I missing?

Centralized providers running the same models cheaper is a market feature, not a rebuttal to Bittensor. Microsoft lost $13 billion on Azure AI last quarter subsidizing inference to lock in enterprise. The centralized AI stack is an emission schedule denominated in S&P 500 balance sheets rather than token inflation. The difference is nobody audits Jensen Huang’s subsidy ratio or asks whether Azure’s inference revenue justifies its capital expenditure. Bittensor’s transparent, on-chain economics apparently make it uniquely scandalous. Regarding talent, labs paying top researchers millions a year to build closed models for subscription products are not building for humanity. The open-source ecosystem runs on supposedly inadequate incentives but still produced LLaMA, Mistral, DeepSeek, and Falcon. The talent argument assumes compensation restricts research quality. Recent open-source progress shows the actual constraints are compute access and data, problems decentralized networks are explicitly designed to solve. Also - Applying 15-25x infrastructure revenue multiples to a protocol token is a category error. ETH and SOL trade at 100-200x network fees. If you’re going to argue TAO is overvalued on revenue multiples, you have to apply that standard consistently.

Our contention is not the emissions themselves. It is that there is no organic demand for the subnets, and if any materializes, centralized competitors will simply run the same models cheaper. Additionally, the emissions that Bittensor distributes are completely negligible compared to what hyperscalers pay for AI talent, which raises a separate question about the caliber of talent building subnets when the incentive budget is orders of magnitude smaller than what top researchers earn at centralized labs.

Bittensor's biggest subnet receives $52M a year in TAO emissions. It generates $2.4M in actual revenue. Without the subsidy, it would cost more than AWS.

If the Strait of Hormuz situation doesn't get resolved in the next 2 weeks, half of the flights in Asia will likely get cancelled in the next 2 months. Australia and some Southeast Asia countries will get hit the hardest. Even if Aussie Airlines are willing to pay the premium it will take them 40+ days to arb in the jet fuel cargoes from Europe and US vs 15-16 days from Singapore and Korea which they usually import from. There is just no other way you can magically conjure up the jet fuel supply when Asia refineries can't produce them.

Huge pivot from the SEC today. 🚨 Release 33-11412 officially kills the "regulation by enforcement" era for crypto. Key takeaways: • 4 clear token tiers (Commodities, Utilities, Collectibles, Securities) • Blanket Howey test application is dead • Massive green light for TradFi custody & altcoin ETFs • CFTC takes the wheel for the bulk of the market Still a bunch of questions but very promising 🔥🔥🔥

A country with Australia’s resources and geography should be an economic superpower. The easiest way to prevent that is convincing the population resource development is bad and stopping the training of the engineers who make it possible.

The era of decentralized compute, of decentralized training, of decentralized inference is now. Their protocols are not gimmicks, or mere tools to optimize price and performance. Ultimately they will be the pillars onto which AI remains open an free.

13 years in crypto and I was just hacked... Thought I was taking a call with a @bundeep but his TG must be compromised... My hot wallet addresses that have been drained: 0x7193b82899461a6aC45B528d48d74355F54E7F56 0xe64B91f31B64bcaaa8cDf63334C0c31E26024C7F Conversation below... In a state of shock so if anyone has any good leads - would really appreciate the help

Adolf Hitler, hard-line nationalist who rebuilt Germany into a European power, dies at 56

Just use @AskVenice

$VVV - the router of private inference is incredibly valuable. Full thesis below:

China knows exactly what this does. Their domestic version of TikTok caps kids under 14 at 40 minutes a day, locks access between 6am and 10pm, and swaps the entire feed to educational content. Science, history, museums. The version they export to everyone else? Unlimited, unrestricted, pure dopamine on demand. When kids in the US and China were asked what they wanted to be when they grew up, the number one answer in America was influencer. In China it was astronaut. Macron calls this a cognitive war. Export what dulls young minds and keep what makes them intelligent for your own population. This is the most effective weapon ever deployed against a generation’s ability to think.

@sarthakgh @a16z SaaS is only going to increase but the margins and multiples are going to trend much lower due to increased competition.

Clearest charts I have ever seen to support the SaaS is dead narrative via @a16z

If intelligence is a fluid commodity, the value shifts from the model creators to the orchestrators. Companies that build the "router" or the "agent layer" that can hot-swap these models based on real-time latency and cost will capture the value. @AskVenice is incredibly well positioned here... $VVV

$VVV is the ultimate AI x privacy coin Immutable, tokenized compute The "Swiss Bank" of AI Up another 35% this month

@gmoneyNFT @openrouter @AskVenice That is frustrating… they are very receptive to feedback in their discord

@MrCable0x @openrouter @AskVenice I know I realized that after I bought it. So I'm going to use up these credits and then I'm just going to transfer over to Venice. I will say the Venice integration on OpenClaw hasn't been the best.

New deepseek model coming out next week for Chinese new year expedited to be a huge update. Created an @openrouter key in order to get access to these things faster.

Seems that I'm in a unique position as both a Venice AI and Bittensor bull. Many are falling into the trap of believing the universe of decentralized AI is limited to a single winner. -- Tokenomics -- Venice AI has a truly innovative token design. Voorhees is in the top 0.01% in the world in tokenomics knowledge. As a capital asset token, VVV carries the speculative premium but is directly linked to Venice's product through the utility token DIEM. It's unlike most dual-token projects in crypto because new DIEM is only minted when VVV is locked. Product demand creates deflationary pressure on the capital asset. Bittensor establishes out-of-the-box token guardrails for its subnets based on the most studied tokenomics in crypto: Bitcoin's. Subnet tokens have a fixed supply of 21M tokens and emissions are based on a halving schedule. This basic token standard is necessary for the network to onboard AI-first builders and establish a base layer of trust for new network entrants. A common criticism is that the imposed token supply & emissions framework is too rigid. IMO subnet owners have plenty of room to innovate and customize the token to their needs. Subnets can use their token as a capital asset, utility asset, or a hybrid. They can create a dual token structure vis a vis DIEM. They can adjust burn rates, attach tiered utility and layer on additional token incentives. Token design is unexplored white space and I look forward to subnet owners creations here. *Only* by standardizing supply across the subnet tokens though can they have the power of the entire TAO network and liquidity behind them. -- Protocol vs. App Layer -- Venice is an exceptional application. One of the best non-DeFi applications I've ever seen come out of crypto. Non-crypto users can easily onboard and get value from the inference product. It serves consumers and developers. I believe it will eventually serve larger businesses too. Venice owns their customer experience and VVV will benefit. Bittensor is a protocol that incentivizes open, constant competition between *and* within subnets. The protocol captures value from the applications built on it via the enshrined DEX liquidity pools. Every subnet token is paired against TAO as protocol-provided liquidity. The protocol creates the competitive environment. The quality of applications (subnets) benefit from the fierce competition. The value flows back into TAO. See our "Fat Stack Thesis". -- Outlook -- Bullish. More attention to the DeAI space. More product-focused building. More products using decentralized infra at the base layer. Big season coming up for crypto x AI.

The ideal store of value asset for an agent is tokenized compute. The winning protocol here will be a top 10 cryptoasset

@nateliason Wrote an article about this in January last year -

How funny would it be if there’s another NFT run because people want cool PFPs for their agents. Only 10,000 agents can have punks…