Dropping a podcast with @vikramskr, writer of Vik's Newsletter and co-host of Semi Doped, on AI agents, memory, and the hardware behind the data center buildout. Buying more GPUs is only one piece. Getting data to them fast enough is another. Vik explains why memory and interconnects can leave compute sitting idle, why AI data centers need more optics, and how training and inference create different hardware demands. We also get into a question I found interesting. Do agents always need the highest possible speed, or should we be optimizing for how much useful work they get done per token? In this conversation we discuss - AI agents and what they actually need to run - AI spending and the risk of overbuilding - Training versus inference workloads - Copper, optical interconnects, and co-packaged optics - SRAM, DRAM, high bandwidth memory, and storage - Useful work per token and agent context - New inference chips and the memory bottleneck Timestamps 00:00 Highlights 01:18 Vikram Sekar and Semi Doped 02:55 AI agents and hardware demand 18:57 AI spending and the risk of overbuilding 25:10 Training versus inference 36:25 Why distance limits copper bandwidth 41:18 Co-packaged optics and the push beyond copper 46:04 New approaches to optical interconnects 52:24 The AI memory hierarchy 56:39 Useful work per token 57:51 How much context does an agent need 1:03:35 DRAM, HBM, and memory on logic 1:05:54 The next generation of inference chips 1:08:25 AI's biggest engineering bottleneck Enjoy.

Dropping a podcast with @armaniferrante, Founder and CEO of @Backpack, on tokenized stocks and the future of global capital markets. Tokenizing a stock is only one piece. The bigger opportunity is connecting it to the liquidity and price discovery that already exist in public markets. Armani explains why redeemability matters, why Backpack chose the licensed path, and how the team is trying to build one global liquidity layer. We also get into Backpack's own BP token, a potential IPO, and the equity-conversion option Armani describes for contributing token holders. In this conversation we discuss - The different types of tokenized stocks - How redeemability connects Solana with traditional stock markets - Licensing and one global order book - Hyperliquid, Kraken, and regulated derivatives - Competing with the largest exchanges - The BP token, growth milestones, and optional equity conversion Timestamps 00:00 Highlights 01:26 Armani Ferrante and Backpack 03:18 Why tokenized equities matter 13:05 The different models for tokenized stocks 19:33 Backpack's approach to tokenized stocks 28:58 Backpack's place in the market 31:47 Why Backpack chose the licensed path 35:25 One global liquidity layer 40:00 Hyperliquid, Kraken and derivatives 49:29 Competing with the incumbents 53:47 Why a token launch is a starting point 1:00:04 Backpack's three allocation buckets 1:02:31 The treasury lockup and a potential IPO 1:07:07 The option to convert BP tokens into equity Enjoy.

Dropping a podcast with @pequityresearch Mr. P has been doing some great work breaking down the AI buildout and following where hyperscaler capex actually goes. In this episode I wanted to walk through the full stack with logic, memory, power, and networking. We get into why memory could become the largest line item in the AI bill, what old GPU rental prices tell us about compute demand, and where value is going to accrue as the physical constraints get harder to solve. At the center of the conversation is his view that AI spending cannot grow forever. Long-term contracts might change the shape of the next memory downturn, but they don’t eliminate the cycle. And signing a 10-year contract does not mean anyone can actually see 10 years of demand. We also get into why he’s excited about optics, where NAND and HBF fit as agents use more memory, and his views on Chinese open source and the future of US model development. We discuss: - Why he thinks 10-year demand visibility is bullshit - Memory’s growing share of hyperscaler capex, and why estimates vary so much - Why older GPUs are still renting and what that says about compute demand - How long-term agreements, pricing floors, and prepayments actually work - Power as a bottleneck, and why identifying a constraint isn’t the same as finding an investment - Copper vs optics, and where networking value accrues - Agents, NAND, and where HBF fits in the memory hierarchy - CXMT, Chinese open source, and the risks of slowing frontier model development Timestamps: 0:00 – Why AI Spending Can’t Grow Forever 1:13 – P Equity Research’s Background 5:48 – Where Hyperscaler Capex Actually Goes 8:00 – Is Compute Still Tight? 12:08 – Memory’s Share of the AI Bill 17:20 – Why the Memory Cycle Isn’t Dead 18:37 – Inside a Long-Term Agreement 28:30 – What Happens If Customers Cancel? 32:17 – The Rising Cost of the AI Buildout 33:43 – Copper vs Optics 38:49 – Power and Gas Turbines 39:29 – US Models and Chinese Open Source 42:36 – Agents and NAND 47:00 – Where HBF Fits 52:23 – What P Is Most Excited About 56:27 – CXMT and China’s Memory Industry 1:04:00 – Closing Thoughts Enjoy!

If just 1% of the ~$69T U.S. stock market moved onchain, that would be ~$690B in tokenized equities Roughly 275x today’s ~$2.5B onchian tokenized stocks That’s one asset class, in one country. Now add bonds, commodities and derivatives We’re still very early

My conversation with @tarunchitra As a co-founder of Gauntlet and GP at Robot Ventures, Tarun has one of the sharpest frameworks for understanding market structure across both crypto and AI. In this episode we dig into why open source AI is unbundling faster than most people expect, and how the resulting stack looks surprisingly similar to DeFi. We spend time mapping the AI infrastructure layers directly onto crypto primitives and examining where value is actually going to accrue as models, harnesses, routers, and inference providers separate. At the center of the conversation is the belief that AI’s unbundling is creating a new competitive order-flow market (data centers competing like nodes, MEV-like dynamics for tokens/GPUs) while crypto itself has settled into a more mature “TradFi plus+” phase focused on trading, payments, and bringing real assets on-chain. We discuss: - The current state of crypto as TradFi+ and the decline of speculative narratives - Why AI is killing Bitcoin mining economics and weakening ETH value accrual - The architectural parallel between AI stacks and DeFi (Harnesses = Wallets, Routers = DEX aggregators, Models = Protocols, Inference Providers = LPs) - Why open source models are unbundling faster than traditional software - Agents as the next interface layer and the potential unbundling of ETFs -Cryptography, verifiable compute, and turning GPUs into digital assets - Onchain compute trading as the real crypto × AI opportunity - Sustainable business models and where value will ultimately capture Timestamps: 0:00 – Introduction & State of the Crypto Market 2:00 – Speculative Narratives Fade, Trading & Payments Remain 7:00 – AI’s Impact on Bitcoin Economics & Data Center Opportunity Cost 9:00 – ETH Value Accrual, Solana Positioning & DeFi Token Sustainability 15:00 – Trading Design Space & On-Chain Volume Upside 25:00 – AI Unbundling Thesis: Open-Source Models vs Data Centers 35:00 – The DeFi Mapping (Harnesses, Routers, Models, Inference) 48:00 – Agents, Preference Expression & Unbundling Traditional Products 1:00:00 – Real-Time Harness Generation & Active Learning 1:05:00 – Cryptography, Verifiable Compute & On-Chain GPU Markets 1:10:00 – Closing Thoughts: Where Value Accrues Next Enjoy!