
AI Agents, Zcash Mania & Crypto’s Ownership Crisis
AI Summary
The Delphi Hivemind podcast covers market trends, AI, and crypto, with an unfiltered perspective. The current market sentiment is generally positive, with stocks and crypto showing upward momentum, seemingly shrugging off geopolitical concerns in the Middle East. This resurgence in risk appetite is reflected in tech and AI stocks hitting new all-time highs, and a renewed interest in crypto, particularly in meme coins and AI-related tokens. However, underlying concerns about inflation persist, with bond yields rising, indicating pressure in that sector.
A significant discussion point is MicroStrategy's (MSTR) recent earnings call, where CEO Michael Saylor entertained the idea of selling Bitcoin. This was a surprising development, as Saylor has historically been a staunch advocate for holding Bitcoin. The rationale behind this potential sale is linked to MSTR's preferred stock (STRC) and its ability to sustain dividend payments. The STRC structure involves repurchasing shares to maintain a peg, which requires significant capital. Saylor's mention of selling Bitcoin is seen as a way to generate cash to support the STRC dividend, thereby instilling confidence in the STRC holders and the overall MSTR strategy. This move is viewed as a positive signal, suggesting a strategic approach to capital allocation rather than a sign of distress. The market appears to be interpreting this as a net positive for both MSTR and STRC, as it opens up a new source of funding without significantly impacting Bitcoin's overall supply.
Beyond MSTR, the conversation shifts to specific altcoins, with Zcash (ZEC) being a prominent topic. Despite past underperformance, Zcash is seeing renewed interest due to its long proof-of-work distribution, a recent breakout, and the emergence of narratives around privacy coins. The involvement of Multicoin Capital, a prominent venture firm, in building a position in Zcash is seen as a high-conviction signal. The resurgence of privacy as a theme in crypto, with Monero and other privacy coins also performing well, supports the bullish outlook for Zcash. Its strong fundamentals, including a well-distributed token supply, lack of VC dumping concerns, and a robust development team, are highlighted as key strengths. The potential for Zcash to become a significant altcoin winner, similar to past narratives like Solana and Hyperliquid, is discussed, with institutional interest and strong technology as further catalysts.
Another altcoin gaining attention is VVV (presumably a ticker for a specific project), described as a fundamentally strong business with significant ARR (Annual Recurring Revenue) growth, driven by AI-related tailwinds rather than purely crypto speculation. Its relatively low market cap and high FDV (Fully Diluted Valuation) suggest it might be under-owned. The project is focused on flowing value through a token, with a strong emphasis on reinvesting in growth and avoiding VC unlocks that could lead to dumping. Its clean chart performance year-to-date, especially its AI overlap, makes it an attractive proposition, though some argue that pure AI plays or equities might offer clearer exposure to the AI narrative.
The discussion then delves into the complexities of AI and decentralized AI infrastructure, specifically mentioning Venice. The speaker expresses skepticism about Venice's product, questioning the fundamental value proposition of a de-anonymization layer for AI when user context, not just prompts, can leak sensitive information. While acknowledging potential niche use cases like adult content, the speaker believes it doesn't make sense for most privacy-conscious users. The conversation touches upon the pass-through model, where revenue generated often goes to underlying model providers, drawing parallels to companies like Cursor. The acquisition of Cursor by SpaceX for a reported $60 billion is presented as a significant validation of the AI model business, despite the revenue pass-through concerns. The rationale behind SpaceX's acquisition is seen as a strategic move to bolster their own AI capabilities (Grok) and acquire talent and data for their composer model, aiming to compete in areas like coding, which is identified as a potential killer app for AI. The valuation of Cursor is contextualized against the high valuations of OpenAI and Anthropic, suggesting it might not be as exorbitant as it initially appears.
The conversation shifts to agent harnesses and the evolving landscape of software built for agents. Hermes is highlighted as an example of an agent harness gaining traction, potentially stealing market share from competitors. The idea that agents are becoming the primary interface for interacting with services, rather than humans directly, is explored. Software products that are "agent-first" are seen as the future, with companies like Parallel, a search engine for agents, emerging as examples. The challenge of collaboration is also discussed, with the concept of agents communicating with each other rather than humans directly. This paradigm shift suggests that traditional SaaS models built for human interaction might become obsolete, with value accruing to backend infrastructure and agent-centric software. The difficulty of building defensible moats in the AI model space is contrasted with the potential for strong businesses in the agent orchestration and services layer.
A significant portion of the discussion is dedicated to the Houdini Swap acquisition and its implications for token holder rights. The core issue is the perceived rugging of token holders when Soul Strategies acquired Houdini Swap's equity without any indication of valuing or considering the token. This event is seen as a stark example of the lack of liquid funds and the inherent risks associated with early-vintage crypto tokens. The speaker criticizes the dual equity token structure and emphasizes the need for protocols that enforce token holder rights, such as Meta-X or similar models. The argument is made that without such protections, confidence in the crypto market will remain low, hindering the growth of liquid funds. The speaker believes that market participants, including venture firms and developers, need to be punished for participating in such rug pulls to incentivize better practices.
The conversation then moves to Pump, another project that recently announced a reduction in its buyback program from 100% to 50% of revenue. While the burn of bought-back tokens is considered irrelevant, the reduction in buybacks has sparked debate. The speaker acknowledges that companies often retain earnings for growth, but emphasizes the importance of clarity and a long-term commitment to token holders. The lack of a legally enforceable claim on cash flows for token holders, unlike equity holders, is a recurring concern. The potential for Pump to adopt an "ACE" (Acquisition, Conversion, Equity) model, similar to what Meta-X offers, is suggested as a way to provide greater security and transparency to token holders. The speaker also touches upon Pump's revenue generation, its origins in meme coin trading, and the challenge of sustaining growth in a rapidly evolving market. The difficulty of predicting the future of AI and software, especially when it comes to agent-centric applications, is a concluding theme.