Feed — Buzzberg

For most investors, buying TAO provides a diversified 'mutual fund' exposure to the entire Bittensor ecosystem, similar to investing in a venture fund. Picking individual subnets is more like angel investing and requires significant research. He notes that subnet investing is open to non-accredited investors, unlike traditional startup equity.

Invests in specific Bittensor subnets that demonstrate product-market fit, revenue, and mature teams, similar to evaluating traditional startups. Examples include Ridges, Shoots, Hippius, Vid.io, and Score. Also invests in research subnets like NOVA (pharmaceutical molecule discovery) which are highly speculative but could be enormous if successful.

BitMind (Subnet 34) is an AI security company focused on deepfake detection and proof-of-human verification. It uses a decentralized, permissionless system where miners compete to develop detection models and generate data to fool them, creating a continuously learning system. The company has raised $3M from VCs and targets enterprise and consumer use cases in media, social media, and government.

IOTA (Subnet 9) aims to orchestrate distributed compute worldwide to train frontier-scale AI models at significantly lower cost. It uses interruptible compute from sources like idle laptops or data center surplus, allowing training with 'fractal compute' (short, variable intervals). The project offers a train-at-home application for easy participation and targets the high cost and inefficiency of centralized training.

Bittensor (TAO) is a decentralized network that allows for global competition to provide AI services at the lowest possible price, which is deflationary. The current governance issues, like the Covenant AI departure, are part of the learning process and will lead to improved subnet governance, making the network stronger. Buying TAO is like buying an ETF into the Bittensor ecosystem, providing exposure to decentralized global intelligence.

The speaker's tool, "Death by Claude," analyzes companies for AI replaceability, scoring them. It identifies "AI wrapper" businesses (simple SaaS layers on top of a model) as highly vulnerable, giving examples like a code-generation startup a 78/100 "dead" score. As underlying AI models (e.g., Claude) become more capable and directly accessible, middle-layer "wrapper" services that don't add significant unique value, data, or network effects can be easily replaced by a prompt or a fine-tuned SLM. Investors should AVOID undifferentiated "AI wrapper" companies in the consumer and business services space, as they face existential, margin-crushing competition from the very platforms they depend on. A wrapper company develops a profound network effect, proprietary data flywheel, or deep regulatory integration that cannot be easily replicated by a base model, creating a sustainable moat.

The speaker states Anthropic's new "Mythos" model is so powerful at hacking software it's essentially a "cyber weapon of mass destruction," leading the company to withhold public release and work only with a consortium of critical partners. This capability creates an existential, two-tier dynamic in the AI race and national security. The speaker infers this forces a conversation about nationalization and covert government use, comparing it to the Manhattan Project. The situation demands close monitoring (WATCH) because it represents a pivotal, high-stakes moment for the company, the AI industry, and geopolitics, with unpredictable outcomes for valuation and strategy. The core claims about the model's danger could be overstated for IPO narrative purposes. An open-source model could achieve parity faster than expected, undermining the strategic advantage.

The speaker argues that Small Language Models (SLMs) are rapidly improving in "intelligence density" and will be capable of handling 90% of common enterprise work tasks by 2030, at a dramatically lower cost than frontier LLMs. This enables massive cost savings (citing AT&T cutting costs by 90%) and could lead to "hyperdeflation" in AI inference pricing. It empowers small teams to serve niche markets profitably, potentially eroding the economic moat of frontier model providers. The entire technology services sector built on AI applications should be watched closely, as the underlying cost and accessibility of intelligence are shifting, enabling new business models and threatening incumbents reliant on expensive API calls. Frontier models continue to advance at a pace that maintains a significant capability gap for complex, novel tasks that SLMs cannot handle, preserving their premium pricing power.

Jason Calacanis explicitly states that Apple has not released any inspiring products since Steve Jobs' death, citing specific failures like the lack of AR glasses, cancellation of self-driving cars, and Siri's incompetence. Without continuous innovation and visionary leadership, Apple is milking past successes and risks losing competitive edge in future markets such as AI and hardware. AVOID because the company appears to be stagnating and not addressing emerging opportunities, making it unattractive for growth-oriented investment. Apple's strong financial position, ecosystem loyalty, and potential for strategic acquisitions could offset innovation gaps.

Salen stated the energy grid is "broken" and the U.S. has flat energy generation growth vs. China's 10x, creating a critical need. He invested in Base Power, which uses distributed home batteries for grid arbitrage and resilience. Fixing energy generation and distribution is a prerequisite for economic and national goals. Hardware-focused companies that rethink fundamental infrastructure (generation, distribution, storage) are addressing the "largest market in the world" with little recent innovation. LONG on the energy minerals/infrastructure sector, as it's a critical, massive market being reopened for innovation through new hardware and technology approaches. Technological hurdles; scaling manufacturing; regulatory and permitting challenges; competition from legacy utilities.

When forced to choose between Anthropic and OpenAI for an all-in investment, Salen said "Anthropic," citing that "all the founders I know are using Claude Code" and it has "captured lightning in a bottle." He perceives strong product traction (Claude Code) among a key user base (founders) and sees it as having momentum, despite acknowledging both companies have issues. LONG on Anthropic based on current product-market fit and developer momentum in a forced-choice scenario against its primary competitor. Rapidly shifting competitive landscape in AI; execution or regulatory challenges specific to Anthropic.

Delian argued that terrestrial wireless carriers (T-Mobile, Verizon, AT&T) have a moat based on physical ground infrastructure, but Starlink's direct-to-phone capability will break that moat. He thinks "all [carrier] valuation gets transferred to SpaceX over the next... 10 years." The advent of space-based broadband (Starlink) with global coverage directly to devices threatens to obsolete the core infrastructure advantage of legacy telecom carriers. AVOID the incumbent telecom carrier sector, as their business model and valuation are under existential, long-term threat from SpaceX's Starlink. Carriers successfully adopting or partnering with Starlink; slower-than-expected rollout of direct-to-cell technology; regulatory intervention.

Delian stated SpaceX has built "the best telecommunications business ever" through Starlink and is the only entity with the capability to take humans to the moon and Mars, with "uncapped potential upside." The company is on the cusp of an IPO, which will realize immense value for early investors (e.g., Founders Fund's 18-year hold) and has proven business lines with future expansion potential. LONG because it is positioned for "huge wealth creation" and has dominant, growing franchises in telecom and space access with a visionary leader. Execution risk in massively ambitious, capital-intensive projects like interplanetary travel. Valuation risk at a $1.75T target.

Larsen stated, "I think that we are [in a defense tech bubble]," citing an "over supply of capital" paying "very, very high prices for things that may not pencil." He compares it to previous hype cycles (crypto, SaaS) where early investors succeed but late capital enters at inflated prices. Winners will emerge, but selectivity and price discipline are critical. WATCH because the sector is hot and topical, attracting excess capital which creates valuation risk. Investors must be highly selective about entry points and business model viability. Companies failing to justify high valuations; dependency on prolonged government budgets and procurement cycles.

The host Alex explicitly states he is considering taking $10,000 (a "stack") to buy and stake TAO tokens as a "learning experiment," expressing positive sentiment toward the Bittensor ecosystem's growth and diversity. Bittensor's value is tied to network utility and adoption. The showcased subnets (drug discovery, creator economy, vision models) demonstrate diverse, high-potential applications that could drive demand for the network's native intelligence marketplace and its TAO token. Direct statement of capital allocation intent indicates a bullish, long-term exploratory view on the asset, based on the belief that the ecosystem is expanding and creating tangible utility. Bittensor is a nascent protocol; the commercial success of its subnets is unproven, and token value is subject to high volatility, regulatory uncertainty, and competition from centralized AI platforms.

Calacanis states he has personally bought ~$500K of TAO and is a partner in a fund (Stillcore Capital) dedicated to the ecosystem. He believes TAO could reach a $500B market cap (a ~200x increase) in 5-10 years, rivaling Solana or Ethereum. He argues Bittensor's model of competitive subnets (e.g., for compute, storage, lead generation) creates a decentralized mechanism to commodify and improve digital services, driving costs down and quality up through global, incentivized competition—a foundational shift akin to cloud computing. LONG because he sees it as the "better Bitcoin" for the AI era—a protocol that turns wasted compute (like Bitcoin's PoW) into productive services—and believes crypto regulation has matured, de-risking the asset class. He is willing to risk total loss on his position for this asymmetric upside. The entire ecosystem fails; subnets don't gain traction, produce low-quality outputs ("slop"), or the technical/cryptoeconomic model doesn't scale as envisioned. High volatility is guaranteed.