Podcast Summary: This Week in Startups — “Bittensor’s (alleged) $10M rug pull” (feat. Mark Jeffrey) | Ep. 2275
Date: April 14, 2026
Host: Jason Calacanis
Guests: Mark Jeffrey, Ken Miyachi, Will Squyers, Stefan Cruz, Lon (co-host and producer)
Episode Overview
This episode dives deep into the recent controversy surrounding Bittensor, a decentralized AI/crypto network, following an alleged $10 million "rug pull" by a prominent subnet operator. Host Jason Calacanis is joined by longtime friend and crypto OG Mark Jeffrey, along with subnet founders Ken Miyachi (Bitmind) and Will Squyers & Stefan Cruz (Macrocosmos). The conversation analyzes the mechanics and implications of Bittensor’s subnet structure, incentive systems, and the vulnerabilities exposed by the alleged rug pull. The episode also showcases exciting subnet projects, debating how decentralized AI and crypto intersect, and wraps up with broader discussions about risk, investment, and even pop culture.
Key Discussion Points & Insights
1. The Bittensor Rug Pull Incident
-
Background & Allegations
- Subnet operator Sam Dare (“Templar”/Covenant AI) allegedly dumped 37,000 TAO (≈$10M) and walked away from the project, impacting holders (13:37, Lon).
- Quote:
- "He dumped 37,000 TAO on his own subnet holders. It's about $10 million in money. And he sort of walked away from the entire project." (13:37, Lon)
- The abrupt move caused TAO’s price to drop ~25% (15:04, Lon).
- His explanation centered on “decentralization” concerns, but most of the community sees it as opportunistic and damaging (21:07, Mark Jeffrey).
-
Mechanics of a Rug Pull in Bittensor
- Subnet owners accumulate sizeable subnet and TAO tokens as rewards, which remain liquid and accessible.
- Quote:
- “If you have enough of them, and there’s liquidity... you can at any time just sort of flood Uniswap with your tokens and exchange it back out for TAO...and now you’re gone. Which it looks like... is what Sam did.” (16:04, Mark Jeffrey)
-
Community Reaction & Root Cause
- Broad repudiation from Bittensor subnet operators—“I am Spartacus” moment reinforcing community integrity.
- Mark Jeffrey calls Sam’s justification a post-hoc rationalization rather than legitimate grievance (21:07, Mark Jeffrey).
- “No, I do not. And you can see this, the evidence, like nobody in the ecosystem believes that.” (21:07, Mark Jeffrey)
-
Systemic Vulnerabilities & Proposed Solutions
- Current flaw is information asymmetry and limited controls—single operators can abscond with funds before anyone notices.
- Various proposals:
- Owner lockups: Smart contracts that restrict immediate access to emission rewards.
- Dynamic “proof-of-conviction” staking: Subnet ownership goes to those who lock up the most tokens for the longest period, thus aligning incentives with long-term value creation (23:32, Mark Jeffrey).
2. Evaluating Bittensor’s Ecosystem and Investment Models
-
Uniqueness & Signals
- Mark Jeffrey praises Bittensor’s early ecosystem as having “more early signal” than any comparable crypto or DePIN (Decentralized Physical Infrastructure Networks) project (03:23, Mark Jeffrey).
- Subnet projects are producing real, valuable hypotheses with strong founding teams—a contrast to much of DeFi’s “nonsense”.
-
Application Layer & Value Creation
- Jason is interested in crypto moving beyond speculation and money transfer, looking for consumer and enterprise use cases (02:24, Jason Calacanis).
- Bittensor’s decentralized AI ambitions are compared to the “Linux moment” in OS history—open-source, distributed, permissionless (09:20, Mark Jeffrey).
-
Investment Approaches: Token vs Equity
- Investing in TAO tokens: Broad mutual-fund-like exposure to the protocol and subnets—ideal for those wanting to participate without the bandwidth to pick individual projects (36:29, Jason Calacanis).
- Investing in subnets: More active, “angel investor” style; potentially higher upside and higher risk (38:15, Mark Jeffrey).
- Traditional venture/equity investment is still possible (e.g. in Bitmind), but harder for non-accredited investors; tokens level the playing field for the public (39:31, Ken Miyachi).
-
Risk & Speculation
- Clear warnings: “90% of the bets I make go to zero” (36:29, Jason Calacanis).
- Importance of time horizon, risk appetite, and understanding the mechanics before investing.
3. Showcasing Bittensor Subnets & Operators
a. Bitmind (Ken Miyachi, Subnet 34) — Deepfake Detection & Proof-of-Humanity
- Mission: AI security for deepfake detection; applications in media, social platforms, government, and enterprise.
- Mechanism: Miners compete to build the best detection models; adversarial “red team” miners generate and attempt to fool systems, creating a rapid evolutionary pressure (26:38, Ken Miyachi).
- Quote:
- “It’s a continuously learning system trying to improve over time.” (26:38, Ken Miyachi)
- Commercial Value: APIs and bots for media/news; business communications security; proving content authenticity is increasingly crucial (29:05, Ken Miyachi).
- Investment & Openness:
- Raised $3 million from crypto and AI VCs (33:28, Ken Miyachi).
- Participatory—anyone (not just accredited investors) can buy subnet tokens and support the project.
b. Macrocosmos (Will Squyers & Stefan Cruz) — IOTA & Distributed Training
- IOTA (Subnet 9):
- Goal: Orchestrate global decentralized compute power (like SETI@Home) to train frontier AI models (44:33, Stefan Cruz).
- Ambition: Enable cost-efficient, large-scale distributed training rivaling the capabilities of data centers at a fraction of the cost.
- Innovation:
- “Train at home”: Install simple software, donate compute, earn tokens passively (46:36, Stefan Cruz; 50:13, Will Squyers).
- Focus on making “meatloaf” with distributed compute, not just “prime rib” datacenter GPUs—democratizing participation.
- Mechanics: Handles highly interruptible, variable, and fractional compute—unlike centralized infrastructures (54:22, Stefan Cruz).
- Quote:
- “We actually want to be able to make... indistinguishable training compute from this mixture of compute nodes.” (46:36, Stefan Cruz)
- Business Model:
- Fully bootstrapped, company-based with deep roots in Bittensor.
- Ethics Post-Rug Pull:
- Strong emphasis on trust and fiduciary duty; owner lockups seen as vital for ecosystem maturity (58:10, Will Squyers).
Notable Quotes & Insights (with Timestamps)
-
On the Rug Pull Allegation:
- “If you have leadership and you have responsibility, you can mess with it.” – Jason Calacanis (00:06)
- “…the incentivization alignment engine worked spectacularly well, but then it fell out of alignment when there was sudden success. It created this sort of overwhelming temptation.” – Mark Jeffrey (00:15, echoed at 16:04)
- “It looks like… this is what Sam did.” (00:32 & 16:04, Mark Jeffrey)
- “He dumped 37,000 TAO on his own subnet holders. It’s about $10 million in money. And he sort of walked away from the entire project.” – Lon (13:37)
- “Do I believe that Sam’s epistle about why he left because of decentralization is correct? No, I do not.” – Mark Jeffrey (21:07)
- “The Bittensor version of ‘I am Spartacus.’” (21:25, Mark Jeffrey)
-
On Smart Contract Solutions:
- "The architecture in crypto is smart contracts. ... The answer is to lock those tokens so he can't spend them. ... If you’re operating the subnet, then your tokens are locked." (23:32, Mark Jeffrey)
- “The chain will reward capital conviction in a subnet.” (23:32)
-
On Investment Models:
- “You can own the equity in Ken’s corporation, or you can own the tokens. ... It just depends on which you believe in more.” (35:20, Mark Jeffrey)
- “There are people who want to back others to go play poker tournaments, and those who want to play in the tournaments.” (36:29, Jason Calacanis)
- “90% of the bets I make go to zero. So you need to have a certain constitution…” (36:29)
- “If you’re a normal person in the normal world and you’d like to play in this arena, it is open to you; you don’t have to be accredited.” (38:15, Mark Jeffrey)
-
On Incentives:
- “This is the most ferocious form of capitalism ever invented, as far as I can tell.” (42:01, Mark Jeffrey)
- “[Bittensor’s] incentive alignment engine is its core invention and everything within it is contests within contests within contests.” (42:01, Mark Jeffrey)
- “It's a way for you to invest in the AI future as an unaccredited investor that you normally don't get.” (38:15, Mark Jeffrey)
-
On Distributed Compute (IOTA):
- "We want to be able to make... indistinguishable training compute from this mixture of compute nodes." (46:36, Stefan Cruz)
- “Just download this thing, set it up, and now you’re mining Bittensor.” (45:49, Mark Jeffrey)
Important Timestamps
- 00:00–00:51: Rug pull intro & context
- 04:32–06:17: Uniswap/DeFi explanation and significance
- 13:37–15:08: Details of Templar subnet, Sam Dare’s actions, price drop
- 16:04–18:46: Smart contract loopholes, possible remedies
- 21:07–22:42: Discussion on Sam Dare’s stated grievances vs real motives, proper resignation process
- 23:32–24:39: Smart contract upgrades & new governance proposals
- 25:24–30:53: Bitmind project (deepfake detection), adversarial mining, enterprise/government use cases
- 36:08–40:14: Investment vehicles for public vs accredited investors
- 44:33–54:42: Macrocosmos, IOTA, distributed training, democratizing compute
- 57:19–58:49: Emotional/ethical aftermath of the rug pull; solutions going forward
Tone & Atmosphere
The conversation is lively, technical but accessible, and often candid—reflecting both the optimism and wariness of veteran entrepreneurs when navigating new, rapidly evolving ecosystems. Frequent analogies (Linux, venture capital, poker tournaments, Airbnb) help frame the concepts. The mood is simultaneously analytical and community-minded, with a strong bent toward practical risk management, trust, and incentive alignment.
Memorable Moments
- Metaphor for crypto investing:
- Backing a fund (buying TAO) vs. being an active angel (buying specific subnets).
- “There are people who want to back others to go play poker tournaments, and those who want to play in the tournaments.” (36:29)
- On Bittensor’s capitalist Darwinism:
- “Contests within contests within contests.”
- Train-at-home analogy:
- “Ribeye steak” (prime GPU) vs “meatloaf” (fractal distributed compute). (46:36, Stefan Cruz)
- Community's “I am Spartacus” moment in the wake of the rug pull — subnet operators rallying to reaffirm trust.
- Meta discussion:
- Comparison of spectacle vs. authenticity in art/entertainment during "off-duty" segment (Bieber vs. Sabrina Carpenter at Coachella; 65:10+).
Conclusion & Takeaways
- The Bittensor rug pull exposed important architectural and governance vulnerabilities common to decentralized systems—information asymmetry, liquidity, and incentive misalignment.
- The problem is not unique to crypto, echoing early failures in venture/startup history (“shutting down and taking the money to Vegas”).
- The remedy is likely institutionalized via smart contracts — requiring “skin in the game” and minimizing single-actor risk.
- Despite setbacks, the Bittensor ecosystem is showing robust early signals and cultivating a unique, meritocratic, and incentive-driven environment for decentralized AI.
- Individuals can participate in multiple ways — from passive mutual-fund-like vehicles (TAO) to directly engaging with/operating subnets or funding corporations when possible.
- The episode underscores the importance of trust, community, and aligning long-term incentives in all forms of decentralized, collaborative innovation.
For listeners:
If you’re interested in the intersection of AI, crypto, and decentralized governance—or just want to understand how these rapidly evolving models work—this episode provides both a cautionary tale and a showcase of bleeding-edge innovation.
