Meeting Summary: BitTensor Deep Dive and Propaganda Subnet Brainstorm

Overview

Extended dinner and strategy session with Grant El-Deir (friend from Capital Factory, met February 2025) at a Korean restaurant and then Gary's house. Covered Grant's Gittensor subnet progress, BitTensor ecosystem analysis, life updates from both sides, and deep collaborative brainstorm on a "propaganda subnet" concept for incentivizing viral content creation. Gary and Grant identified strong alignment on game theory thinking and agreed to collaborate on subnet development.

Key Topics Discussed

Grant's Gittensor Subnet Update

Current Status:

• Top 30 subnet by emissions, paying out $150,000+/month to miners
• Top miner makes $40,000/month
• Team just became full-time as of this week
• Market cap around $30-35 million
• 97% success rate for code generated to code used in production
• 166 followers on subnet X account (minimal marketing, relies on PR attribution)

Technical Approach:

• Every PR requires "contributed by Gittensor" tag with website link for organic discovery
• Tier system: must make 3 PRs to 3 separate repos in "forums" before earning
• 70% credibility threshold required (merged PRs vs closed PRs)
• Building ELO rating system for coders similar to chess rankings
• Uses validators as judges, miners as participants, subnet owner as competition designer

Key Quote:

"We have a 97% success rate. So it's pretty incredible to have that high code generated to code used in production."

BitTensor Ecosystem Critique

Grant's Assessment:

• Most subnet owners are "first generation" who made money early and are now complacent
• Robin and others just buy compute and resell it - "no innovation or skills required"
• Only real "second generation" builders: Grant, Andrew Landon, and one other (Shaq)
• General vibes down because price is down
• Token value fundamentally about directing the "workforce of the world"

Why BitTensor Works (Grant's Explanation):

"Tau is denomination of value. Subnets - these competitions - the value of that competition is priced in Tau. So if someone has a competition for cups and someone else has a competition for identifying deepfakes, there is no objective way to say which one is more valuable. This allows for the value of these subjective things to be priced directly against each other."

"The more Tau you have the more you can say 'Hey, the world needs more of these cups' or 'Hey, the world needs more of these deepfake image detections.' You're basically able to direct the workforce of the world."

Proper Subnet Design:

• Must be 100% skill-based, not just capital-based
• Should benefit from permissionless competition
• "Anything that can benefit from a permissionless competition should be a subnet"

Propaganda Subnet Concept ("GSN")

The Opportunity:

• Propaganda/narrative control is "arguably the most valuable thing in the world"
• Current content creator funds (TikTok, YouTube) are essentially existing subnets but per-platform
• Layering on top of all social media creates unprecedented narrative control
• "Unbannable" since no single account - just an invisible hand incentivizing content

Proposed Mechanism:

1. Fork BitCast (subnet 93) as starting point
2. Change from rewarding all creators to rewarding top 3 per "brief"
3. A "brief" is the narrative/topic to push for a set period (e.g., 2 weeks)
4. Reward based on views via social media API calls
5. Extreme virality requirement eliminates bot problem ("can't bot 100 million views")
6. Start with TikTok (easiest for new accounts to go viral), expand to all platforms

Key Innovation:

"You control the narrative on all social medias. Because we're in a society now that's driven by the extremes. And we're going to lean into that by saying we're going to reward the extreme virality of this propaganda."

Gary's Vision:

• Could start with Christian-aligned propaganda
• Or start amoral/opportunistic to prove concept, then branch Christian-specific subnet
• Gary as game master picking narratives, holding significant alpha tokens
• Mike Ajouz (wealthy Christian funder) could be interested in backing

Grant's Enthusiasm:

"I think you are genuinely probably the perfect person in the world and perfectly positioned to do this."

"You have every attribute that one could desire to be good at doing this thing. And you have the experience on the player side of it. You've been in that competition before."

Funding:

• Subnet slot costs ~250 tau (less than $100,000)
• Grant willing to invest or partner
• Could potentially get free slot from community member who bought one to give away
• No external capital strictly needed

Gary's Activism Writing Connection

Gary played back old blog posts from 2023 about propaganda and "Activism 2.0" (massively multiplayer movement building games):

Key Insights from Gary's Previous Writing:

• "Honest propaganda" vs dishonest - the word itself has been propagandized negatively
• Participatory propaganda creates ownership and momentum
• Movements need clear game masters, structured competition, ELO-like rankings
• Best historical example: Office of War Information in WWII

Grant's Reaction:

"This is almost like a blueprint for what needs to be done. Even the language of it - incentivized. It's freaky how... I hear the word synchronicities."

Life Updates

Gary:

• Got baptized a few months ago (Gulf of Mexico, near Houston)
• Manufacturing startup idea with pastor didn't work out
• Now doing Applied AI community work with Travis Oliphant (event January 29th)
• Staying with housemate who does Andrew Huberman podcast editing
• Dating someone who works at NSA (visiting tomorrow)
• Strong spiritual journey, seeing God's hand in reconnections and timing
• Read and wrote extensively about propaganda theory in 2023 without knowing why

Grant:

• Team just went full-time this week
• Moving to Waterline (74-story building) once ready
• Started modeling on the side 3 months ago
• Father lives in Rio; planning visit
• Brother graduated, now trading and building course with friend Dylan
• Sleeps 9 hours/night (11pm-8am), credits creativity to isolation
• Podcast at episode 78

Other Business Ideas Discussed

Grant's Previous Ideas (not pursued):

• Chalpedia: Wikipedia alternative with incentivized accurate information
• Lunch Club: Meet-up group for downtown workers → upsell beef and rice delivery
• Hardware innovation subnet: Benchmarks and competitions for physical hardware improvement (Mike and his friend hated this idea)

Random Thoughts:

• OnlyFans crypto clone (how Gary and Grant originally bonded)
• Decentralized church concepts
• Incentivizing school choice voucher signups
• Converting celebrities through content without their knowledge (measuring influence)

Geopolitical Discussion

Brief discussion on:

• Venezuela/Maduro situation - US being "refreshingly honest" about taking over
• Tariffs and their effects
• George Soros as antichrist figure (Gary's contact Mike Ajouz partied with him)
• Nick Fuentes blowing up
• Jews and their PR problem with Israel/AIPAC ("Jews are making it really easy to hate them")

Vibe Coding & AI Tools

Both agreed on transformation:

• Building has become almost purely imagination-level
• "The worst it will ever be" mentality when using Bolt.me 1.5 years ago
• Grant vibe codes all day making internal tools
• Gary revisited and improved old tools he built
• Demographic that benefits most: creative people who couldn't code before

Grant's Quote:

"All I do is vibe code now. Like all day. I went from sitting on the charts all day to just vibe coding all day. For the past like six months at least."

Key Quotes

On BitTensor's Potential:

"It's just more efficient economics. So there was capitalism. And then the next thing is going to be incentivism."

On Propaganda Value:

"Propaganda is... this is the case for it being the most valuable thing possible. Because if you apply it to children. If you create what they desire in life. You now have programmed people to now dedicate their life to a specific mission."

On Subnet Design:

"The way to build a subnet is make it 100% skill-based. So it's like, actually, innovation can occur and it benefits from being a permissionless competition."

On Gary's Unique Positioning:

"You have the vibe code, you have the virality experience. You have like all the pieces for it. And it's just like, everything's coming together at the perfect time."

On Social Validation:

"I care about monetary validation... I don't need someone else to tell me I did good or I did bad. It's like, I understand that I did good or I did bad."

On Faith and Future:

• Gary: "I think I'm going to be one of the most influential Christians in the world."
• Gary: "God has great plans for me... there's no way that the story goes: Gary hung out with a lot of very successful people and then he crashed out."
• Grant wants huge family (jokingly said 50 kids)

Action Items / Takeaways

1. Gittensor Onboarding: Grant to send Gary docs and website; Gary to potentially become miner making agent-based contributions (Sunday afternoon call planned)
2. Propaganda Subnet Development: Gary to write out proposal brief; Grant willing to invest, partner, or provide technical support
3. Applied AI Live Event: Grant registered for January 29th event
4. Continued Collaboration: Aligned on game theory thinking and subnet philosophy; Grant considers Gary "perfect person" for propaganda subnet
5. Mike Ajouz Pitch: Gary to potentially present subnet concept to Mike for Christian propaganda funding

Spiritual/Philosophical Themes

Gary's Framework:

• God reconnecting people at exactly right moments
• Recognizing patterns in seemingly unrelated past work (2023 propaganda writing)
• Faith in God's plans despite setbacks
• "I laugh with God more and more these days"

Grant's Philosophy:

• Isolation breeds creativity by avoiding "contamination of conditioning"
• Monetary validation over social validation
• Everything is a game to be optimized
• Not contaminated by preconceived notions about how BitTensor should work

Shared Values:

• Thinking big, executing patiently
• Teaching kids game theory/incentive design
• Building Christian community for families
• Taking action over analysis paralysis