Ep. 101 - Seby Rubino
Seby
@sebyrubino is building Instant
@thinkinstant, Subnet 46
Timestamps
1:30 - Highs, Lows & Staying Sane in Bittensor
2:28 - Token Equity vs. Subnet Tokens
5:16 - Resi to Zipcode to Instant
8:37 - Monetizing Miner Emissions
10:20 - Research vs. Revenue: The Nova Case
12:05 - Why Accelerated Inference
12:58 - Groq, Cerebras & the Taalas Demo
16:29 - Why Inference Speed Failed Before
17:14 - Miner Collateral & Alpha Slashing
19:21 - Beyond TEEs: TOPLOC & Managed VNet IR
22:44 - The Midjourney Discord Lesson
26:36 - Does Anyone Actually Pay for Privacy
29:39 - Picking a Model from Open Router Demand
31:36 - NVIDIA Inception & Hardware Access
32:46 - Exclusive: Tokenized GPU Lending
37:03 - Financing Terms: B300 Pods & tinybox
38:41 - Competing With Banks & Venture Capital
40:35 - Delaware SPVs, 50% LTV & Defaults
42:36 - Lender Risk if Hardware Prices Fall
45:10 - Is Compute in a Speculative Mania
45:42 - Anthropic's IPO & OpenAI's Exit
48:33 - One-Bit Models & Frontier Intelligence
51:31 - Ternary Models, Extropic & Memristors
52:48 - Frowned Upon to Normalized in Bittensor
53:54 - Buccaneer, Not Pioneer: Starting on Third Base
57:06 - Filtering Noise as a Subnet Builder
1:00:32 - The Exploit Debate & What's Next