Sabitlenmiş Tweet

Sat down with @0xDeltaHedged [ Cofounder of @OpenGradient ].
He worked at NASA, Google, Two Sigma, Palantir... had a career people dream of.
left it all to build something bigger.
Conversation started with AI and data privacy.
but ended with an answer we've all been asking: who owns your digital life?
Verifiable AI, data ownership, the future of agentic commerce... we went deep on all of it.
it's a kind of founder conversation that reminds you why this space exists.
Timestamps:
00:00 — Intro: No one should trade a lifetime of memories for AI output
00:32 — Welcome to the Base Show + Matthew's background
01:06 — NASA → Google → Two Sigma: the psychological shift to building
02:43 — What is Open Gradient? Making AI verifiable explained simply
04:10 — Is the data privacy threat real? The Cambridge Analytica parallel
05:28 — Cursor's Chinese AI model controversy & hidden threats
06:42 — "No one should trade a lifetime of memories for AI output"
07:37 — MemSync: aggregating your digital footprint privately
08:06 — Is decentralized AI actually decentralized? The Bittensor debate
09:01 — OPG token launch: what it unlocks
10:22 — Traction: 4,000+ models, Sentient, Flock, Filecoin partnerships
12:02 — Breaking it down for the average Joe: how OPG protects you
13:34 — The HTTPS standard for AI: Open Gradient's endgame
15:14 — Can everyday builders fight AI monopolization by big tech?
16:33 — The aha moment: DeFi was primitive, ML could fix it
18:21 — Uniswap research grant: reducing impermanent loss 1,670% annually
19:47 — Who's backing Open Gradient: a16z, Coinbase Ventures, Ilya, Sandeep
21:14 — The next 12 months: mainnet, decentralization, model incentives
22:51 — Why agentic commerce isn't hype — but also isn't ready yet
26:16 — Sahib's marketing agency moment: how do you market TO agents?
28:33 — The Virtuals partnership and Titan Launch Program explained
31:00 — How verifiable AI makes agentic commerce trustworthy
33:22 — Matthew's personal AI stack: Opus, Codex, Kimi K2.5
36:13 — The belief Matthew got wrong: LLM intelligence would plateau
38:46 — Dark truths of AI: job replacement and sycophancy
42:14 — The most dangerous company in AI centralization (his answer may surprise you)
44:14 — Will decentralized AI win — or is big tech going to dominate?
46:23 — One piece of advice for young builders entering AI
50:01 — Speed and relationships: the only real moat left
53:30 — Token launch week: the vibe on the ground in Seoul
55:00 — Closing thoughts + follow-up in 4-6 months
Enjoy the podcast.
English



















