Bitcoin Takeover Podcast podcast

S17 E52: Michael Sutton on Kaspa, Covenants & Research

0:00
2:59:03
Rewind 15 seconds
Fast Forward 15 seconds

Michael Sutton has been with Kaspa for eight years: first as a researcher, then as the engineer who rewrote the entire codebase in Rust and unlocked 10 blocks per second. He is the co-author of the DAG Knight paper with Yonatan Sompolinsky, and in this conversation we cover the full arc of his work: the early L2 research that predated Kaspa's launch, the post-launch stabilization, the Tokata hard fork that brought covenants to L1, and the current sprint toward a DAG Knight testnet by end of year.

The centerpiece is a masterclass in covenants. Michael defines them precisely (authority over a UTXO that is not unlimited), explains why recursive proofs for lineage become a nightmare at scale, and presents KIP20, Kaspa's covenant ID system that solves the lineage problem locally. He then walks through the two compilers built on top: Silver Script for single-contract covenants and Argent for multi-contract, multi-app systems with atomic cross-covenant operations. The chess implementation that started as a design challenge became the proof of concept for the entire multiplex architecture.
We also discuss the year of hacks (France kidnappings, the Binance hot wallet, hardware wallet exploits) and why covenants matter more than multisig for long-term security, the role of AI in both accelerating Kaspa development and defending against attackers, the ZK opcodes that enable both scaling and privacy, the emission schedule and the utility imperative, and what success actually looks like for a project that has the technology but is still building the ecosystem.

The episode includes a live experiment where the Kaspa community sends roughly $30 to a NanoGPT payment address on air, with mixed results and a support ticket pending, plus a hands-on look at KaChat, the encrypted chat and payments system built on Kaspa's blockDAG, where I receive the "takeover" domain name during the recording.

Sponsored by Cake Wallet, Orange Rock, Braiins Hashpower, Layer 2 Labs, and SideShift.
Follow Michael: @MichaelSuttonIL on X | GitHub

Kaspa Resources: kaspa.org | Kachat | Kaspa Dollar Initiative

Previous Kaspa Episodes: S16 E41 with Yonatan Sompolinsky | S16 E45 with Shai Wyborski

Timestamps

0:00 Intro
0:52 Welcome: Michael Sutton, Kaspa core dev & DAG Knight co-author
1:59 Michael's Bitcoin origin story (2018)
3:30 How Kaspa generalizes Bitcoin
4:14 HPC background & joining DAG Labs
5:37 First project: Arbitrum-style L2 research
7:27 DAG Knight: why not at launch?
9:02 The Rust rewrite & the path to 10 BPS
10:25 BlockDAG & K-clusters explained
12:26 The core dev challenge: 100 blocks per second
13:17 Real-time decentralization explained
15:22 What it takes to reach 100 BPS
17:15 The loyalty test: why stay with Kaspa?
20:04 What would make Michael work on Bitcoin?
23:18 Sponsors: Cake Wallet, Orange Rock & Braiins Hashpower
26:28 What is a covenant?
28:30 Vaults, inheritance & the two covenant families
29:20 Lineage & the recursive proof problem
33:29 Proving L2 state on L1
39:28 CTV, Jeremy Rubin & the Schnorr trick
41:03 Why recursive proofs become a nightmare
42:00 KIP-20: the covenant ID that simplifies everything
44:13 Silver Script & Argent: the two compilers
45:29 The year of hacks: why covenants beat multisig
51:06 Sponsors: Ecash & SideShift
52:57 AI in Kaspa development & research
56:07 The eager student vs. the responsible adult
59:25 Zcash's inflation bug parallel
1:02:07 Will AI replace human coders?
1:06:45 Kimi K3 & the open source verification loop
1:09:35 What should the Kaspa community build?
1:10:34 Chess as a covenant: the design challenge
1:11:44 The multiplex design & never scanning for checks
1:18:40 Blitz chess & ping-pong on-chain
1:20:49 Order book DEX & prediction markets
1:22:27 The RPG voting game: a crazy idea
1:29:47 "Michael always talks about chess": the criticism
1:31:15 KUSDT bridge & the Kaspa Dollar initiative
1:33:10 Bitcoin's empty mempool & the payments problem
1:36:02 Live experiment: sending Kaspa to NanoGPT
1:38:04 Transaction failed: debugging on air
1:45:02 Four UTXOs & the $10 second attempt
1:51:51 Kaspa.org, wallets & hardware support
1:58:17 The Moonsettler debate: network effects vs. features
2:00:27 Saylor's "covenants add complexity" & the OP_CAT fear
2:06:20 Kachat: chat & payments on the blockDAG
2:09:10 CPU mining days, Ice River & ASIC availability
2:10:32 RTD: miners as oracles
2:11:41 Running a Kaspa node: hardware requirements
2:14:51 Why NanoGPT rejected the deposit (probably)
2:16:21 The Zcash lesson: product-market fit takes a decade
2:18:47 Can Kaspa have shielded pools?
2:20:29 The window for privacy is closing
2:23:30 DAG Knight: testnet by end of year?
2:25:15 BlockDAG is the problem, DAG Knight is the solution
2:26:39 No delay bound: speed & resilience
2:28:52 Iran cuts the cables: the partition scenario
2:30:50 Downloading Kachat live & getting "takeover"
2:36:43 The emission schedule question
2:39:11 Can Kaspa do smart contracts like Solana?
2:41:36 The definition of success
2:45:50 Replacing Visa & the ultimate sequencer vision
2:49:28 "You want to replace Visa but can't replace Litecoin"
2:52:41 The core retreat & Coder of Stuff
2:54:14 Where to follow Michael
2:55:34 The Supra guy accepts Kaspa
2:58:53 Outro

More episodes from "Bitcoin Takeover Podcast"