S15 E48: Paul Puey on Edge Wallet & Privacy Layers

Paul Puey has been building self-custody software since 2013, back when his wallet was called Airbitz and "encrypted by default" sounded like a marketing slogan rather than a design principle. The last time I sat down with him was for a Bitcoin Magazine video interview in 2019, and it produced the cleanest breakdown of cryptocurrency privacy I had heard in years of doing interviews. So when I bumped into him in Nashville a week before this livestream (the same trip where I picked up a sweet Free Ross T-shirt), I asked him to come back on the show and do it again.

He did. He also did something I did not expect: he made one of the most persuasive cases I have heard for the idea that Bitcoin's base layer could, over several decades, simply fade away while bitcoin lives on somewhere else. Paul is the co-founder and CEO of Edge, a mobile, multi-asset, privacy-focused wallet. He is not a Bitcoin maximalist, and he says so freely. "I'm not a Bitcoin Maxi," he told me, "but I'm a self-custody maximalist."

What follows is two hours of wallet architecture, Monero view keys, Thorchain, miner economics, MimbleWimble and Lightning routing, compressed into a single read.

A Self-Custody Exchange, Not a Wallet

The first surprise came when I mentioned the swap function I had discovered while testing the app that morning. Paul corrected my framing right away. "The goal of Edge is actually to be less a wallet," he said. "Really our goal is to be a self-custody exchange and to compete with the centralized exchange services."

Since we last spoke, Edge has integrated somewhere between 12 and 16 swap providers ("gosh, I've lost count"), including more than half a dozen decentralized exchanges such as THORChain and aggregators and bridges like LI.FI and Rango. Next on the radar were Maya Protocol, which he expected to bring Zcash swaps, and Serai, whose founder he had talked to about decentralized Monero trading.

The logic is simple. Today a user has to get a wallet, secure the keys, open an exchange account, buy the coins, then do a WalletConnect dance just to reach a DeFi protocol. Edge wants to fold all of that into one app, the way Coinbase does, "but of course we fundamentally feel like we need to do that in such a way that people still hold their own keys, have their own funds and have a high level of privacy."

Building Privacy Tech in America After Nashville

I asked what it feels like to run a privacy-first wallet from the United States, given how rough the last couple of years had been for privacy projects. Paul pushed back on the word I used. Edge, in his view, is not helping anyone circumvent KYC. "In the self-custody space you simply, that requirement isn't there," he said, describing crypto as "one gigantic big gray area" where you build what people want first and work with regulators later.

His privacy philosophy has two halves: hide what is sensitive, and better yet, never ask for it. "If you don't require it in the first place, then you've significantly lowered the odds of that sensitive information leaking out into the world." And this is not only about privacy. When your holdings are tied to your name and address, "you effectively become a target," so for Paul, "privacy equals security."

We were recording a week after Bitcoin 2024, and I admitted my worry that the industry was betting everything on one candidate. Paul thought the speech set a lower bar than people believed ("leave us alone and Gary Gensler out") and that the real shift was the groundswell of pro-crypto representatives in Congress. His own preference at the time was RFK Jr., whom he considered more pro-Bitcoin at heart. My concern was about the other kind of bar: what politicians would now expect this industry to donate in exchange for being left alone.

Open Source, Contact Lists and the Permissions Panic

Someone on Twitter wanted to know if Edge is still open source. It is. The code lives on GitHub under EdgeApp, and Paul was upfront that it is not a trivial build, because "it has many different code bases that need to come together with a very specific setup." People have built it anyway, and some have launched their own projects on top of it.

He made a point I think more wallet users should hear. Open-sourcing only the key management library is not enough, because "the interface could easily compromise that." Everything between what you see on the screen and the moment the app hits the network needs to be auditable.

That matters most when privacy-minded users panic over permissions. Why does Edge ask for the camera, photos and contacts? Camera and photos are for scanning QR codes. Contacts let you tag a transaction with a friend's name, and that tag is encrypted on your phone. "That data in clear, unencrypted, is never sent to Edge," Paul said. His bigger lesson: "what permissions an application asks for are not necessarily permissions that the company is asking for." The only way to prove it is to publish the code.

Xpubs, Blockbook and the Monero View Key

Light wallets need someone else's server to learn about the blockchain, so I asked what public key data Edge sees. Paul's answer was precise. "Edge the company doesn't collect any public key data," though some can end up in their hands depending on what you do. For balances and transaction history, Edge queries open infrastructure. It moved from Electrum servers to Blockbook servers, the open-source indexer built by Trezor, run by various people and companies. Because those endpoints are open, a server operator "might see that this IP address went and queried for a balance," but cannot tell it came from Edge.

The bigger privacy choice is what Edge does not send. Many wallets ship your xpub to their backend, which is efficient, but "the server that you send it to can derive every address you will ever have till the end of time." Edge queries individual addresses across dozens of chains, which is more costly to engineer. "Is it perfect? No," Paul admitted, "but it does contribute to a higher level of privacy."

Monero is trickier. Because transactions are encrypted, a wallet has to scan the whole chain to find its own outputs, which burns battery and bandwidth on a phone. Edge uses the approach pioneered by MyMonero, the company of Riccardo Spagni (fluffypony): hand a view key to a server and let it tell you which transactions are yours.

The server learns your transaction IDs and incoming amounts, but not who sent them, where your outgoing payments went, or your true balance. Paul estimates that keeps "about half to two-thirds of Monero's privacy." If you want all of it, you can point Edge at your own light wallet server, and he mentioned a small box with a full node and LWS being developed by Doug Tuman, the organizer of MoneroTopia.

What really excites him is a future version of view tags, which would let a server know a transaction is yours and nothing else. "To me, that's the holy grail," he said: "90 to 99% of the privacy but with 100% of the usability."

Why Your Full Node Can't Talk to Your Phone

People often ask me why they cannot simply point a mobile wallet at the Bitcoin Core node running at home. Paul gave the best plain-language answer I have heard. A full node stores blocks and verifies them, but "a full node has a very poor language by which you can ask it for information." It can hand you block 500,000. A light wallet does not want blocks. It wants transactions and unspent outputs for specific addresses.

An indexer like Electrum or Blockbook "takes the entire blockchain and then it puts it into a database," so you can ask friendlier questions: give me the balance for this address, give me only the transactions after block 500,000, give me 50 at a time so a low-bandwidth connection in the middle of Africa does not time out.

For a single-coin Bitcoiner, a node-connected wallet can still be workable. For an app juggling Bitcoin, Litecoin, Dogecoin, BCH and more, it would be too heavy. And Paul sees that diversity as a feature. "I challenge anyone to tell me what's the drastic difference between their checking account with Wells Fargo and Chase." Software wallets, by contrast, can be radically different from each other.

Thorchain, Sidechains and the Rootstock Era

Since this show was sponsored by StarkWare and Layer Two Labs, I asked whether sidechains still have a future now that ZK rollups get all the attention. Paul thinks sidechains were never that bad, and that the real objection back in the Rootstock days was the small federation controlling the peg.

His example was THORChain. "Name anyone that's called Thorchain a sidechain. No one has," he said. But technically, instead of a federation of around 15 signers, he described 40 to 80 nodes that rotate keys weekly to lock bitcoin and mint a synthetic version, for half a dozen chains at once. Rollups are "a more trustless way to do it," he conceded, but this model is cheaper and "pretty darn good."

I added my usual line that almost everything built on other chains today was first proposed for Bitcoin, and I made the case for Paul Sztorc's drivechains: not just a block size increase without a hard fork, but sidechains for every use case that pay fees back to miners. Peter Todd and others think it hands miners too much political power. I think the security budget question makes it worth discussing anyway.

The Bitaxe Argument and the Security Budget

Paul agreed the security budget is the conversation nobody wants to have, especially "the Uber Maxis that don't want sidechains, they don't want rollups, they only think Lightning." His math was blunt: "If Bitcoin doesn't at least double every four years, then we have effectively a net downward security budget."

I told him about a Twitter exchange where someone argued we could let the big miners go bankrupt and secure Bitcoin with home devices like the Bitaxe and the NerdMiner. Paul took it seriously and dismantled it. If miners with cheap ASICs and flared gas cannot turn a profit, why would a hobbyist paying full retail electricity? "So what would make me think that I'm going to be more profitable than the other miners?"

When I raised ideological miners who treat it like a lottery ticket, he called that a tiny share of hash rate and pointed to the original elegance of Satoshi's design: it relies on rational selfishness. "We've inverted that if we're relying on ideological miners that mine at a loss."

On tail emission, the idea Peter Todd champions, Paul has no objection in principle. He just does not think it will ever happen, "because Bitcoin's a religion and that's breaking their religion."

Why the Layer 1 Might Quietly Die

Then he laid out the scenario that became the heart of the episode. Imagine an L2, a rollup or something like a giant threshold multisig, where bitcoin is pegged in and all the economic activity happens. Ten years out, that L2 collects all the fees. "Once you've pegged your Bitcoin into an L2 and that's where all the economic activity is, why would you ever peg it back to the L1?" The base layer keeps minting fewer and fewer coins, and maybe a century from now, "the L1 effectively just dies, and that's okay because Bitcoin still continues on."

I told him he had just made the best case for drivechains I had ever heard. He then explained why the L1 could not easily fight back. Threshold signatures operate at the cryptography level, so 80 cosigners can look like "a single key transaction" to every node and miner. "What are you going to do, L1, to try to prevent an L2? Tell me." His closing question for maximalists: 50 or 70 years out, if the L2 holds the whole security budget, what is the long-term need for the L1?

He also noted that there is more wrapped bitcoin on Ethereum than BTC on Lightning, which in a way makes Ethereum a layer two for Bitcoin already. And Lightning, in his view, keeps dragging you back to the base chain to open channels and rebalance.

Why Privacy Coins Missed the Bull Market

About an hour in, I finally steered us to privacy, and to why Monero and Zcash looked so bad price-wise during the last cycle. Paul blamed the narratives. 2017 still carried the "spend Bitcoin, use Bitcoin" story. 2021 was DeFi yield, then NFTs: "here's my .eth address and look how many NFTs I have," which he called "the polar opposite of privacy." And this cycle was driven by the ETFs.

He is convinced privacy demand returns with payments. At PorcFest, he said, "everybody there was paying with Monero," for hamburgers and silver, nothing illicit. "Once we get into the transactional use case again, I think the need for privacy and the desire for privacy" will push those markets up.

A project called Zano asked on Twitter whether Edge would add them. Paul said Edge already supports more privacy coins than any app he knows of (Monero, Zcash and Pirate Chain) and that each took "an insane amount of engineering effort." Some chains they evaluated "didn't live up to the marketing." Before any listing, Edge wants SDKs and node tooling, plus swap liquidity, so users can actually acquire and trade the asset, not just hold it.

The Four Layers of Privacy

This was the part I invited him back for. I complained that people on Twitter tell me CoinJoin is "just like Monero," when amounts stay visible and consolidating mixed coins can undo the whole thing. Paul broke it down into four layers, as he did five years earlier: privacy for the sender, privacy for the recipient, privacy for the amounts, and network-level privacy.

CoinJoin gives sender privacy and nothing more. Stealth addresses give recipient privacy. Paul remembered his team looking at them years ago while working with Amir Taaki of Dark Wallet. Confidential transactions, which he credited to Greg Maxwell, shield amounts. "These are the three key technologies that are used in all of the true privacy coins," he said. "If it's missing any one of those, you have a major gaping hole." The fourth layer is IP privacy through VPNs and Tor, which hides from your ISP that you are even touching a privacy coin.

His harshest words were for silent payments (he called them "secret payments" on air, then admitted "the nomenclature skips me"). In his view they are stealth addresses rehashed, giving roughly one-third of the privacy while carrying "100% of the user experience challenge" of private coins, because wallets must again scan the chain to find their transactions. "It's like all of the detriment in user experience with only a third of the benefit." I made the same broader argument about partial solutions in my essay On The Limitations of Bitcoin Privacy.

MWEB: From Chatty to Non-Interactive

I asked him about Litecoin's MimbleWimble Extension Block, which I covered at length with Charlie Lee. Paul had written MimbleWimble off long ago because it was "very chatty": sender and recipient had to exchange data off-chain before a payment. "I literally remember the developer saying, oh it's simple, you just email this file. Are you kidding me?"

Litecoin changed his mind on that front. "They've made it non-interactive," he said, so you can publish an address and receive while offline. "Now it becomes interesting." He then claimed ignorance on how its privacy compares with Zcash's, "because I'm stepping outside of my pay grade."

His real objection is to the soft fork route. If you have to peg into the extension block and peg out to transact on the main chain, "it's just another L2," and soft forks are "a giant political war" plus "insanely complex code." Zcash, by comparison, lets transparent and shielded funds mix in the same transaction. If Bitcoin ever gets two successful L2s, Paul imagines one could be a Zcash-style privacy layer and the other an EVM smart contract layer.

Lightning, Custody and the Anonymity Set

Some people pitch Lightning plus Chaumian ecash on custodial nodes as Bitcoin's privacy and scaling fix. Paul was having none of it. "Once it's a custodial solution, it's dead. It's dead to me." He called it "a cop out to the narrative that was presented to us years ago."

He credits Lightning with privacy "way well above Bitcoin," but the network's incentives push it toward big, liquid hubs. With one hop between us, that hub knows both IP addresses and who paid whom. More hops mean more privacy, but in a five-hop route, "five additional Bitcoin have to be locked on the network for my one Bitcoin to successfully get over to you." Capital efficiency and privacy pull in opposite directions.

So no Lightning in Edge. His team tested the Breez SDK, which he praised, and still "quickly found routing issues very easily at very low amounts." For a self-custody exchange, a roughly $2,000 practical limit simply does not fit. And multi-asset wallets that do support Lightning usually do it through a custodian: "We're not going down that path ever."

Seed Phrases, Signal and the People Who Don't Care

When I installed Edge, it never asked me to write down 12 words. Instead you create a PIN and an account. Paul explained that it "feels like an account but it's fully client-side encrypted," backed up and protected by two-factor, with a real private key underneath. Edge is not competing with other wallets, he said, but with "the Krakens, Coinbases, the Geminis." Fiat partners tell Edge that, compared with other integrations, its users skew heavily toward newcomers and older people.

He is open about who this alienates: "the 24-word Maxis" who want words in a safe. Hardware wallet signing is not on the roadmap for now. Paul prefers threshold signatures and social recovery across your own devices, because he worries people underestimate the risk of losing or exposing a written seed.

The reason he cares about users who do not care is the anonymity set. If only the two of us used self-custody, "we would just be the two people in the woods with our private keys written on a piece of paper." He has onboarded friends to Signal who do not care about privacy at all, and now their travel chats give cover to those who do. "Building tools even for the people that don't have that ideology is vastly more important than people realize."

Controversy, Namecoin and a Self-Custody Maximalist

I asked for his most controversial opinion about Bitcoin. He had trouble picking. Raise the block size, give up on Lightning in favor of L2s, the ETF sucks, and of course: "I think the L1 will die in several decades," and "it would be a net good for Bitcoin." He laughed that it is hard to be controversial "if you're not hard set on the religion of Bitcoin."

I tried to call him a maximalist anyway, since he helps Bitcoin scale and probably ends up back in BTC. He refused the label. Most of his value sits in bitcoin today, but he holds and promotes other coins and makes no promise it stays that way. I reminded him that Satoshi was not a maximalist by today's standards either, having contributed to Namecoin. "So I guess I'm closer to Satoshi in that regard," he said, then caught himself: "that's pretty hefty bragging."

Closing Thoughts

One small correction came at the end: Edge is mobile-only. My desktop copy was the iPhone app running on Apple Silicon, which Paul uses daily and which Apple translated surprisingly well. A native desktop app is on the horizon, but as he put it, "the present is mobile."

You can download Edge and find Paul and his co-founder at edge.app, with a direct Android APK for phones without the Play Store. Because Edge does not track its users, he asked everyone for feedback, including me: "we have no other way of knowing." If you want more on this topic, revisit my Charlie Lee MWEB conversation, and for the drivechain side of the miner fee debate, my talk with Paul Sztorc on scaling.

Listen to Paul Puey on Spotify, Apple Podcasts, YouTube, and More

Watch on YouTube

Time stamps:

00:00:47 – Introducing Paul Puey
00:02:08 – Overview of Edge Wallet Features
00:04:00 – Decentralized Exchange Integration
00:04:58 – Challenges of Privacy in the U.S.
00:06:05 – Privacy and User Security
00:08:05 – Political Climate and Crypto
00:10:01 – Expectations from Politicians
00:12:33 – Open Source Status of Edge Wallet
00:14:31 – Importance of Transparency
00:15:56 – User Privacy and App Permissions
00:18:51 – Edge Wallet's Architecture
00:19:19 – Public Key Data Handling
00:19:58 – Transition to BlockBook Servers
00:20:59 – Privacy Enhancements in Edge Wallet
00:21:59 – Understanding Xpub and Privacy Trade-offs
00:23:12 – Monero's View Key Explained
00:25:17 – Running a Private Monero Node
00:26:19 – Future of View Tags in Monero
00:28:04 – Sideshift ad
00:28:26 – Comparing Electrum and Full Nodes
00:30:24 – Blockchain Data Management
00:32:25 – Querying Blockchain Efficiently
00:34:38 – Challenges of Running a Full Node
00:35:33 – The Usability of Multi-Chain Apps
00:36:38 – Comparing Banking Systems to Crypto
00:37:20 – Sponsorship Message from StarkWare
00:38:09 – Excitement for L2 Developments
00:38:25 – The Future of Sidechains vs. zk Rollups
00:38:50 – Improvements in Sidechain Technology
00:40:03 – The Potential of Sidechains for Multiple Cryptos
00:41:17 – Challenges for Bitcoin's Functionality
00:41:45 – Market Demand for Bitcoin Innovations
00:42:22 – Debate on Sidechains and Miners' Political Power
00:43:16 – Regulatory Discussions on Miners' Compensation
00:44:44 – The Viability of Small-Scale Mining
00:45:53 – The Reality of Mining Profitability
00:48:19 – Ideological Mining vs. Profit-Driven Mining
00:49:41 – The Concept of Tail Emission in Bitcoin
00:50:21 – Challenges of Implementing Tail Emission
00:51:46 – Funding L2 Validators through Transaction Fees
00:52:53 – The Future of Bitcoin's Layer 1 and Layer 2
00:54:30 – Discussion on Layer 1 and Layer 2
00:55:05 – Multisig and Threshold Signatures
00:56:52 – Long-term Viability of Layer 1
00:57:52 – Future of Bitcoin
00:58:30 – Ethereum as Layer 2 for Bitcoin
00:59:24 – Challenges of Lightning Network
01:00:59 – Hybrid Proof Systems
01:03:05 – Transition to Privacy Discussion
01:03:35 – Market Trends Affecting Privacy Coins
01:07:00 – Privacy Coins and Transactional Use
01:08:01 – Integration of New Privacy Coins
01:10:01 – Lengthy Response on Listing Coins
01:11:00 – Clarifying Privacy Expectations
01:12:21 – Privacy Layers Overview
01:13:28 – Stealth Addresses Explained
01:14:35 – Confidential Transactions
01:15:45 – Challenges of Privacy Coins
01:18:59 – IP Address Privacy
01:19:06 – Bitcoin's Privacy Potential
01:20:17 – Zcash and Scalability
01:21:07 – Litecoin's MimbleWimble Extension Block
01:22:38 – User Experience and Protocol Complexity
01:25:30 – Soft Fork vs. Sidechain
01:28:22 – Lightning Network's Privacy Claims
01:29:43 – Anonymity Set in Lightning
01:30:39 – Understanding Centralized Entities and Anonymization
01:31:51 – Challenges of Multiple Hops in Lightning
01:32:37 – Integration of Lightning Network into Edge Wallet
01:34:22 – Self-Custody Focus of Edge Wallet
01:34:49 – Bringing Self-Custody to the Masses
01:36:13 – Importance of Privacy Tools
01:37:09 – Controversial Opinions on Bitcoin
01:40:47 – Skepticism Towards Bitcoin's Future
01:43:16 – Market for Non-Bitcoin Functions
01:44:33 – Hardware Wallet Integration in Edge Wallet
01:46:29 – Onboarding Users Without Seed Phrases
01:47:03 – User Experience and Security in Edge Wallet