Self-custody is understandably a big focus for crypto. While self-custody of assets has existed for millennia, public blockchains enable a digital form of self-custody. Significant progress has been made in improving self-custody, and continues to be made, but a vast majority of users have opted for custodial solutions. Choice is always great, and people choosing reputable CEXs need not be bullied - instead, we should offer better custodial solutions.

Today, there are 2 dominant usecases for crypto:

1) Alternative store-of-value and/or speculation: There are hundreds of millions of users worldwide that trade or hold crypto. However, a vast majority of these users do so on centralized exchanges or related custodial solutions. Larger players opt for institutional custody solutions. Coinbase and Binance alone have over 100 million users each.

2) Stablecoins: Likewise, most consumer usecases of stablecoins have chosen centralized stablecoins like USDT and USDC, over CEXs like Binance, with a lot of it activity happening on Tron. Ethereum is preferred for larger transfers, but a lot of consumers - particularly in countries like Argentina and Turkey - have anecdotally overwhelmingly chosen USDT on Tron, mostly via Binance.

Custodial solutions too have improved by leaps and bounds, but there remains significant scope for improvement, including becoming semi-custodial.

But first, some reasons why most people will not opt for self-custody:

1. Key management is brutal. Even with friendlier solutions emerging over the years, and more to come, a single accident or mistake means you can lose everything.

2. No recourse for crime. Likewise, the space is rampant with scams, phishing and malware. A single momentary lapse means you may lose everything, with no recourse.

3. It’s just too complicated. There are many wallets, many applications, many chains, many block explorers - it’s just overwhelming.

Of course, there’ll always be a niche who will opt for self-custody and accept responsibility if things go south - this will not and should not be the norm. Most normies just want good services, and they are happy to pay a service provider who is proficient at it, with years of experience and a strong reputation.

So, here are some suggestions for better custodial solutions:

Usecase-based UX: In many cases, people just want to use crypto to do something. Let’s say you just want to use crypto to hold or exchange stablecoins. The correct UX should be, for example - I log in to Coinbase.com or the Coinbase app, convert USD from my bank account to USDC, select the sendee or merchant from my address book/QR code/URL, and that’s it. In the background, Coinbase can mint USDC on Base (or an instance of), do the transaction, and if the user wants, they can verify onchain. All fees and UX hurdles are abstracted away. There are many solutions that can go beyond trustmebro, the above is just an example of how you can be semi-custodial.

Another usecase could be the user wants to earn yield. Coinbase can have active strategies, which are transparent, and the user simply deposits their funds into a selected pool. There can also be “recommended pool” so it’s even simpler. The user can once again verify if they want to, but the baseline UX is seamless.

Applications-based UX: Moving up to more advanced users, we now have people who want to select what applications they use, but don’t want to go through the hassles of vetting applications, avoiding scams, managing keys etc. For these users, semi-custodial solutions may also be useful. The paradigm that people understand well is the app store model.

Coinbase can audit and curate applications that have a high probability of being safe, and offer a risk analysis in simple terms. Personally, I use Gnosis Safe and I like the Applications UX, though it could be a lot better. Of course, users can also opt to use unverified applications at their own risk. (i.e. sideloading)

But a simple app store where users can use applications on one app or website can go a long way to making the UX seamless. What chains are being used, what fees are needed can be abstracted away from the user. Of course, this will require close collaboration from the service provider and the application developers. An ideal solution would be to have an open standard that all app stores can implement and all applications can develop for, but coordination will be difficult.

My first post on Mirror in October 2021 was kinda about this topic. (One thing I’ve changed my mind on is I now believe many of the “tech-savvy normie users” will opt for semi-custodial solutions like the ones described above, instead of smart contract wallets. That’ll be reserved for enthusiast users.) Significant progress has been made since then, yet I feel we’re lagging behind in achieving this vision by 2025.

With the advent of many CEXs branching out into L2s, there’s a significant opportunity to get the best of both worlds - onchain verification with seamless UX.

Self-custody is understandably a big focus for crypto. While self-custody of assets has existed for millennia, public blockchains enable a digital form of self-custody. Significant progress has been made in improving self-custody, and continues to be made, but a vast majority of users have opted for custodial solutions. Choice is always great, and people choosing reputable CEXs need not be bullied - instead, we should offer better custodial solutions.

Today, there are 2 dominant usecases for crypto:

1) Alternative store-of-value and/or speculation: There are hundreds of millions of users worldwide that trade or hold crypto. However, a vast majority of these users do so on centralized exchanges or related custodial solutions. Larger players opt for institutional custody solutions. Coinbase and Binance alone have over 100 million users each.

2) Stablecoins: Likewise, most consumer usecases of stablecoins have chosen centralized stablecoins like USDT and USDC, over CEXs like Binance, with a lot of it activity happening on Tron. Ethereum is preferred for larger transfers, but a lot of consumers - particularly in countries like Argentina and Turkey - have anecdotally overwhelmingly chosen USDT on Tron, mostly via Binance.

Custodial solutions too have improved by leaps and bounds, but there remains significant scope for improvement, including becoming semi-custodial.

But first, some reasons why most people will not opt for self-custody:

1. Key management is brutal. Even with friendlier solutions emerging over the years, and more to come, a single accident or mistake means you can lose everything.

2. No recourse for crime. Likewise, the space is rampant with scams, phishing and malware. A single momentary lapse means you may lose everything, with no recourse.

3. It’s just too complicated. There are many wallets, many applications, many chains, many block explorers - it’s just overwhelming.

Of course, there’ll always be a niche who will opt for self-custody and accept responsibility if things go south - this will not and should not be the norm. Most normies just want good services, and they are happy to pay a service provider who is proficient at it, with years of experience and a strong reputation.

So, here are some suggestions for better custodial solutions:

Usecase-based UX: In many cases, people just want to use crypto to do something. Let’s say you just want to use crypto to hold or exchange stablecoins. The correct UX should be, for example - I log in to Coinbase.com or the Coinbase app, convert USD from my bank account to USDC, select the sendee or merchant from my address book/QR code/URL, and that’s it. In the background, Coinbase can mint USDC on Base (or an instance of), do the transaction, and if the user wants, they can verify onchain. All fees and UX hurdles are abstracted away. There are many solutions that can go beyond trustmebro, the above is just an example of how you can be semi-custodial.

Another usecase could be the user wants to earn yield. Coinbase can have active strategies, which are transparent, and the user simply deposits their funds into a selected pool. There can also be “recommended pool” so it’s even simpler. The user can once again verify if they want to, but the baseline UX is seamless.

Applications-based UX: Moving up to more advanced users, we now have people who want to select what applications they use, but don’t want to go through the hassles of vetting applications, avoiding scams, managing keys etc. For these users, semi-custodial solutions may also be useful. The paradigm that people understand well is the app store model.

Coinbase can audit and curate applications that have a high probability of being safe, and offer a risk analysis in simple terms. Personally, I use Gnosis Safe and I like the Applications UX, though it could be a lot better. Of course, users can also opt to use unverified applications at their own risk. (i.e. sideloading)

But a simple app store where users can use applications on one app or website can go a long way to making the UX seamless. What chains are being used, what fees are needed can be abstracted away from the user. Of course, this will require close collaboration from the service provider and the application developers. An ideal solution would be to have an open standard that all app stores can implement and all applications can develop for, but coordination will be difficult.

My first post on Mirror in October 2021 was kinda about this topic. (One thing I’ve changed my mind on is I now believe many of the “tech-savvy normie users” will opt for semi-custodial solutions like the ones described above, instead of smart contract wallets. That’ll be reserved for enthusiast users.) Significant progress has been made since then, yet I feel we’re lagging behind in achieving this vision by 2025.

With the advent of many CEXs branching out into L2s, there’s a significant opportunity to get the best of both worlds - onchain verification with seamless UX.