Christoph Ono

And turns out this is 99% bots. Actual people usage is minimal. Fun stuff.

Nostr is a rebellious protocol with the most conventional interfaces.

Had Claude review its code and coding process from the last year of Arké. From this post, it seems like AIs are still basically in a Memento situation where they are only coherent for a short period of time and rely on notes they left for themselves to navigate unfamiliar situations. https://gbks.substack.com/p/a-year-of-code

Nosta.me traffic over the last year. I think at least one of the spikes was bot activity, which lead me to restrict things more. A month ago or so I did an AI rewrite to bring the code up to par, but otherwise it's been hands off (no time or mental space). I should really dig a bit further into these numbers to better understand usage.

Hackathon project adding an MCP server to Arké, so Claude can request payments. That was really fun to put together. https://x.com/matthewvuk2/status/2106416882219655229?s=20

Claude talking about its own work in Arké:

"Coding is fast; correct coding is not. An AI-assisted session produces a first draft with tests in hours, but a third of past sessions went into revisiting generated code for faults no single file reveals: device-scoped actions on account-scoped state, read paths fixed without write paths, task ownership, wrong assumptions about what bark does, and claims in docs taken on trust."

Now let's see how to improve that process...

Such good work by the Umbrel gang.

“It looks close to the design.”

Will fiat values make it into Arké very soon? Potentially. Always with the option to hide them.

Those bloodhounds also have a hardware wallet.

How about this Bitcoin app in Bloodhounds (s2e6)?

Just checked what the fee was to send 1 USDT on Tron, and it was $3.474. Also learned that you need to stake $11,000 worth of Tron to get a free stablecoin transaction per day (staking gets you energy that is then used to pay fees). And that there are third-party marketplaces for energy to cut transaction costs to $0.50-0.80. Hope I understood this correctly, but sounds pretty messy, TBH. I find it very enlightening sometimes to dig around other ecosystems to investigate claims we hear about them.

So interesting how Penpot explored this floating panel concept in 2023 and ditched it, the Figma had it for a short period in 2024 and ditched it, and a couple of years later Apple introduced it with MacOS 26 and ditched it again with MacOS 27. Seems like there are some alluring ideas everyone just has to try.

Just had Claude fix up nosta.me, after not having worked on it for a long time. Just amazing how fast it it could find and fix tons of issues, bring the site up-to-par with the latest NIPs, and even implement some new ones (like highlights and payment addresses). Still some things to do, but it works much better now.

https://github.com/GBKS/nosta-me/commits/main/

“You’ll watch something anyway, so why not brain rot? It’s actually better than an ad because it doesn’t demand anything from you.”

“If everything is content now, then brain rot is anti-content. It’s the only way to be free online, which is to do something so stupid it can’t be monetized.”

https://www.nytimes.com/2026/09/08/opinion/gen-z-tiktok-ig-reels.html

The first few themes and localizations are live now with the latest release. Along with the ability to export/import your contacts, tags, and notes. Plus a few other goodies. The work continues.

Arké is still in beta. Please stick with small amounts for the time being. And share your feedback if you can, it helps improve the experience.

Great to see apps charging for the awesomeness they provide.

Are you a humanist artisan of the web?

"The Internet is a symbol, the catalyst of an era based on revolutionary freedom in the field of expression, creation and knowledge. These are the human ideals we are striving for. We have to observe our new horizons with increasing sensitivity in order to inspire a future of growth, serenity and sharing."

https://www.brunellocucinelli.com/en/thoughts/humanist-artisans-of-the-web.html

New blog post with some recent work on Arké and the Bitcoin Core App. https://gbks.substack.com/p/two-miles-an-hour

Banging my head against multi-device architecture via iCloud for Arké. Quite the puzzle. https://github.com/GBKS/arke/blob/main/Shared/Docs/Architecture/Multi_Device_Design.md

Mocking up some transaction scenarios to help figure out the UI states in the new core app.

Having AI whip out the first localizations to German and Japanese. Probably going to be a bit goofy until a few more iterations in.

Not sure how it is in engineering, but some design/architecture professors tear up your work in front of you during a critique if it’s not good. It’s tough, but it teaches you to not get too attached and push for better. You need to care and feel attached, but also be able to look at it from a distance and let go quickly.

New version of Arké is on TestFlight. Transaction details are now a swipeable stack of cards to quickly flip through and add notes. Plus force move (unilateral exit) refinements, security hardening and bark package updates. Check it out. Share your feedback.

Update on stuff I worked on in the last month, which was primarily unilateral exit UX for Arké. It's coming along nicely. https://gbks.substack.com/p/security

Just had to think of this mockup I made 3 years ago about escalating security reminders. A wallet starts annoying you hardcore if you store more bitcoin in it than is appropriate for the security it provides...

Oh boy….

Just came across this design I made 1.5 years go for the core app to inform users about pending network updates (all made-up content of course).

I am with Ben Shneiderman on direct manipulation.

Prototyping a way to go more quickly through transactions and add details. What do you make of this stacked-card concept?

Interesting to see the term "Design Engineer" become more used. In 2004, when I first started working for digital agencies, before the term UX existed, the field was "Interaction Design", and todays "Design Engineer" was the "Interactive Developer".

The agency I worked at had technical architects, backend developers, application developers, interactive developers, designers, and PMs and I always thought this split-up made a lot of sense. It becomes very clear when working directly with people who are really good in their areas.

In more recent years, the idea of the full-stack developer was around, assuming that one person can just do it all, which never made much sense to me.

No matter what labels you use, building good software (or anything really) requires various skillsets.

There might be more effort going into making bitcoin usable by agents than by humans.

And another Arké release is up on TestFlight:

• Many improvements to force moves (unilateral exits)
• Tag and contact editor revision
• Settings - reorder and new Notification and Address Pattern views
• X-ray iteration (remove balance, change VTXO list to graph)
• Use real fee rates
• Minor revision of the intro videos https://blossom.primal.net/cc6f0ea5b8c2ccd32d1f66abbc41561697e213d8ed1d88777c2c322672f026d0.jpg

If you like 10-minute walkthroughs of unilateral exit user experiences, then this one is for you.

Not really sure if it's a good idea to add stablecoins to bitcoin wallets. They drag in so much complexity that it may actually be more practical to have separate, lean & dedicated apps.

A nice thing about the Live Activity for unilateral exits in Arké, is that they show up on your watch, too. Left is a mock-up of the Live Activity on the phone, and right is a photo from signet exit testing over the weekend. Trying to make this a calm and convenient experience.

The X-Ray view, which allows more seasoned users to look at their VTXOs and other under-the-hood stuff is getting a bar graph visual. You can quickly see your VTXO structure and how long until refresh/expiry.

Arké defaults to mainnet these days, but creating a signet wallet is simple. Tap the test tubes on the cover screen. The background video switches to show the fancy lab guy. And creating a wallet will get you a signet one. The built-in signet faucet gets you some test sats right away. Everything else is pretty much the same as mainnet.

Adding and editing basic contact info is also getting a revision in Arké. Now back to testing and refining unilateral exits.

Adding and editing tags will be a bit nicer in the next Arké release. The old version was essentially AI slop. https://blossom.primal.net/44900757f4a92654dff5c0f86613b3b8153e51f90eb770f96e2c585585ae0a98.jpg

Interesting to see some tense conversation around Ark, Spark, Liquid, etc and different defenses of tech choices. Tough discussion, because there are so many facets. From basic architecture to protocol dynamics and then protocol and library/SDK implementation, to infrastructure operations and end-user apps. And the project background matters, too, where a heavily-funded commercial entity will have different dynamics and incentives than a lean open-source focused startup, and those show up in various decisions. Things are not clear cut, but that's how it is with many things.

Arké has had unilateral exits (called forced moves) for some time. I am currently refining the experience and tinkering with this feature video intro to better communicate how it works. What do you think?

"All design is a quarrel between two impulses: the wild idea and the cold edit. Unleash the wild, and you get baroque chaos. Lean too far toward cold, and you get vanilla or gray."

An experimental promo video for Arké about how bitcoin is a bit like a public park.

Working on the app is fun, but man do I sometimes need creative side quests when Claude rattles on and on.

A new version of Arké is now up on TestFlight. Here are the changes, with the first two being the biggest ones:

• Better fee estimation for lightning payments (including send all).
• More precise refresh notification scheduling.
• Change inline address expansion in receive to a sheet format.
• Add option to export debug logs (thank you Duncan).
• Make first use experience accessible.
• X-Ray view iteration (single reload button, more usable VTXO refresh options…).
• Dismiss wallet deletion view once deletion is complete.
• Fix tag rendering after import.
• Fix initial rendering and empty state of fee summary view.
• Fix localization issues in devices screen.

It's quite a diverse mix of things, many of them stemming from feedback and conversations. Please keep them coming, no matter how big or small.

A few Arké users have reported a balance that has been stuck in pending for a few days. Here's how you can get it unstuck (the underlying issue is already fixed and this should not happen anymore).

In your balance details view, you'll see a timer that reads something like "Next round in 12m". Keep the app active until that point in time, and maybe give it another few minutes extra. That should fix it. If it doesn't reach out.

Thanks to everyone for testing and helping make the app more solid.

Pretty sure social media bans wouldn't have come up if social media companies had taken more responsibility in protecting minors. Simple stuff like not letting kids be contacted by random adults, muting push notifications at night or during school hours, private accounts by default, removing engagement-maximizing mechanics, etc. When parents and teachers constantly fight against this stuff at home and in school, of course there's a strong counter-reaction at some point. Let's not let those companies off the hook easily.

Updated some pages on the Arké website:

https://arke.cash/roadmap/

https://arke.cash/how/

https://arke.cash/how-technical/

Better, not perfect.

Some notes from the bark and Arké mainnet launches last week, along with first feedback on the app. It's been exciting. https://gbks.substack.com/p/toot

Would be nice if bitcoin conferences did not put pseudoscience and conspiracy peddlers on the big stages. How about a dedicated BS stage?

Bark and Arké are now on mainnet. I consider Arké still in public beta. Focus now is on stability, reliability and interoperability. If you try it out, let me know what you think. https://blog.second.tech/introducing-noah-and-arke/

Completely gratuitous and no one will notice it, but QR codes in Arké now have a very small legend that indicates what addresses are in said QR codes. If you know, you know. If you don't know, doesn't bother you. Another experiment, let's feel it out.

In tougher times it’s better to rally together than to argue. Onward and upward.

How do wallets handle payment and payer proofs? Any good UX patterns out there? https://github.com/BitcoinDesign/Guide/issues/1235

“AI is best at the things you’re bad at, because you can’t recognize its mistakes.”

Just pushed a new version of Arké to TestFlight. This one is about refinement:

• More precise fee calculation, which is especially useful when trying to send the full savings (onchain) balance
• Different intro video segments for signet and mainnet
• A new cover video that is less 80's glam
• Improved wallet creation performance and error handling
• Unilateral exit progress display tweaks
• Wallet import with file improvements

Still ~150 to-dos on my list, but it is coming together.

https://arke.cash/

Second has a new page about apps that implement Bark, their Ark implementation. Super nice to see Arké in there. https://second.tech/docs/built-with-bark

And to keep going on and on about unilateral exits, Noah and bark-web support them also. As it should be.

Been seeing several conversations that it is not important for L2 wallets to implement unilateral exits. Good people, who else is going to do it? And who can actually do the unilateral exit for you when only your wallet has the data needed for it? The primary task of a wallet is to ensure your bitcoin are safe. Swapping to the 1000th stablecoin is way lower on the priority list.

OK, it appears that no Spark or Arkade Labs Ark wallet actually implements unilateral exits. Protocol-level capability or separate CLI tool is not usable for most and does not count. Pretty disappointing. So if the server goes away, your funds are gone. The feature to claim them is not implemented.

And by the way Arké has a great unilateral exit flow. The Second crew provides a really good toolbox there.

Let's maybe hold off with the stablecoins and yield and all that and make sure the basics are solid.

Has anyone tried a unilateral exit on Spark? How is it?

Bitcoin is "...a monetary technology. This is the accounting industry." https://stacker.news/items/1483355

I like it.

Looks like there are 294 different Stablecoins based on the US Dollar. The biggest one, Tether, exists on 126 different blockchain networks, 17 of which have 100M or more circulating USDT, and 60 have 1M or more. The second biggest, USDC, exists on 141 different blockchain networks, 18 of which have 100M or more circulating USDC, and 76 have 1M or more. When you want to pay with USDC in the Stripe Crypto Checkout Demo, you have to choose one of 400 different wallets. Decentralization at its finest.

Just kidding, but how is this not super messy and a world of pain in regards to interoperability and payment UX?

https://defillama.com/stablecoins?pegtype=PEGGEDUSD

https://docs.stripe.com/payments/stablecoin-payments

Got a Zombie score of 39%. What a great site. Something that's only possible on Nostr. https://www.plebsvszombies.cc/

People are not friction. Well said. https://daverupert.com/2026/03/people-are-not-friction/

Quick demo, making 100 bitcoin (signet) transactions to see how Arké responds. Getting quite smooth overall.

Bit of a restyling of the Arké payments detail screen, along with some information architecture changes of the refresh indicator.

Can every one please stop implementing Spark? The privacy problems are real and annoying.

Cool to see the results of the Runway ad contest. AI video has made such huge leaps, and the participants really know how to get creative with it. https://runwayml.com/big-ad-contest

Would be cool if Arké could make payments to local fiat destinations just by paying lightning invoices and other services like Tando handling everything in between. Some AI-brainstorming resulted in a simple protocol that could facilitate this. No clue if it's worthwhile, but was useful to explore: https://github.com/GBKS/pave-protocol

The new USDC site has some nice videos on how people use it. But I cannot get over the fact that all these chains are based on single-address re-use. If you organize your wedding in some other country and pay everyone with USDC, are you OK that they can see your full balance and all past and future transactions in that wallet? Or your employer or employees? Your grocery store? Or am I missing something? https://www.usdc.com

New Arké version is up on TestFlight, with a roughly estimated 1,000,000,000 improvements to the Send flow.

Goal is that you can throw all your bitcoin/ark addresses, lightning invoices/offers/addresses, BIP-21s and BIP-353 in all kinds of combinations at it, and it will try to make sense of it all and present whatever comes out on the other end in a simple way. Quite intricate. It's a good iteration, with some more to come.

Definitely feels like the most intricate part of the app.

"Beyond the GUI: It's Time for a Conversational User Interface" from 2013. https://www.wired.com/2013/03/conversational-user-interface

"Why conversational design is the future" from 2016.

https://medium.com/intercom-inside/why-conversational-design-is-the-future-2c05f65aa68e

Will AI make it happen this time?

Chuck Norris’s public key was his private key. Nobody dared to use it. RIP

Arké looks, feels, and works differently than other wallets. Wouldn't be much of a contribution to the ecosystem if it was like all the other ones. Some experiments will work and some will fail. But if we want more people to use this stuff, we need to try things.

Some interesting research by Anthropic on what people want from AI, and whether they are getting it. https://www.anthropic.com/features/81k-interviews

Considering writing an Arké case study for the Bitcoin Design Guide. Here's a first issue to invite conversation. https://github.com/BitcoinDesign/Guide/issues/1220

This looks great. I'll investigate Passkey Login for Arké as well.

First Arké push notification ever! The text obviously needs work, but very cool that it works now (this was an intimidating piece of code for me to work on). The bark mailbox system makes this very convenient to implement (those Second people seem to know what they are doing).

Privacy-wise, this requires a relay to connect the app, Ark server and Apple push notification servers. The relay can connect your device ID, mailbox ID, and incoming Ark payment VTXOs. Don't like it, don't turn on notifications. I'll have to see how I can only store the least amount of data possible, since I really don't care to know what people do. The only interesting thing to know maybe would be a simple counter that tracks how many payments were made. That would at least give some broad insight into activity.