Hey there,
It has been quiet here for a while. My first child arrived, and around the same time the team took over the building, working from the user stories I wrote before I stepped away. Last time I wrote about making those stories plain enough to need no explaining. They seem to have been plain enough, because for the first time in a long while I have had space that nobody needed me to fill.
A Familiar Kind of Quiet
I have been through this kind of quiet before. Something shifts in my life, the writing stops for a few weeks, and when it comes back it is pointed somewhere slightly different. So I was not worried when the weeks went by without an entry. I knew roughly what it was.
What I did miss was the journaling itself. Writing here every week is where I find out what I actually think, and without it the thoughts piled up with nowhere to land. That made me a little sad, and it made the quiet louder.
Becoming a dad is the obvious change, and it is the one I will mostly keep for myself. But it was not the only thing moving. Sitting on the sideline of my own company, at home and often alone with my thoughts, I could feel a second change underneath the first. It had to do with what I want to write about, and for whom.
Two Episodes and a Gap
It surfaced while I was listening to podcasts. Two of my favourites are All-In and My First Million, and they are a large part of how I keep up with technology and business. Lately AI runs through nearly every episode: the models, the race between labs and between nations, the companies being built on top, the jobs that may disappear.
Twice I noticed the same absence. In the episodes I was listening to, nobody connected the AI agents they were excited about with the blockchain infrastructure being built at the same time.
In the spring I finally wrote here about blockchain, after nearly a decade of keeping it out of my work. That entry was about letting go of the original vision and following the current towards the institutions instead. What it did not do was look closely at what happens when the agents arrive on those rails. That is the half the podcasts left out, and the second time I noticed it, it stayed with me.
What Moved While I Was Quiet
When I wrote that spring entry, Davos had just spent a week on the tokenisation of assets. It still sounded like talk. In the months since, a good part of it has turned into plumbing.
In May, DTCC, the depository at the centre of American securities settlement, announced that assets tokenised by its depository are expected to be available on Stellar, a public blockchain, in the first half of 2027. The first candidates are the most ordinary assets there are: Russell 1000 shares, index funds, US Treasuries. In September, The Clearing House, which runs the CHIPS and RTP payment networks for America's largest banks, chose Quant to build the interoperability and orchestration layer for its network of tokenised bank deposits, also aimed at the first half of 2027.
The regulators have been moving underneath. Last December CFTC staff set out how tokenised Treasuries, corporate bonds and money market funds can serve as collateral for futures and swaps. And in September the Bank for International Settlements published a working paper with a prototype on the XRP Ledger. It records a fingerprint of a published dataset on the ledger, together with the publisher's key, so anyone can check that a statistic really came from the institution that released it and has not been changed since.
None of this is a price story. It is settlement, collateral and proof of origin, the dull parts, which is usually where something has become real.
The Missing Half
Agents are starting to act for people. Not only answering questions, but booking, buying, signing and sending. Every one of those actions needs authority behind it. Someone allowed this agent to do this thing, within these limits, and someone can take that permission away again. Today most of that authority lives in API keys, which work like a door key: whoever holds one is treated as the owner. This year's breaches have shown again and again what happens when those keys walk out of the building.
The rails above offer something different, and I think this is where the two halves fit together. Tokenised money can carry its own conditions, settling only when what was agreed has actually happened. The Clearing House describes its network in exactly those terms. A ledger can bind an action to a key, and a key to a known party, the way the BIS prototype binds a dataset to its publisher. Put an agent on top of that and its permission no longer has to live in a settings page. It can sit in the money it is allowed to move, with a record of who granted it, when, and for how much, that nobody can quietly rewrite afterwards.
That is the same question we keep asking at Askara Solutions, one layer down: who owns what, who said yes, and can we show it quickly when it matters. We built our agents around that question for compliance. I suspect every agent being built now will run into it the moment it touches money.
What a ledger cannot do is carry the reason. It can prove that a payment was signed. It cannot tell you whether it should have been. That part stays with the person who gave the agent its permission, and I do not think any amount of clever infrastructure will move it.
I should be plain about where I stand. I hold XRP, XLM and QNT, the tokens behind the three networks above, along with Algorand, Hedera and a few others I think belong in the same picture. I expect to benefit if these shifts in the financial system play out the way I think they will. That is part of why I find them so exciting, and part of why I want to write about them, so read what I say here knowing it. None of it is investment advice.
Picking Up the Instrument
A while ago I wrote here that AI is an instrument. It has no intent of its own, and what it does depends on the consciousness of whoever is holding it. I meant it then as a reflection on our own work. Over these quiet weeks it turned into something closer to a frame for everything I want to write next.
So the journal is going to change a little. The building story stays, because it is the proof that any of this is real. Around it I want to look at what the AI frontier produces each week, the useful and the dangerous, treat each development as an instrument that can be turned either way, and keep coming back to what makes us human. And I want to keep filling in the half the big podcasts leave out: what happens when agents start to run on rails that remember.
I am writing this at home, in one of the quiet hours these weeks have been full of. Both changes are only weeks old. I do not know yet where the second one leads, only that it has started to ask for words again.
With care, Ben



