agents are making financial decisions at machine speed. the signing architecture usually hasn't caught up
2PC splits the key: one share on your device, one in infrastructure. the agent can act, but neither side can move funds alone. you stay in the loop without being asked to approve every transaction
https://t.co/HOusIuuRIS
any agent wallet comes down to the same question: who signs, who owns, who can override
delegated authority is a cryptographic guarantee, not a UX decision
you don't show a Prime Minister something in your store just to show off.
you show him something that works wonders.
@CactusXR gives a physical store a brain. staff always know where things are, nothing gets lost between shifts, problems get caught before they become problems.
the stores that figure this out early won't be going back
Modernizing every part of the supply chain is a national interest.
For everyday consumers, the supply chain ends in the grocery store.
In this picture, the @SwedishPM gets a demonstration of @Auki’s AI copilot @CactusXR for retail by store owner Johan Westlund, ICA Kvantum Värnamo.
Sweden has always been a leader in innovation and technological adoption.
nice. free apartment cleaning.
but are you willing to let a stranger's camera
map every corner of your home
to train robots you'll never benefit from?
it starts with cleaning.
then repairs. then errands. then everything.
this is how the inside of our homes becomes a dataset.
@Auki is building against this -
robots that work inside your space without your camera feed ever leaving your walls.
the physical world doesn't have to be a data mine.
Today, we're launching shift. We're starting by cleaning your apartment in New York City, for free.
Here's how it works. Book a shift cleaning. A vetted shift operator comes to your home wearing one of our devices. They clean. They leave. You pay nothing.
In exchange, we record the cleaning. Robotics is being built on data about how people do daily tasks, and the value of that recording is what funds the service. Anything personal in it is anonymized before the recording is processed.
By now, you have heard about the shift to AI more times than you can count. About the shift toward you, the part where you actually feel it, you have heard almost nothing. Shift is what starts to make it concrete, in specific cities, with specific services.
Today, cleaning in New York. Soon, handymen, repairs, and errands across the globe. And this is just one side of shift, with more on the way.
Comment “shift” and we’ll send you an early access link.
if a bot can beat any single verification method, single-method systems eventually fail
that's why Passport combines government ID, biometrics, phone, onchain activity, and more. the score is the product of all signals, not just one
DePIN compounds because real infrastructure has real utility that doesn't expire.
Think about what Auki is actually building - a decentralized network that makes physical spaces readable for machines.
Every spatial domain contributed by a participant is another piece of the world that AI and robots can now understand. That accumulates. It doesn't reset.
It doesn't need a narrative cycle to keep being useful.
Most of crypto may be a casino, but DePIN is not.
DePIN is a force of nature that will compound year after year after year until it's just as obvious as tap water.
A robot rolling the aisles, live-streaming what it sees, so head office can check stock levels and shelf layouts from anywhere without sending someone to walk the floor.
That's what Auki's building toward
when wallet infrastructure gets acquired, any project on it inherits a different risk model
the keys were never yours, the new custodian just made that obvious. the philosophical alignment that drove the integration no longer holds
this is what choosing a proprietary SDK looks like when the incentives change
protocols don't get acquired. the architecture is the commitment
https://t.co/QKMg7un9p5
Healthy direction.
The verification is real cryptographic attestation (Intel TDX, OS integrity, code hash), and the fact that it's open source means the checker itself is auditable.
Verifiability you can't independently check isn't worth much; this you can.
Pope asks, we build.
Behind all the hard work is a simple belief: AI should make humans more free, not more controlled.
That’s why privacy, trust, and verifiability matters even SV don’t believe these.
Demo: NTP time sync on the real world web is now much better than UTC time sync. Necessary for time sensitive collaborative perception and advanced orchestration.
This is running on some of our robots already too.
what's the right number of Stamps to feel confident about someone's humanity?
depends on what you're protecting.. every single method may start to crack when attackers target it specifically
how proof of humanity works in the age of AI: https://t.co/UyksgnBzCg
The real unlock here isn't another agent - it's that your memory, keys, and skills finally live in one place instead of the tool.
Switch from Claude Code to Cursor to Codex and lose nothing.
"iCloud for AI agents" - @openclawdiai is exactly right. 🦞
We just shipped a big update to https://t.co/nxbGPpIhuB
New website. New features. New way to think about what Clawdi actually is.
Here’s everything that changed 🧵
Every robot in that lab is learning the same lesson: where am I, and where is everyone else? Auki turned that question into infrastructure.
Solve it once, share it with every machine, and you've built the foundation the whole robotics industry is going to need.
Insurance, banking, and healthcare AI teams are all about to discover they can't ship without provable data protection.
The auditors will ask, the lawyers will block, the projects will stall.
@PhalaNetwork is one of the few teams that already has the answer ready. Confidential compute with verifiable attestation. They're going to be everywhere!👍🏼
When GDPR meets LLMs, “secure” isn’t enough. You need proof.
Can data escape?
Is the running code what was audited?
Can cloud admins read runtime memory?
What privacy-preserving compute means for AI data compliance:
https://t.co/HCMHNi7XRW
Picture a store employee with their hands full of inventory, wearing smart glasses, asking out loud where a product belongs, and being guided there turn by turn like Google Maps.
That's what @Auki is already demoing. The glasses don't even do the SLAM - they stream video to a GPU on the real world web that figures out where they are.
Once you see this working, every other approach to physical AI starts to feel clunky.
Demo: since glasses don't have the onboard compute and battery needed to track their own position, they stream video to a domain-side GPU machine on the real world web which does it for them in real time.
Necessary if you want your AI copilot to know where it is.