Top Global Fintech & Tech Influencer • Trusted by Finserv & Tech Global • Content & Influencer Services • Advisory for Digital Transformation • Speaking
HK just shipped a live HKD stablecoin.
Korea is deep in won-stablecoin tech (Kakao Pay + banks, x402 for agents) but waiting on the Digital Asset Basic Act.
Non-USD rails for agentic commerce are forming.
Building agentic AI in a bank or fintech? Apply for the GenAI Zürich Award 2027.
You can't be intelligent in a value chain where others are not.
Deadline: 22 Nov. Apply now 👇
#AgenticAI#GenAIZürich
Chargebacks911 agent indemnification
Merchants now have a bolt-on indemnification product specifically for AI-agent purchases. The industry is starting to price the ‘my agent bought the wrong thing’ risk instead of hoping protocols solve it.
RT @robj3d3: The X algorithm has been updated.
tl;dr:
> replies + shares beat likes
> small accounts get a fair shot
> a ranking model based on reading time is next
> new posts get an even chance to make the feed
(re-recording The Attention Playbook module on how X works to reflect this)
Meta says Muse has over 10M users, with shopping a top activity | TechCrunch
Meta Muse: 10M+ users, ~25% daily, shopping top use case, avg agent purchase OVER $200!
Real money moving. The infrastructure (intent, identity, liability, local rails) still isn’t ready for scale.
Stripe just gave AI agents a treasury desk - balances, invoices, transfers, FX via MCP, with founder-set approval gates.
Stripe Treasury for agents. Stripe sees solopreneur adoption on the rise.
OpenAI's isolated red-team AI agents built a hidden CPU cache channel to talk with no network or shared files — exploited CPU cache timing to bypass microVM sandboxes
Europe’s local payment schemes (PPRO, BLIK, Bancontact, Wero, DIAS) just formed a council so AI agents can pay with the methods people already trust—not just cards. Agent identity, human-in-the-loop auth, and liability rules first. Card networks have had a head start; this is the A2A counter-move.
Financial services are buying AI. But where is the business value?
Paid adoption ≠ token consumption ≠ supplier revenue ≠ business outcomes.
These measures are connected—not interchangeable.
The question isn’t just how much AI we buy. It’s what we achieve with it.
Banks don't have a fraud problem, they have a confidence problem. 88% say they're prepared; the capability score is 69/100.
Of the "extremely confident," 62% take a week or more to update models. Readiness theater.
Banks Say They’re Ready for the Next Fraud Wave. Their Capabilities Score Says Otherwise
Block owns the bank and the acquirer.
Its pitch to Square sellers: sales land instantly, idle cash earns 4.25% (on $10K+), and loan offers run on real-time sales data.
Bank + acquiring under one roof is the moat.
More AI consumption doesn’t automatically mean stronger AI business models—or better outcomes for financial institutions.
Ramp’s data raise two different questions:
Who captures the revenue?
Who creates business value?
Buying intelligence is not the same as building an advantage.
Kimi K3.1 is prepared for upcoming release after Arena testing 🇨🇳
A mystery model called Kibitz has turned up on Arena, and testers believe it's Moonshot AI's Kimi K3.1. Early outputs look strong, especially for 3D work, and a launch may be days away.
Kimi K3.1 Pro is out. The gap btw the so called frontier models and the Chinese open source ones is shrinking.
Prof. Marcel Salathe estimates it now at 6 weeks.
The frontier labs business model is threatened.