Top Global Fintech & Tech Influencer • Trusted by Finserv & Tech Global • Content & Influencer Services • Advisory for Digital Transformation • Speaking
Dear @Google, I asked my AI agent to only READ my fintech newsletters. Your consent screen offered everything or nothing.
Telling an agent “read only” is a request. A read-only permission from Google is a lock.
Allow users choose the lock. #AgenticAI#DigitalTrust
NVIDIA’s GR00T 2 transfers human skills to robots with just five task demonstrations 🤖
💬 NVIDIA’s GR00T 2 learns physical tasks from human videos and other robots, achieving 95% success after five demonstrations per task.
https://t.co/TwCMQFcnfy
NEW CONCEPTS: The Bank of England's "𝐦𝐢𝐧𝐢𝐦𝐮𝐦 𝐯𝐢𝐚𝐛𝐥𝐞 𝐛𝐚𝐧𝐤" : "bare metal" functions that keep serving through a cyber-attack. PRA chief Katharine Braddick calls it "an urgent question".
In an agentic world, resilience is not a cost centre. It is the product.
https://t.co/piM63KBfy8
💎 This week's video: In conversation with Aser Blanco, Global Banking Lead at NVIDIA.
Click to enjoy the full video interview : 👉 https://t.co/ERxH0iZBqy
❝ The first thing
OpenShell brings to the industry
is that the controls over what
that agent can do reside outside
of the sandbox. It is not the
agent monitoring the agent.❞ — Aser Blanco, Global Banking Lead at NVIDIA
One day after Jensen Huang’s GTC Berlin keynote, where NVIDIA launched the OpenShell Reference Architecture, I sat down with Aser Blanco, Global Banking Lead at NVIDIA, to talk about what it means for banks.
We cover:
🔺 Why the controls have to sit outside the agent’s sandbox, and the extra layers Vera CPUs and BlueField DPUs add at the hardware and network level
🔺 What financial services still needs on traceability and audit trails
🔺 Bank foundation models: why Aser says it takes two data scientists and a few days, not billions
🔺 Why fraud is the first use case, with payback in four to eight months
🔺 From data silos to one intelligence layer, and why banks may finally become customer-centric
🔺 5,000 agents for a data transformation, 20,000 lines of legacy code in an afternoon, and why the excuses have run out
🔺 Why this is a business revolution that the CEO has to lead
📍 @nvidia | GTCBerlin
Today is 11/11, World Frequency Day.
While silicon intelligence is being engineered at extraordinary speed, we remain energy beings in the quantum field.
At 11:11 your local time, pause with me to the UNITY track. 🙏
https://t.co/bRtccwtOt0
#WorldFrequencyDay#TunedInTogether
Visa’s new Agentic Commerce Benchmark shows that 54% of failed AI shopping runs still led to a purchase.
The underlying problem? Shoppers rarely know (or articulate) all their requirements. If AI agents just execute surface-level prompts instead of digging for missing data, buyer’s remorse is inevitable.
https://t.co/T3d6SB57dJ
This remake is brilliant. This makes me think that more businesses will follow what Goldman Sachs decided:
𝐫𝐞𝐦𝐨𝐯𝐞 𝐂𝐥𝐚𝐮𝐝𝐞 𝐜𝐨𝐝𝐞 𝐚𝐧𝐝 𝐭𝐡𝐞 𝐥𝐢𝐤𝐞𝐬 𝐟𝐫𝐨𝐦 𝐧𝐨𝐧-𝐭𝐞𝐜𝐡𝐧𝐢𝐜𝐚𝐥 𝐛𝐮𝐬𝐢𝐧𝐞𝐬𝐬 𝐞𝐦𝐩𝐥𝐨𝐲𝐞𝐞𝐬 𝐥𝐢𝐤𝐞 𝐭𝐫𝐚𝐝𝐞𝐫𝐬 𝐚𝐧𝐝 𝐛𝐚𝐧𝐤𝐞𝐫𝐬
(even though GS is a lead underwriter for Anthropic's upcoming IPO)
Samsung Wallet now sends USDC abroad from 82M US Galaxy phones. Dollars in, local currency out, chains invisible.
Whoever owns the wallet screen owns the customer.
But “zero-fee” is only Samsung’s side. The FX spread on the payout decides whether this beats Western Union.
https://t.co/D6EKyMfztO
Reality check: of the top 10 dollar stablecoins, only USDC and PYUSD would meet GENIUS Act reserve rules effective Jan 2027 (BPI). USD1 and Ripple USD fall short, backed by money market funds.
If agents pay in stablecoins, reserve quality matters.
https://t.co/xZJDJkz6hu
@ST_Automation It seems that most of became tech debt. Code that never reached production scale and maybe even became dangerous from a governance perspective.
UBS staff run 12,700 personal AI agents. Goldman pulled Claude Code back from bankers and traders. Morgan Stanley advisors made nearly 50,000 attempts to build agents; most never stuck.
Claude Reads the Account. Regulators Rank the Agent That Pays.
https://LINK
Claude Fable 5.2 builds a working operating system with 1,000 agents in eight days 🤖
Anthropic’s Claude models are taking on engineering challenges that go far beyond writing code.
https://t.co/eeAywUTQmq
The resilient Excel goes agentic in Brazil first! 📊⚡
Microsoft & Elo are piloting an AI agent that executes corporate payments directly inside Microsoft Excel via Pix, Open Finance, & Azure AI.
No more app-switching for SMB accounts payable.
The real test? Seeing if embedded governance & audit trails can prevent fraud at scale.
🔗 https://t.co/Q2XNU5fIb2
#Fintech #OpenFinance #AIAgents #CorporatePayments
Science Is Littered With AI Slop. A New Detector Could Filter It Out. 🧬 🤖
Rather than relying on text alone, the algorithm looks for clues in a paper’s reasoning.
https://t.co/rH2cQdLvud
🚨 Retro Revival alert: Sony revives the original Walkman headphones, adds USB-C 🎧
Would you be tempted to purchase these? Yes/No/No way! 🤔
https://t.co/ELxNJpKlsv
Agents in Banking: what I’m watching
Outside agents:
Liability: who keeps the evidence, who insures it
The push route: A2A via AP2, and a refund?
The bank’s own agents:
3. Where the agent register sits
4. Agentic PFM+: does Smart Cash go live?
Read my latest article https://t.co/62a24tZDQx
💎 This week's video: In conversation with Aser Blanco, Global Banking Lead at NVIDIA.
Click to enjoy video: 👉 https://t.co/ERxH0iZBqy
💬 Banks are preparing to run 10,000, 20,000, even 30,000 AI agents. Asking an agent “please don’t delete my files” is not a control.
One day after Jensen Huang’s GTC Berlin keynote, where NVIDIA launched the OpenShell Reference Architecture, I sat down with Aser Blanco, Global Banking Lead at NVIDIA, to talk about what it means for banks.
We cover:
🔺 Why the controls have to sit outside the agent’s sandbox, and the extra layers Vera CPUs and BlueField DPUs add at the hardware and network level
🔺 What financial services still needs on traceability and audit trails
🔺 Bank foundation models: why Aser says it takes two data scientists and a few days, not billions
🔺 Why fraud is the first use case, with payback in four to eight months
🔺 From data silos to one intelligence layer, and why banks may finally become customer-centric
🔺 5,000 agents for a data transformation, 20,000 lines of legacy code in an afternoon, and why the excuses have run out
🔺 Why this is a business revolution that the CEO has to lead
📍@nvidia | GTCBerlin