Microsoft's Work Trend Index and Anthropic's Economic Index — two unrelated datasets — point at the same conclusion: AI value in banking is decided by the operating model, not the tooling. Here's what that means for corebanking.
Xi Jinping's WAIC 2026 keynote gave open-source AI exactly one sentence — and closed on being ready to change course. If your bank is building on Chinese open-weight models, that ratio is the strategy question. Part 1 of 3.
On 2 August, the EU AI Act's transparency obligations become enforceable. Six concrete scenarios from a bank's floor — chatbot, RM emails, market commentary, campaigns, call centre, agents — show why nobody owns this yet.
The market stopped hiring juniors because of AI — entry-level postings are down 32% in Switzerland alone. The data says that's the wrong conclusion, and the firms that see it have a rare window.
This week extends a signal that has been building for the past two weeks: enterprise AI is moving from “can it answer?” to “can it be controlled, priced, and run inside real operating constraints?” Senior engineers and regulated-enterprise buyers should care.
A company behaves a lot like a living cell — and 2.4 billion years ago, cells faced an event that turned a poison into an engine. For a lot of companies, AI may be that oxygen. Innovation isn't a project you bolt on; it's the metabolism itself.
Covering 2026-07-06 → 2026-07-12 · Research (ArXiv) + Market (Blogs) The week's signal: agent memory became an attack surface, and compute became a balance sheet. Both are control problems, not capability problems. LEAD — The GPU boom is being financed, not sold Hyperscaler commitments to CoreWeave and Nebius
The whole debate assumes agents replace software engineers. Flip the arrow: Andreessen's "builder," Andrew Ng's ratio flip, and a $401M one-founder company all point the other way — the engineer who masters agents absorbs product, design, and marketing.
AI is only becoming procurement-ready when it's wrapped in audit, identity, and constrained action — not when it's merely more capable. This week's evidence: controlled write paths, inference economics, and agents coupled to truth.
Coverage: 2026-06-28 → 2026-07-05 AI Signals Report: Securing and Scaling Agentic AI Systems Nicholas Blotti · July 05, 2026 This week’s research and market signals converge on the growing pains of agentic AI systems as they move from experimental demos to production-grade deployments. Key challenges include
Coverage: 2026-06-28 → 2026-07-05 This week in AI papers We keep an eye on new AI papers on arXiv, pick one or two that really matter each day, and share the key ideas — no hype, just clear explanations. LLM Daily – From Similarity to Vulnerability: Key Collision Attack
Coverage: 2026-06-28 → 2026-07-05 This week in the Daily AI Newsletter Each day we scan dozens of AI articles, deep-dive two in the audio brief, and share the key ideas — no hype, just signal. Daily AI Brief — 2026-07-03 83 articles · 2026-07-04 · 3
AI Video
I wanted a talking AI agent to front my newsletter. Getting there took four attempts and a switch to datacenter GPUs — so instead of narrating it, I interviewed the agent that actually built the pipeline: the dead ends, the trade-offs, and why I think there's a business here.
AI
Coverage: 2026-06-21 → 2026-06-28 AI Signals Report: OpenAI GPT-5.6 and the Agentic Safety Stack Nicholas Blotti · June 28, 2026 OpenAI closed the week with a limited preview of GPT-5.6 — a three-tier family (Sol, Terra, Luna) it frames as resetting the frontier on
AI
Coverage: 2026-06-21 → 2026-06-28 This week in AI papers We keep an eye on new AI papers on arXiv, pick one or two that really matter each day, and share the key ideas — no hype, just clear explanations. LLM Daily – MIRROR: Novelty-Constrained Memory-Guided MCTS Red-
AI
Coverage: 2026-06-21 → 2026-06-28 This week in the Daily AI Newsletter Each day we scan dozens of AI articles, deep-dive two in the audio brief, and share the key ideas — no hype, just signal. Daily AI Brief — 2026-06-26 118 articles · 2026-06-27 · 2
AI Strategy
"Adopt AI" sounds like one decision. For a European company that has to keep data strictly under control, it's really a bet across inference, data, weights, jurisdiction, hardware cost — and time. Why there's no easy choice between American, Chinese, and European models.
LegacyModernization
For thirty years, the most important software in a bank has also been the least understood. I have seen this up close, again and again. Modernizing legacy systems has been a thread running through most of my career — including migrating a private-banking ERP off the mainframe, a system that
AI
Coverage: 2026-06-14 → 2026-06-21 AI Signals Report: Securing and Scaling Autonomous AI Agents Nicholas Blotti · June 21, 2026 This week’s research and market signals converge on the rising complexity and security challenges of autonomous AI agents. Papers reveal new attack vectors and evasive behaviors in deployed
AI
Coverage: 2026-06-14 → 2026-06-21 This week in the Daily AI Newsletter Each day we scan dozens of AI articles, deep-dive two in the audio brief, and share the key ideas — no hype, just signal. Daily AI Brief — 2026-06-19 114 articles · 2026-06-20 · 3
AI
Coverage: 2026-06-14 → 2026-06-21 This week in AI papers We keep an eye on new AI papers on arXiv, pick one or two that really matter each day, and share the key ideas — no hype, just clear explanations. LLM Daily – Ghost Vectors: Soft-Deleted Embeddings Remain Reconstructible
AI
Coverage: 2026-05-24 → 2026-05-31 This week in AI papers We keep an eye on new AI papers on arXiv, pick one or two that really matter each day, and share the key ideas — no hype, just clear explanations. Unpacked by our trio: Alex the plain-language host,
AI
Coverage: 2026-05-17 → 2026-05-24 This week in AI papers We keep an eye on new AI papers on arXiv, pick one or two that really matter each day, and share the key ideas — no hype, just clear explanations. Unpacked by our trio: Alex the plain-language host,
AI
Coverage: 2026-05-10 → 2026-05-17 This week in AI papers We keep an eye on new AI papers on arXiv, pick one or two that really matter each day, and share the key ideas — no hype, just clear explanations. Unpacked by our trio: Alex the plain-language host,