Field notes.
What I read, what I test, what I ship. Articles on quantitative finance, AI applied to markets, risk management (ALM), and the tools I use daily: Python, Claude, n8n, Pine Script, Bloomberg.
OpenAI Files for IPO: The Hidden Cost of AI Models
On June 8, 2026, OpenAI filed confidentially for an IPO, a week after Anthropic. What going public will expose about the real cost of AI models.
Read article →GEO, GSO, AEO: what is the difference, and which one you need?
GEO, GSO and AEO name the same fight: getting cited by AI. The only question that matters is whether ChatGPT and Perplexity cite your site or your competitors'.
Read article →AI agents: whoever owns the semantic layer wins
On June 7, 2026, a SiliconANGLE analysis marks a shift: the value of AI agents now sits in the data semantic layer, not the model. Microsoft, Snowflake and Databricks are fighting to own it.
Read article →Apple picks Gemini for Siri: the AI model is now a setting
On June 8, 2026, Apple handed Siri to Gemini and will let users pick Claude or ChatGPT. Why the AI model is becoming a swappable component, and what it means for companies.
Read article →AlphaSense hits $7.5bn: research becomes an AI agent
On June 3, 2026, AlphaSense raised $350M at a $7.5bn valuation. The real moat isn't the model, but 500 million documents and the SuperAnalyst agent.
Read article →ZoomInfo GTM.AI: data as the bedrock for AI agents
On June 1, 2026, ZoomInfo launched GTM.AI, a verified context layer that feeds AI agents over the MCP protocol. A bet on data while the stock fell 50%.
Read article →Anthropic files its S-1: enterprise AI goes public
On June 1, 2026, Anthropic filed a confidential IPO draft with the SEC at a $965bn valuation. What its S-1 reveals about the real business of enterprise AI.
Read article →SEO vs GEO: what changes for your rankings in 2026
SEO vs GEO in 2026: SEO ranks you in Google's blue links, GEO gets you cited inside the AI answer. Why your competitors are already there, and you are not.
Read article →Getting cited by ChatGPT, Perplexity and Google AI Overviews
If ChatGPT does not cite your site, your competitor wins the client who asked. Why your site is not cited by AI, what it costs you, and how to check where you stand.
Read article →What GEO is, and why your site needs it
GEO decides whether ChatGPT, Perplexity and Google AI cite your site or your competitors'. What it is, what it costs you, and how to see where your site stands.
Read article →Microsoft Build 2026: 7 in-house AI models cut OpenAI ties
On June 2, 2026, Microsoft unveiled seven in-house AI models at Build, led by MAI-Thinking-1 (35 billion active parameters). The long-standing Azure-equals-OpenAI equation formally breaks.
Read article →GitHub Copilot moves to tokens: the end of flat AI pricing
On June 1, 2026, GitHub switched Copilot to token billing. The unlimited flat plan is gone, replaced by AI Credits priced against each model's API rate, and the first invoices are jumping ten times or more.
Read article →Snowflake buys Natoma to govern AI agents
On May 27, 2026, Snowflake acquired Natoma, a 27-person MCP gateway. Granting data access is no longer enough; the real stakes are controlling what agents are allowed to do.
Read article →OpenRouter hits $1.3bn: the end of single-model AI
OpenRouter raises $113M Series B led by CapitalG at a $1.3bn valuation. The model-routing layer between 400+ AI providers is becoming the strategic infrastructure of enterprise AI.
Read article →Anthropic targets $900bn and leapfrogs OpenAI
Anthropic closes $30bn this week at a $900bn valuation. Sequoia, Altimeter, Greenoaks lead the deal. Why the market is paying that price, and what it means for buyers.
Read article →Intuit cuts 17% of staff to accelerate AI
On May 20, 2026, Intuit announced 3,000 layoffs while reporting a 10% jump in quarterly revenue. This is not a crisis. It is the playbook for what a profitable SaaS vendor looks like when it rewrites itself around AI.
Read article →OpenAI x Dell: Codex leaves the cloud for on-premises data centers
On May 19, 2026, OpenAI signed a multi-year deal with Dell to run Codex in hybrid and on-premises environments. The first frontier offering outside the public cloud, built for regulated sectors.
Read article →KPMG signs with Anthropic: three Big Four on Claude in seven months
On May 19, 2026, KPMG rolled out Claude to 276,000 staff. With Deloitte and PwC, three of the Big Four have signed alliances with Anthropic. Hourly-rate consulting is ending here.
Read article →Anthropic buys Stainless: who owns the AI agent plumbing
On May 18, 2026, Anthropic acquired Stainless for over $300 million. The SDK toolkit powering OpenAI, Google and Cloudflare APIs has changed owner, and changed sides.
Read article →SAP bets on Claude as the brain of its autonomous enterprise
On May 12, 2026, in Orlando, SAP made Anthropic's Claude the primary reasoning model of its Autonomous Suite. The move resets the back-office stack and raises a rare question in enterprise software.
Read article →ChatGPT plugs into your bank: OpenAI takes on personal finance
On May 15, 2026, OpenAI wired ChatGPT directly into 12,000+ US banks through Plaid. The move looks routine; it shifts the layer where personal financial advice gets delivered.
Read article →Anthropic overtakes OpenAI in enterprise AI
On May 13, 2026, Ramp's AI Index put Anthropic at 34.4% of business adoption, ahead of OpenAI at 32.3%. The first crossover since ChatGPT walked into the office, the same week PwC committed to certifying 30,000 staff on Claude.
Read article →Slack becomes Salesforce's free CRM front-end
On May 11, 2026, Salesforce confirmed that every new customer gets a free Slack workspace, already wired into the CRM. A strategic shift that moves the CRM interface out of the browser and turns Slack into the orchestration layer for Agentforce agents.
Read article →ServiceNow brings AI agents to every business function
At Knowledge 2026 on May 5, ServiceNow stopped selling copilots and shipped Autonomous Workforce specialists that close cases across IT, CRM, HR and security. The L1 IT desk resolves 99% faster. The control tower becomes the real product.
Read article →Sierra raises $950M: the AI customer agent becomes a category
On May 4, 2026, Sierra closed a $950M Series E at $15.8bn valuation. Bret Taylor turns ARR (from $100M to $150M in six months) and 40% of the Fortune 50 into a self-service platform. The customer-facing agent war just changed gears.
Read article →Anthropic and OpenAI launch their AI consulting arms
On May 4, 2026, within hours of each other, Anthropic and OpenAI rolled out two joint ventures with Blackstone, Goldman Sachs, TPG, and Bain Capital. Anthropic raises $1.5bn, OpenAI targets $10bn. Shared model: Palantir-style forward-deployed engineers. Shared target: McKinsey, BCG, Accenture, Deloitte.
Read article →Salesforce Agentforce Operations: the back-office turn
On April 29, 2026, Salesforce shipped Agentforce Operations, its first AI agent platform built for the back office. Audit cycles cut by 50% to 70%, 30 prebuilt blueprints, native Flow integration entering beta this May. The CRM giant takes the fight to SAP and Microsoft.
Read article →Mistral Workflows: France's bet on enterprise AI orchestration
On April 28, 2026, Mistral AI opened Workflows in public preview, a durable orchestration engine on Temporal already running in production at La Banque Postale, France Travail, ASML and CMA-CGM. The missing piece in Europe's stack against OpenAI and Anthropic.
Read article →Bloomberg AskB: Anthropic Claude lands inside the Terminal
On April 28, 2026, Bloomberg opened AskB in beta to 125,000 of its 375,000 Terminal users. Under the hood: Bloomberg-trained models and Claude. Why this AI agent rewrites the Terminal and what it signals for ALM teams.
Read article →ECB's QRF: how a random forest has quietly steered inflation calls since 2022
Four ECB economists confirmed on April 21, 2026 that a Quantile Regression Forest has been preparing monetary policy decisions since late 2022. The model called the 2025 inflation surprises within 20 basis points.
Read article →ECB April 30: stagflation forces a new ALM stress test
On April 30, 2026, the ECB holds at 2.00% with 2026 inflation revised up to 2.6% and growth cut to 0.9%. Why this call reopens, for European ALM teams, the scenario nobody wanted to write down.
Read article →Taiwan: 16 banks unite to build a sovereign banking LLM
On April 23, 2026, sixteen Taiwanese banks pooled funds to build a sovereign banking LLM for NT$40-70 million. CTBC leads, the FSC supervises. What this signals for Europe's financial AI strategy.
Read article →Aspect Capital: Martin Lueck pushes back on AI quant trading
On April 23, 2026, Martin Lueck (Aspect Capital co-founder, $7.8bn AUM) refuses to hand portfolio decisions to AI. Why a quant pioneer's warning also hits banks.
Read article →Google TPU 8: the agentic era lands on silicon
On April 22, 2026, Google unveiled its TPU 8t and 8i; Anthropic committed to 3.5 GW. Inference is now the economic front line for AI in finance.
Read article →MCP: the flaw catching up with banks on agentic AI
Anthropic's MCP design flaw, OX Security's 200,000 exposed instances and the bank-side fallout: who owns the third-party risk now?
Read article →Anthropic-Amazon: $100bn and 5 GW to keep Claude running
On April 20, 2026, Amazon commits up to $25bn more in Anthropic and Anthropic pledges $100bn in AWS spending over a decade. Behind the number: a circular model and a new risk line for finance.
Read article →Systemic AI: the Bank of England moves to simulations
On April 16, 2026, Deputy Governor Sarah Breeden confirmed the BoE is now stress-testing AI for financial stability. A methodological pivot that opens the era of AI-specific stress tests for banks.
Read article →Wall Street Q1 2026: 5,000 fewer jobs, more AI
The six largest US banks posted $47.3bn in profit while cutting nearly 5,000 jobs. Behind the efficiency narrative, the AI line item keeps growing.
Read article →