In partnership with
Welcome to the iPrompt Newsletter
That's all it took to hijack ServiceNow's AI agents—bypassing MFA, SSO, and every security control in between. One hidden flaw. Full platform takeover. If your company uses ServiceNow (and statistically, it probably does), your IT team needs to know about this before your adversaries do.
Meanwhile, OpenAI announced ads are coming to ChatGPT—your conversations will now influence what you're sold. Meta dropped $2 billion on an AI agent startup that China is investigating. And Boston Dynamics just revealed why betting your automation strategy on a single vendor is a mistake.
Here's what you need to know—and do—this week.
What investment is rudimentary for billionaires but ‘revolutionary’ for 70,571+ investors entering 2026?
Imagine this. You open your phone to an alert. It says, “you spent $236,000,000 more this month than you did last month.”
If you were the top bidder at Sotheby’s fall auctions, it could be reality.
Sounds crazy, right? But when the ultra-wealthy spend staggering amounts on blue-chip art, it’s not just for decoration.
The scarcity of these treasured artworks has helped drive their prices, in exceptional cases, to thin-air heights, without moving in lockstep with other asset classes.
The contemporary and post war segments have even outpaced the S&P 500 overall since 1995.*
Now, over 70,000 people have invested $1.2 billion+ across 500 iconic artworks featuring Banksy, Basquiat, Picasso, and more.
How? You don’t need Medici money to invest in multimillion dollar artworks with Masterworks.
Thousands of members have gotten annualized net returns like 14.6%, 17.6%, and 17.8% from 26 sales to date.
*Based on Masterworks data. Past performance is not indicative of future returns. Important Reg A disclosures: masterworks.com/cd
What you get in this FREE Newsletter
In Today’s 5-Minute AI Digest. You will get:
1. The MOST important AI News & research
2. AI Prompt of the week
3. AI Tool of the week
4. AI Tip of the week
…all in a FREE Weekly newsletter.
ChatGPT Gets Ads—Your Conversations Are Now Inventory
OpenAI will begin testing ads in ChatGPT within weeks. Free users and the new $8/month "Go" tier will see sponsored content based on their conversations. OpenAI claims ads won't influence answers—but the incentive structure just changed. When your AI assistant makes money by showing you ads, its loyalty splits. Plus, Pro, and Enterprise remain ad-free.
Your move: If you use free ChatGPT for anything beyond casual queries, this is your prompt to switch.
ServiceNow Flaw: One Email, Full Company Takeover
Security researchers discovered CVE-2025-12420—a critical vulnerability in ServiceNow's Virtual Agent. With only an email address, attackers could impersonate any user, bypass MFA and SSO, and execute privileged AI workflows. ServiceNow serves 85% of the Fortune 500. Cloud customers are patched. On-premise users: update now.
Your move: Forward this to your IT team with one question: "Are we patched?"
China Investigates Meta's $2B Manus Acquisition
Meta acquired Singapore-based AI agent startup Manus for over $2 billion—and China's Ministry of Commerce is now investigating for export control violations. Manus hit $100M ARR in eight months by building AI agents that actually complete tasks. The deal explicitly bars Chinese ownership, but Manus originated from Chinese startup Butterfly Effect.
Why it matters: The AI agent market just became a geopolitical flashpoint. Expect more scrutiny on cross-border AI deals.
Boston Dynamics Sells Out 2026—Here's the Real Signal
Boston Dynamics revealed its production-ready Atlas humanoid at CES. Every unit for 2026 is already committed to Hyundai and Google DeepMind. The robot lifts 110 pounds, swaps its own batteries, and starts factory deployment in 2028.
But here's what matters for you: Hyundai isn't buying Atlas to replace workers. It's building the proving ground to validate humanoid ROI before selling to everyone else. If you're planning automation investments, the playbook is emerging—and the leaders are testing internally first.
Our Angle: The Infrastructure Play Nobody's Talking About
While OpenAI chases ad revenue and Meta acquires flashy agents, Anthropic is executing a quieter strategy. Claude Code hit $1 billion in run-rate revenue in six months. The Model Context Protocol (MCP) now processes 100 million monthly downloads and has become the de facto standard for connecting AI agents to tools—adopted by OpenAI, Microsoft, and Google.
Here's what most coverage misses: Anthropic isn't trying to win the foundation model race. It's building the plumbing every agent runs on. Think AWS, not iPhone.
OpenAI's ads mean ChatGPT's incentives now partially align with advertisers. Claude's subscription-only model keeps its incentives aligned with users. In enterprise deals where trust compounds, that gap widens every quarter. Anthropic just reported it's outpacing OpenAI in enterprise adoption. Now you know why.
The CRM that saves teams hours every week
It's not about working harder — it's about having a CRM that actually thinks ahead. HubSpot Smart CRM learns how your team operates and adapts to make everyone more effective. Streamline day to day tasks and track the activity that actually matters to your business. The result? Your team gets back hours every week to spend on growth instead of admin work. Start free. See the difference.
AI Prompt of the Week
What it does: Forces AI to give you actionable intelligence instead of generic summaries.
The prompt:
I'm researching [TOPIC]. Act as a senior analyst briefing an executive who will make a $10M decision based on your analysis.
For each finding:
1. State the fact
2. Explain the second-order implication most people miss
3. Give me one specific action I can take in the next 7 days
Avoid: generic observations, hedging, anything in the first page of Google results.
Prioritize: contrarian insights, timing-sensitive info, specific numbers.
Why it works: The "$10M decision" framing activates more rigorous reasoning. The 7-day action requirement forces concrete outputs. The "avoid" list blocks the safe, useless defaults.
Real example: A reader used this to analyze a potential vendor acquisition. The AI surfaced that the target's key patent expires in 18 months—buried in SEC filings. That single insight changed the deal terms.
AI Tool of the Week
NotebookLM
What it is: Google's AI research tool that turns your documents into an interactive, queryable knowledge base—with citations.
Why you need it: Unlike ChatGPT or Claude, NotebookLM only answers from sources you upload. No hallucinations about your data. It can turn a 200-page report into a 10-minute AI-generated podcast, create flashcards, build data tables, and export to Sheets.
One-liner pitch: "It's like having a research assistant who's read everything you've uploaded and never makes things up."
Rating: ⭐⭐⭐⭐⭐ (5/5)
Key features:
Source-grounded answers with citations (zero hallucination risk on your docs)
Audio Overviews: turn any document into a podcast-style briefing
Data Tables: extract structured info from messy PDFs and export to Sheets
Free tier is genuinely useful; Plus tier bundled with Google One AI Premium
Best use case: Before your next big meeting, upload all relevant docs and ask: "What are the three things most likely to come up that I'm not prepared for?"
AI Tip of the Week
The "Critic Model" Technique
The tip: Run every high-stakes AI output through a second prompt that finds flaws—then fix them.
Why it works: Google just admitted single-model AI safety is broken. Their fix? A second AI that watches the first. You can do this manually in 30 seconds. LLMs are better at finding errors in existing text than avoiding them while generating.
How to do it:
Review this for: factual errors, logical gaps, missing context, and anything that would embarrass me if an expert read it.
Be harsh. Then provide a corrected version.
Limitations: Doesn't catch consistent blind spots. Works best for factual content. Adds ~30 seconds to your workflow.
Pro move: Use a different model for the critic step. Claude checking GPT (or vice versa) catches more errors because their blind spots don't overlap.
Your Move
You just learned:
ChatGPT's business model just changed—your conversations are now ad inventory
One email address could compromise your entire ServiceNow instance
The real AI infrastructure race isn't about models—it's about the plumbing
Now implement one.
Here's your 30-second action: Open a new email. Type your IT lead's address. Subject line: "ServiceNow Virtual Agent - are we patched?" Body: "CVE-2025-12420. One email address = full platform takeover. Cloud customers auto-patched. If we're on-prem, we need to update. Can you confirm?" Send.
That's it. You just reduced your company's attack surface. Most readers won't even do that.
Stay curious—and stay paranoid.
— R. Lauritsen
P.S. Still using free ChatGPT for work? Ads are coming. Claude Pro is $20/month. Your privacy is worth 67 cents a day.
P.P.S. Hot take: Meta overpaid for Manus. They bought a "wrapper" around other companies' models for $2B. Agree or disagree? Hit reply—best responses make next week's issue.