-
Anthropic’s $80B Loss, Explained
Anthropic reported an $80.7B net loss, but the business lost under $1B. Here is where the other $72B came from and how to explain it on a call.
-
The PHP Flaw Hit “70% of the Internet.” Here’s What It Means for You
Two critical PHP flaws hit headlines as “70% of the internet.” For most companies it means very little. Here’s how to tell, and why AI made it critical.
-
Your AI Has a Shutdown Date: How to Maintain an AI Solution
OpenAI turns off GPT-4o and GPT-4.1 in its API on Jan 14, 2027. How finance teams can test, pin and monitor the models under their AI tools.
-
AI Is Paying Off. The Data Isn’t Ready. Here’s Why That Takes Months
77% say AI is paying off, but only 5% say their data is ready. Why AI on the numbers that matter takes months of data work in regulated finance.
-
Bank of America Hired 1,000 Contractors. Your AI Shortcuts Are the Lesson
Bank of America added 1,000 contractors to patch flaws AI models exposed. For finance teams, the lesson is that AI shortcuts are now real risk.
-
28% or 75%? How Much of Your Job AI Will Do, and How to Score It
Anthropic asked US workers and its own staff how much of their job AI will do by 2030: 28% vs. 75%. Here’s how to score your own job, task by task.
-
AI Beat 12 CPAs. Here’s What the Study Actually Shows
Claude scored 100% and 12 CPAs averaged 37% in Mercor’s study. What it shows about AI, what it doesn’t show about accountants, and the sycophancy problem.
-
Open vs. Closed AI Models: What Finance Teams Need to Know
Open-weight or closed AI models? What each controls, who handles safety, and why the free model isn’t free. Plus how harnesses let you use both.
-
Whose Permissions Is Your AI Agent Using?
Service keys vs. user-based authentication for AI agents: when to use each, where shared API keys break your ITGCs, and what to check before go-live.
-
AI’s Edge Has an Expiration Date
AI hedge funds beat their peers by about 6% a year through 2017. Then the edge went to zero. What that means for AI business cases in accounting.
