AI NEWS
AI news, read by someone who has built it
Most AI news tells you what was announced. Almost none of it tells you what to do about it.
Latest
Showing 9–16 of 42
AI ANALYSIS / HISTORICAL RELEASE
OpenAI Codex: verifiable evidence is the real code-review feature
Published 25 October 2026. The announcement dates to 2025-05-16. What it established—and what a business must still test.
Read the note →AI ANALYSIS / HISTORICAL RELEASE
Claude Code for teams: adoption is a repository practice
Published 25 October 2026. The announcement dates to 2025-05-22. What it established—and what a business must still test.
Read the note →Analysis | 2024–2025 source review
Microsoft Agent Framework: when multi-agent orchestration earns its cost
Microsoft introduced Agent Framework in public preview on 1 October 2025, bringing AutoGen and Semantic Kernel foundations together. This analysis, published on 25 October 2026, examines that original architecture announcement, not a new release. The practical question is whether orchestration makes a business workflow more reliable than a single agent with well-designed tools.
Read the note →Analysis | 2024–2025 source review
MCP integration: build a testable contract between AI and tools
Anthropic introduced Model Context Protocol on 25 November 2024 to connect AI applications with external data and tools. Published on 25 October 2026, this analysis revisits that launch and the dated initial specification. Its focus is technical integration: what a client and server must agree on before an apparently simple tool call can become a dependable business operation.
Read the note →Analysis | 2024–2025 source review
A2A protocol: interoperability needs a task contract
This analysis is published on 25 October 2026 and examines the original 2025-04-09 announcement or report. It is not a claim of a launch this week or a current availability guide. Evaluate agent interoperability through discovery, task state and independently verified completion, not just message exchange.
Read the note →Analysis | 2024–2025 source review
Computer-use agents: choose the UI only when the API cannot do the job
This analysis is published on 25 October 2026 and examines the original 2024-10-22 announcement or report. It is not a claim of a launch this week or a current availability guide. Use visual automation for a bounded integration gap, with observable postconditions and a realistic recovery budget.
Read the note →Analysis | 2024–2025 source review
ChatGPT agent: put approval at the business commitment
This analysis is published on 25 October 2026 and examines the original 2025-07-17 announcement or report. It is not a claim of a launch this week or a current availability guide. Separate preparation from commitment, and require approval of the actual final transaction rather than the original broad instruction.
Read the note →Analysis | 2024–2025 source review
Deep research AI: audit claims, not citation counts
This analysis is published on 25 October 2026 and examines the original 2025-02-02 announcement or report. It is not a claim of a launch this week or a current availability guide. Use a claim-to-source evidence ledger and allocate human review according to the consequence of an error.
Read the note →
Want this as a keynote for your leadership team?
I turn the same material into a 45–60 minute keynote, customised for your industry and the decisions your room is facing.
BOOK AMER →
