INDEPENDENT RESEARCH / PRACTICE NOTES
This is my working notebook on AI engineering, strategy, knowledge infrastructure, organizational transformation, and the human systems that determine whether innovation becomes real capability.
The notes range from emerging research questions to practical operating models. Some will become papers, tools, talks, or products. Others are here because thinking improves when it is made visible.
Anthropic's September 29 release of Claude Sonnet 4.5 highlights stronger coding, computer use, and sustained work on complex tasks. Alongside the model come checkpoints for Claude Code, a memory tool and context editing for longer agent runs, and an agent software development kit (SDK).
The pairing is instructive. Longer autonomy arrives with better ways to see, constrain, and reverse the work.
The Defense Advanced Research Projects Agency's (DARPA) September 24 update describes its Mapping Machine Learning to Physics (ML2P) program. The program aims to relate machine-learning performance to physical electrical characteristics so designers can optimize not only for accuracy, but also for the useful performance returned by each joule.
At the tactical edge, that is not an efficiency exercise. It is mission engineering.
OpenAI's September 15 release introduces GPT-5-Codex, an artificial intelligence (AI) model optimized for agentic software engineering. The announcement emphasizes both quick interactive work and extended independent execution, including hours of iteration on large tasks and test failures.
When an agent can work longer, “the code runs” becomes an even less adequate definition of done.
Anthropic's September 9 announcement says Claude can create and edit spreadsheets, documents, presentations, and Portable Document Format (PDF) files inside a private computing environment. The release marks another step in artificial intelligence (AI) moving from suggesting content to producing the artifacts through which organizations make decisions.
Once the file arrives polished, the management problem becomes easier to miss.
Google DeepMind and its research partners report on September 4 that an artificial intelligence (AI) controller has undergone testing at the Laser Interferometer Gravitational-Wave Observatory (LIGO) in Livingston, Louisiana. Their Deep Loop Shaping method reduces control noise in a difficult mirror-control loop by a reported factor of 30 to 100.
The achievement is technically specific, and that is precisely why it offers a useful lesson for AI in critical systems.
Anthropic's August 27 threat-intelligence report describes artificial intelligence (AI) in use across multiple stages of cybercrime, including an extortion operation in which an agentic coding tool provides active operational support. The report also describes a low-skill actor using AI to develop and sell ransomware.
The change is not simply that attackers have a better advisor. It is that advice, tool use, adaptation, and execution can now be chained together inside a much shorter loop.
The Department of Defense's (DoD) August 20 announcement realigns the Chief Digital and Artificial Intelligence Office (CDAO) under the Under Secretary of Defense for Research and Engineering (USD(R&E)). The stated goal is to accelerate artificial intelligence (AI) transformation by placing data, analytics, and AI closer to the department's broader technology enterprise.
Organizational charts matter. They establish authority, access, and incentives. They do not, by themselves, create adoption.
The National Institute of Standards and Technology's (NIST) August 14 concept paper proposes security-control overlays tailored to artificial intelligence (AI) systems. The work adapts controls from NIST Special Publication 800-53 to generative, predictive, single-agent, multi-agent, and developer use cases.
The practical promise is not a new checklist. It is a better conversation between two professions that too often approach the same system with different maps.