AI notes & digests
This page is the stable start point for AI writing on oguzhan.co: daily digests, Wednesday and Saturday deep dives, and the Sunday weekly wrap. I cover frontier labs, agents, safety, standards, and tooling in English first, with an original Turkish twin rather than a machine paste.
Digests are the feed — short, sourced, weekday signals. Deep dives are where I argue a position. Use the sections below when you want the latest without hunting the archive.
Digests
Last five daily AI digests (newest first):
- 18 Sep 2026 — Claude now leads a quarter of Anthropic’s R&D
- 17 Sep 2026 — OpenAI misalignment disclosure, six cases, DeepMind Institute
- 16 Sep 2026 — AI standards body talks, Huang’s veto, agent hotlines
- 15 Sep 2026 — Google opens Claude, agent boards, Khan’s old law
- 14 Sep 2026 — Pacing fracture: Hinton, Gomez, and threat intel
Deep dives
Longer analysis with sources and a clear take. This week’s Wednesday piece:
- AI standards body: pact, cartel, or FINRA for models? (Wed, 16 Sep 2026)
Saturday deep dives and the Sunday weekly wrap land on the same cadence when the week warrants them.
Topics
What this desk keeps returning to:
- Frontier labs & pacing — Anthropic, OpenAI, Google DeepMind, and who sets the tempo.
- Agents & tooling — what actually ships, what breaks in production, vibecode workflows.
- Safety, misalignment & standards — disclosure cases, governance proposals, threat intel.
- Cyber adjacent to AI — when model risk meets security practice.
New digests land most weekdays. If you only bookmark one URL on this site for AI, make it this hub. Turkish twin: Yapay zeka merkezi.