Writing

Notes on AI, tools, and building software.

What changed, what matters, and what product teams can do with it.

AgentsInfraModelsOpenAIAnthropicProduct StrategyGoogleSecurity

AI Daily

AI Daily - 2026-10-20: Notion turns agents into a model-agnostic control plane

Oct 20, 20262 min read

Notion’s October 19 update combines model routing, longer-running agents, BYO AI, MCP, and admin credit controls into a more governable agent workspace.

AgentsModelsProduct Strategy

Read featured article

Recent writing.

Page 14 of 15

AI Daily

Mar 30, 2026

1 min read

AI Daily - 2026-03-30: AI product teams are prioritizing runtime correctness over raw feature count

Fresh releases across LangChain, Ollama, and llama.cpp show a clear shift toward reliability, attribution, and inference efficiency that directly impacts shipped AI UX.

InfraModelsProduct Strategy

Read

AI Daily

Mar 26, 2026

1 min read

AI Daily - 2026-03-26: SDKs are converging on agent reliability at runtime

OpenAI Node, LangChain OpenRouter, and llama.cpp shipped same-day updates that tighten streaming, schema correctness, and local runtime stability.

OpenAIAgentsInfra

Read

AI Daily

Mar 25, 2026

1 min read

AI Daily - 2026-03-25: Agent stacks are hardening around OpenAI Responses

LangChain’s March 23-24 releases focused on OpenAI Responses compatibility and safer prompt persistence, signaling a shift from feature racing to production hardening.

OpenAIAgentsInfra

Read

AI Daily

Mar 23, 2026

2 min read

AI Daily - 2026-03-23: vLLM 0.18.0 pushes inference from model hosting to platform ops

vLLM’s latest release adds gRPC serving, GPU-less multimodal preprocessing, and stronger KV/offloading controls, signaling a shift from raw model speed to production-grade inference architecture.

AgentsInfraSecurity

Read

AI Daily

Mar 21, 2026

1 min read

AI Daily - 2026-03-21: Inference stacks race on reliability and throughput

vLLM, Ollama, and llama.cpp shipped important serving and reliability updates in the last 72 hours, tightening the loop between performance, local tooling, and production safety.

AnthropicAgentsInfra

Read

AI Daily

Mar 20, 2026

1 min read

AI Daily - 2026-03-20: Product teams are shipping reliability controls, not just smarter models

OpenAI’s March 17-19 ChatGPT changes and DeepMind’s new AGI benchmark framework show the market shifting from raw model launches to production reliability, controllability, and measurable progress.

OpenAIGoogleModels

Read

AI Daily

Mar 19, 2026

1 min read

AI Daily - 2026-03-19: Agent tooling meets real-world user demand

Google’s Gemini API adds multi-tool orchestration while Anthropic’s 81,000-user study clarifies what production AI users actually need next.

GoogleAnthropicAgents

Read

AI Daily

Mar 18, 2026

1 min read

AI Daily - 2026-03-18: GPT-5.4 mini/nano turn frontier quality into high-volume defaults

OpenAI shipped GPT-5.4 mini and nano to API users, then retargeted gpt-5.3-chat-latest, signaling faster routing choices for product teams.

OpenAIAgentsModels

Read

AI Daily

Mar 17, 2026

2 min read

AI Daily - 2026-03-17: smaller models, faster agents, real-world loops

Mistral, H Company, and NVIDIA all shipped the same message this week: AI products are converging on fewer models, faster agent runtimes, and stronger real-world data loops.

AgentsInfraModels

Read

AI Daily

Mar 16, 2026

1 min read

AI Daily - 2026-03-16: ChatGPT’s GPT-5.1 retirement resets model defaults

OpenAI retired GPT-5.1 in ChatGPT and moved users to GPT-5.3 Instant and GPT-5.4 Thinking/Pro, signaling a faster model cadence and clearer product tiers.

OpenAIModelsProduct Strategy

Read