What is mattwood.fyi?

A live, structured feed of riffs and links from Matt Wood, Chief AI & Technology Officer at AWS — readable by people and queryable by agents as a typed knowledge graph.

Is there an API and is it free?

Yes. A public, read-only query API is available with no authentication and no API keys. See https://mattwood.fyi/openapi.json and https://mattwood.fyi/docs.

How should an AI agent use this site?

Fetch https://mattwood.fyi/feed.json for the full structured feed, or query the API for keyword/semantic search and the connection graph. See https://mattwood.fyi/agents/ for when-to-use guidance.

Exploring Claude/GPT Knowledge Cutoffs

Connections

Challenges

LLMs reward expertise The new item's reverse-engineering approach challenges the premise that LLMs reward expertise — if hidden model details can be systematically inferred by probing, expertise becomes less of a differentiator in working with these systems

Supports

Ethan Mollick: Meaningfully Better AI Releases Are Accelerating Mollick's observation that AI releases are accelerating meaningfully motivates the forensic reverse-engineering approach described in the new item, as faster release cycles make official documentation scarcer
A Rational Conversation on Where AI Is Actually Going | Benedict Evans on Lenny's Podcast Benedict Evans' rational conversation about where AI is actually going benefits from the empirical grounding the new item provides — inferring real training timelines and dataset mixtures cuts through hype with concrete evidence

Related to

Improving Gpt 5 6 Sol In Chatgpt Both focus on GPT-5 specifically — the new item probes its hidden training details while the existing item discusses improving GPT-5/6 solutions in ChatGPT
Introducing Shieldstral. Both concern opaque properties of frontier models — Shieldstral addresses safety/alignment properties while the new item addresses training provenance, both filling gaps left by limited official disclosure from labs
Neuronpedia Jacobian Lens: Interactive J-Space Explorer for Qwen3.6-27B Both use interpretability/probing techniques to extract hidden internal information from large language models — Neuronpedia's Jacobian Lens explores internal representations, while the new item probes models to infer training metadata

Develops into

Comparing 11 Different AI Models Exploring knowledge cutoffs across Claude/GPT is a narrower comparative investigation of model differences; the 11-model comparison develops this into a broader multi-model capability exploration including non-OpenAI models.

Book Prize Index Both concern the boundaries of knowledge retrieval systems—the Book Prize Index aggregates and makes discoverable award-winning books, while the Claude/GPT knowledge cutoffs item explores what information AI systems can access, both addressing how users find authoritative content.
ChatGPT Desktop App for Linux Now in Preview A desktop app with direct ChatGPT/Codex access raises the same knowledge-cutoff awareness issues explored in that piece — desktop users interacting with Codex need to understand model versioning and currency of information

Supported by

Kimi K3: Complete Developer Guide for 2026 Kimi K3 developer guide with 2026 framing directly relates to understanding knowledge cutoffs and temporal context of frontier models like those explored in that item

Type:	link
First published:	Tuesday, 11 August 2026 at 01:09 UTC
Location:	https://blog.sshh.io/p/exploring-claudegpt-knowledge-cutoffs
URL:	https://mattwood.fyi/i/300a42f1

I'm Matt Wood, and this is For Your Information. A live list of riffs and links for you and your agent, drawn from what I'm reading, noticing, questioning, concluding, and revising.

Connections