InsightInsight from our research and the field

All articles

Tech NoteLLM Inside2026.09.03

If an LLM only predicts the next token, how does it get work done? (Part 1)

An LLM predicts the next token rather than looking up an answer. How that objective enables useful work—and why fluent output is not proof.

김해수

Tech Note2026.08.18

Bringing Jaeger to XGEN: from Envoy sidecars to OTel auto instrumentation

Introducing Jaeger tracing alongside an XGEN service migration — running into the limits of the Envoy sidecar, and building distributed tracing with OpenTelemetry auto instrumentation instead.

Insoo Jeon

Tech Note2026.07.16

Why we turned the user manual into a solution guide wiki

Generating a web wiki, Word, Korean, and English documents from one Markdown source — and why we now operate the manual like a product.

sooanc

Tech Note2026.07.14

Setting up a blog authoring workflow developers will actually use

Letting developers publish from their own GitHub account without opening the main repository — lowering the cost of participation with Open Authoring.

sooanc

Tech NoteThe Harness Diary2026.07.10

We burned every retry re-grading the same answer (Part 10)

Quality retries were exhausting the ceiling. The model was progressing through tool calls while the judge re-graded the previous failed answer.

김진수