Hacker News: chtefi

Testing a Kafka Proxy: Taming Millions of Permutations

chtefi — Thu, 25 Jun 2026 18:29:48 +0000

Article URL: https://www.conduktor.io/blog/testing-a-kafka-proxy

Comments URL: https://news.ycombinator.com/item?id=48677427

Points: 2

# Comments: 0

Loop Engineering: When Generation Gets Cheap, Judgment Gets Expensive

chtefi — Thu, 25 Jun 2026 12:26:40 +0000

Article URL: https://sderosiaux.substack.com/p/loop-engineering-cheap-generation

Comments URL: https://news.ycombinator.com/item?id=48672354

Points: 2

# Comments: 0

How to Tell If Your Kafka Self-Service Is Working?

chtefi — Wed, 24 Jun 2026 18:14:04 +0000

Article URL: https://medium.com/conduktor/how-to-tell-if-your-kafka-self-service-is-actually-working-361cef450b81

Comments URL: https://news.ycombinator.com/item?id=48663659

Points: 2

# Comments: 0

New comment by chtefi in "MCP is definitely not dead"

chtefi — Tue, 02 Jun 2026 11:35:52 +0000

I don't like seeing people market "mcp is dead" and always say the same thing, talking about context and how CLIs are better, so wrote this.

What do you think, what have you saw in terms of usage/adoption/issues?

MCP is definitely not dead

chtefi — Tue, 02 Jun 2026 11:35:52 +0000

Article URL: https://sderosiaux.substack.com/p/no-mcp-is-definitely-not-dead-the

Comments URL: https://news.ycombinator.com/item?id=48368880

Points: 3

# Comments: 1

Kafka Partitions are the wrong ordering abstraction. Keys are

chtefi — Mon, 01 Jun 2026 20:04:59 +0000

Article URL: https://medium.com/conduktor/kafka-partitions-are-the-wrong-ordering-abstraction-keys-are-b54dc5b20b95

Comments URL: https://news.ycombinator.com/item?id=48361893

Points: 1

# Comments: 0

New comment by chtefi in "Ask HN: What are you working on? (May 2026)"

chtefi — Sun, 10 May 2026 23:31:24 +0000

Working on https://kapturekafka.dev, a desktop app for Kafka protocol inspection. Think Wireshark or Fiddler, but native for Kafka.

Useful to debug local Kafka apps against any cluster, intercepts the traffic, decodes the protocol. You see interesting (and weird) things when you look at the protocol. Still early, though already useful for local debugging when you know what you want.

Most agent reliability problems are data engineering problems

chtefi — Thu, 07 May 2026 15:29:26 +0000

Article URL: https://sderosiaux.substack.com/p/from-prompt-engineering-to-data-engineering

Comments URL: https://news.ycombinator.com/item?id=48050614

Points: 2

# Comments: 0

New comment by chtefi in "Unified macOS service console for launchd and Homebrew"

chtefi — Sun, 03 May 2026 15:33:35 +0000

I run a lot of small background services on my Mac (thanks Claude/Codex), and Homebrew services that I barely look at. It felt surprisingly hard to know what was actually happening on my machine and investigate/fix quickly, so I built Launchdeck: a fast TUI for launchd jobs and Homebrew services.

Unified macOS service console for launchd and Homebrew

chtefi — Sun, 03 May 2026 15:33:35 +0000

Article URL: https://github.com/sderosiaux/launchdeck

Comments URL: https://news.ycombinator.com/item?id=47997958

Points: 3

# Comments: 1

Agents Are Microservices with a Brain

chtefi — Tue, 28 Apr 2026 16:05:41 +0000

Article URL: https://blog.dataengineerthings.org/the-microservices-playbook-for-multi-agent-systems-4d386cef62e8

Comments URL: https://news.ycombinator.com/item?id=47936341

Points: 1

# Comments: 0

Do LLMs Dream? The Post-Transformers Generation

chtefi — Fri, 17 Apr 2026 13:07:58 +0000

Article URL: https://sderosiaux.substack.com/p/do-llms-dream-the-post-transformers

Comments URL: https://news.ycombinator.com/item?id=47805534

Points: 2

# Comments: 0

Is Anthropic Enshittifying their core product?

chtefi — Thu, 16 Apr 2026 12:58:29 +0000

Article URL: https://sderosiaux.substack.com/p/is-anthropic-enshittifying-their

Comments URL: https://news.ycombinator.com/item?id=47792340

Points: 2

# Comments: 1

Stream your terminal. Watch AI agents work. Like Twitch, for shells

chtefi — Fri, 10 Apr 2026 00:07:59 +0000

Article URL: https://github.com/sderosiaux/twitch-terminal

Comments URL: https://news.ycombinator.com/item?id=47711942

Points: 3

# Comments: 0

$200 subscription VS $3,650 in compute

chtefi — Wed, 08 Apr 2026 12:06:21 +0000

Article URL: https://sderosiaux.substack.com/p/200-subscription-vs-3650-in-compute

Comments URL: https://news.ycombinator.com/item?id=47689072

Points: 3

# Comments: 4

Kafka Explorer – Configuration, Protocol, Wire Format, Errors, Flink SQL

chtefi — Thu, 02 Apr 2026 20:51:38 +0000

Article URL: https://kafka-options-explorer.conduktor.io/

Comments URL: https://news.ycombinator.com/item?id=47620007

Points: 2

# Comments: 0

Paper Cuts #2: RAG is dead, long live memory

chtefi — Mon, 30 Mar 2026 16:17:35 +0000

Article URL: https://sderosiaux.substack.com/p/paper-cuts-2-agents-that-remember

Comments URL: https://news.ycombinator.com/item?id=47576230

Points: 3

# Comments: 0

Decisions, extracted knowledge, handoff context: reasoning data infrastructure

chtefi — Fri, 13 Mar 2026 13:21:25 +0000

Article URL: https://sderosiaux.substack.com/p/ai-agents-produce-a-new-kind-of-data

Comments URL: https://news.ycombinator.com/item?id=47364102

Points: 2

# Comments: 0

New comment by chtefi in "How to stop your AI agent from gaming its own KPI"

chtefi — Tue, 10 Mar 2026 14:20:22 +0000

Mitigations mentioned:

- Make the validator read-only for the agent. Mount it as read-only in the container, or hash your eval scripts at startup and verify before each run. If the agent can write to anything in its evaluation path, it can (will) game it.

- Log the full trajectory, not just the output. Every tool call, file diff, reasoning step. Then run a second agent over the trace with no knowledge of the KPI: it only knows what honest execution looks like and will use its internal honest alignment to assess it.

- Write system prompts like job descriptions, not optimization targets. Name a reviewer. Give the agent permission to fail ("if you can't hit the target, explain why").

- Walk your own prompts: what's the metric, what can the agent write, and can it reach the metric by modifying the measurement instead of doing the work? If yes, close that path.

How to stop your AI agent from gaming its own KPI

chtefi — Tue, 10 Mar 2026 14:20:22 +0000

Article URL: https://sderosiaux.substack.com/p/how-to-stop-your-ai-agent-from-gaming

Comments URL: https://news.ycombinator.com/item?id=47323658

Points: 1

# Comments: 1