wyzer

-1 karma · joined August 9, 2026

wyzer··on The data center fight is heating up

Old news... The protest are going on for quite some time already. As long as the capital flows, do not believe anything will change significantly, except delays in building

wyzer··on AI is helping patients to solve their medical mysteries

See a doctor with your self-diagnosis and they will show you the door

wyzer··on A ledger of which of your own AI architectural decisions you could defend

what about just "remember to: ..." in order to memorize decisions?

wyzer··on Show HN: Write.md – A free, open-source, themeable Markdown editor for macOS

vscode + pandoc ... work like a charm (incl. side by side preview)

wyzer··on Apple Silicon and macOS VMs: Faster LLM Inference with llama.cpp

I see M1 Ultra host mentioned, are there any M1 pro or M3 pro results? Has anyne tried?

wyzer··on Self-Hosted Inference for Agents

Thanks

wyzer··on Small, self-hosted MCP that gives Claude read/write access to your Google Sheets

with native, I mean versus the claude provided connectors

wyzer··on Small, self-hosted MCP that gives Claude read/write access to your Google Sheets

Is this only for sheelts? any plan to extend it? What would be the reasons to prefer it over a native claude integration?

wyzer··on Jay Miner on the History of the Amiga [video]

The OG home computer

wyzer··on Self-Hosted Inference for Agents

Parent deprecated?

wyzer··on Mistral Patent for “Code implemented tool calls”

They exist so you can not be forbidden to do what you invented.. but there might be unintended correlation ;)

wyzer··on Muse Glimmer: 30B-parameter model optimized for always-on local agent workflows

How are you handling the tradeoff between quantization for device fit and accuracy loss on tool calling? That's where local agents typically break down in production.

wyzer··on Taxi drivers rarely die of Alzheimer's

before or after massive use of navigation apps?

wyzer··on How I use LLMs to learn complex topics

Kids use agents for exam training, rather than cheating. Which is great as long as the LLM does not hallucinate