Viewing profile — slipwalker

calendar spread.

GLM 5.x coding plan from z.ai, and opencode go plan.

looks a lot like a rag pull...

0. SWE turned manager against my better judgment 1. MacBookPro M4 base model. 2. roughly 10 years ( my last one was a MCP retina 2015, and before that linux notebooks lasting about…

for a long time i have replaced my nginx instances with openresty ( the open source version with Lua scripting ) and the CVEs seem to be less critical and solved faster: https://ap…

replace documentation for an agent/chatbot accessing a RAG with the collection of requirements/design docs plus the code repository. I have been doing a POC at my day job and it wo…

wezterm running herdr.dev with several pi.dev sessions.

i am assuming he is suggesting a 3 year depreciation cycle for the equipment, quite common on enterprises.

"former software developer" does not say much... which field of expertise, which tech stack are you comfortable with ? In current AI trends coding means very little compared to sys…

towards system design and solution architecture.

for my needs, raycast suffices, and it is one less app to install.

homebrew ( indispensable ! all the other apps below are installed/managed via homebrew ): - raycast ( keyboard shortcuts for clipboard manager, basic window management, pomodoro tr…

i have toyed around Tiny Llama 1.1B ( https://github.com/jzhang38/TinyLlama ) for the "conversational brain" of NPCs in a game i was dabbling.... it was fun.

drop horizontal layers architectures and focus on vertical-slices. This alone reduces the needed context window per feature by an order of magnitude.

Role playing… it’s basically lying and pretending with extra steps.

multiple pi.dev with opencode.ai/go plan; inside herdr muxer. works nicely at a very agreeable cost ( USD$10/mo ).

how about opencode's go ( https://opencode.ai/go ) it is very cost effective and the models are surprisingly competent....

just out of curiosity, which GPU are you running on your local machine ?

https://opencode.ai/en/go evaluate the Opencode's budget friendly "Go plan"

ship it !

are you using opencode ?

Elite (1984) used procedural generation to create EIGHT entire galaxies, each with 256 star systems, on a floppy disk that was less than 48 kilobytes!

i like opencode's zen https://opencode.ai/en/zen with opencode's TUI.

maybe this: https://www.skillsdirectory.com/ ?

I set my "personal instructions"/customization for any LLM tool to highly adversarial mode. Explicitly confront and question every word i type.