Cut AI bills by 50% by updating to a newer model, simplifying prompt language, and forcing a response format. Killed two birds with one stone, because this type of image doesn't fuck up the response anymore (the old AI model used to go haywire and try to solve the JS quiz instead of listening to my prompt) and Watchdog is now way more reliable + cheaper to operate #watchdog !private

checking AI work #bagsoup

more AI work over lunch and after work to get smart prompts working

Shipped BoltAI v1.13.6, use AI Command with local LLMs via Ollama 🥳 #boltai

Experimenting with AI for writing blogs #life

working on improving creation ai assitant knoledge base

pay AI dev #interiorai #photoai

tweak #therapistai AI API parameters to make it less repetitive

🤖 played with Aider and it mostly working with Ollama + Llama 3.1 #research

talk about LLM finetuning and Alpaca and ChatGPT being just instruct-text model #life fxtwitter.com/levelsio/status…

write post about ai #makerpath

🤖 spent my evening writing a better console for some more advanced Ollama 3.1 projects. #research

when AI writes multiple lines, turn them into paragraphs, and apply same font-size etc #otter

study AI #glife

study AI #glife

Fixing various issues and preparing to launch my AI tool.

write content using ai, adapt extend and complete it with knowledge #life

#thecompaniesapi improvements on page content server & refactoring on robot ; launching our first LLM scenario in production env ; putting the 4090 to the test :)