🤖 testing Claude 3 for some low hanging fruit code writing tasks

🤖 played around with adding extra context in some local Ollama models. Trying to test out some real-world tasks I'm tired of doing. #research

Exploring a new idea for a chat bot

used ChatGPT to break the ice on a half dozen small projects with a goal of eventually automating them #mylife

🤖 Used ChatGPT to help me finish off a GitHub Issues syncing cli tool because I'm tired of bulk creating issues in their web UI.

#thecompaniesapi run my phi3-128k flow using llama3.1 and the results are mind blowing, it's insane how good llama is at conserving context and original purpose even when supplied with thousands of tokens; also shipped multiple hotfixes in robot UI; about to merge a month of work and then hop on fine tuning

🤖 Wrote a git commit message bot using an LLM to create better commit messages #research

Let ChatGPT code a side project.

got distracted with my writer AIs — testing between llama3.1 vs gemma2 #leifinlavida

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

I finished a personal and customized AI chat platform for me and some friends to use.

add chat bot #billisimo

updated and updated three projects that are mostly bots I use to do other things. they are refreshed with a standardized deployment setup now which makes them easier to maintain #servers

check out Llama 3.1 #life

created several personal AI agents via ollama #leifinlavida

played around with a conversational chatbot tonight that's actually not bad. #mylife

Destroyed my other two Digital Ocean droplets, I now only have the one $6/mo Dokku droplet...told ChatGPT what I'm trying to achieve with this Line+ChatGPT bot and what I already have working, to get some help visualizing how I should proceed...created a Line developer account...enabled the Messaging API...added the Line Go SDK to my Go backend...next up: I think I want to read the example code provided by Line and ChatGPT and figure out how I can fit it into my existing backend code. I think getting the Line "Echo" bot working in my app would be a good next step. Lowered my phone's daily limit for Twitter and Reddit from 60 mins each to 45 mins each.

Worked on getting ChatGPT working with the Line echo bot example...created a separate file to test the OpenAI API call...hit an error, turned out I needed to pay separately for API credits...hit another error from my request not being formatted properly...finally got it working, now I need to try to integrate it with the Line echo bot...updated the code to use the ChatGPT code and deployed it, but the bot isn't responding anymore, I need to examine the logs...checked the logs with `dokku logs api` (`api` being the name of my app) and saw the error: "You didn't provide an API key."...need to add my API key as an env var...got that added with `dokku config:set`...tested the echo bot again and IT WORKED. It is insane insane insane how fast I can get stuff working when I can ask ChatGPT for help at every step to cut through all the usual Googling to figure out how to do something...next up, hmm, I'm not sure, either work on an accompanying Vue frontend or try fleshing out the bot to handle the particular use-case I have in mind...sent an email to another YouTuber to see if I can do a video with him.

try client side web based Llama 3 in JS #life webllm.mlc.ai/