TIL about the Apple Music Classical App. It’s been out a couple of years and I just now DL’d it. Fun surprise this morning.

I love coffee. I don’t understand weak coffee. The best coffee I’ve ever had was in Istanbul. That is how coffee should be made. I subscribed to some Breville coffee club and all the coffee I got was … weak. Coffee should be, black and strong. Beans should be oily. Anything else is tea.

Sharing so I can remember to go back and read 😬

As someone who writes almost exclusively in Python and related frameworks, I got a lot out of this essay. I thought I knew what was going on w/FastAPI paths, but now I feel like I actually know.

Today we’re launching the text challenge of the Frugal AI Challenge for the AI Action Summit: detecting climate misinformation in the media (Press, TV, Radio), sponsored by the French non-profit QuotaClimat.
The goal of the task is to detect climate-based misinformation and to categorize its type 📃

I’ve spent most of the day dialing in a data pipeline that on paper should have been easy. But the most fun I had today was building bash scripts to do the things I needed. That was by far the most satisfying part of the project. #thelittlethings that just work

My position is that if we don’t consider the human element as we design the systems, then no. But if we use ‘human centered design’ then likely the human element will be persistent. Likely we’ll find there are situations that ‘either or’ or both. Good topic for over 🍻

Actually working on a project right now that I said to myself ‘I wonder where I can find data for this’ … now I can source it synthetically. Nice.

Funny building a demo for a POC and then finding a few days later that all the sensors are at 50% average and then you realize ‘oh yeah law of large numbers’ guess I’ll have rewrite the code that spoofs the data to avoid reaching the mean too early. Doh!

7 mile run on the cold and blustery day. Followed by an amazing potato leek soup my wife made and some reading.

Nothing says Christmas like Hanes Addiction Nothing Shocking playing whilst you decorate. Works for me.

Friday nights are for Taylor Swift, veggie chile rellenos and feature engineering experiments

This weekend I’ve been working on getting the #salesforce Agent Lite orchestrator running on @heroku.com So far the agent manager framework is running. The #Agent pool has been tricky, but I think I have a way of storing and training the models that will work. 🤞

I’ve been spending the last few days on GOFAI and I love it. Building apps that use Logistic Regression and another that uses a GAN. Just makes me smile. This weekend project is a CLI app that runs like an agent. Why? Because. 🤣

Last night might be the first time I’ve received a hug after telling someone my first love in code is Rails … he didn’t even get made when I told him my current love is Python for ML.

Spent that few nights building out use cases for agents and genetic workflows. There appear to be infinite possibilities. I love the idea of asking/typing a simple or complex need and having an agent go figure it out and come back. So far my needs are simple - find local trail and reserve parking.

A while back I was lamenting to a friend that there aren’t a lot stage 2 or stage 3 learning materials on #machinelearning … I think that it’s really they weren’t out, yet. Just finished @sebastianraschka.com book Q and AI. That’s what I was on about. A book discussing next steps.

This already feels like what twitter was a few years back. Here to talk about AI, ML, bikes, tattoos, music. If I follow you I’m not a bot, yet.