AI

Turning an X (Twitter) archive export into native WordPress posts – 1150 tweets filtered and merged into 518, media sideloaded, my words back on my own domain.

Anthropic’s Fable 5 available in Claude That didn’t last long..  but what were they thinking to begin with? Marketing it for months as “too dangerous to release”, then when the govt believed them, it’s an “oops”..? I believe the ban won’t hold up though.

At BASSCOM AI conference in Sofia, Anthropic’s Patrick Dreisch delivers interesting insights now He says AI capabilities double every 7 months now This stroke me deep – “people are the bottleneck” by EnduroSat’s Krasi Stoev. Sad but true

Interesting AI model benchmarking at https://contextarena.ai/. I like how it visualizes the loss of attention as context increases – something many don't realize

Testing n8n AI agents: subscribing to YouTube channels, transcribing videos, analyzing sentiment, and emailing summaries. Easier setup than Node-RED so far.. The AI agent now monitors 89 relevant YouTube channels. It collects the transcripts in a spreadsheets and sends me a summary email at the end of the day.

Quite interesting: Dream 7B, the most powerful open diffusion large language model https://hkunlp.github.io/blog/2025/dream/

Testing llama.cpp with the 30B param LLM on my laptop. It can write code, but be prepared for a wait – the following snippet took over 20 minutes to generate.

Bing’s AI says our Tic Tac Toe game is a tie, I am confused?

#AI animated and colorized picture of your truly as of 1980, by http://myheritage.com Cool or spooky?

My profile picture toonified by #AI https://deepai.org/machine-learning-model/toonify TBH I look like a neanderthal with glasses. The rest of the family looks cute though, so there is hope.

Trained my D.O.G. (Digital arrOw Grabber) project's NNN (Natural Neural Network) to fetch back arrows through positive reinforcement #AI

Received my Elements of AI certificate today. A great online course, by the University of Helsinki and Reaktor #elementsofai #AI