My Agents Have Agents

Inspired by → this tweet. Cartoon co-created with Grok. See more of my AI co-creations

"Whereas in ~May (2025) I was amazed that some prompts produced code that worked out of the box, this is now my expectation. I can ship code now at a speed that seems unreal....The amount of software I can create is now mostly limited by inference time and hard thinking...These days I don’t read much code anymore. I watch the stream and sometimes look at key parts, but I gotta be honest - most code I don’t read...The important decisions these days are language/ecosystem and dependencies." - Peter Steinberger, founder of PSPDFKit

Comments

Post a Comment

Popular posts from this blog

Thinking & Reasoning Models

Image
Nikita Namjoshi 's video " How do thinking and reasoning models work? " from the Google for Developers channel explains the concepts behind "thinking models" or "reasoning models," such as Gemini, and how they use more computation at inference time to achieve better results in complex tasks. Video Summary The video focuses on how Large Language Models (LLMs) can be improved to handle complex tasks like coding, advanced mathematics, and data analysis by utilizing more compute power during the generation phase (inference or test time). The Problem: LLMs generate responses by predicting one token at a time. When solving complex problems, the model has to figure out the entire solution in a single pass to generate the correct final answer, which is difficult. The Solution: Chain of Thought (CoT): CoT prompting is a technique where the model is prompted to generate a series of intermediate steps (a "chain of thought") that lead to the final answer...
Read more

Skribbl.io Tips

Image
Skribbl.io is a multi-player competitive doodling & word guessing game with elements of Dumb Charades & Pictionary .  Each game consists of about 3 rounds. In every round someone has to draw their chosen word and others have to guess it to gain points.  When its your turn to draw, you have to choose a word (which can be a noun, verb or adjective) from three options and visualize that word in 80 seconds. Alternatively when somebody else is drawing you have to type your guess into the chat to gain points. The earlier you guess a word the more points you get. The person with the most points at the end of game will then be crowned as the winner . As a long time doodler, I like how the game is designed.  As the game requires you to visualize a word and draw on the spot, it is a great fun test of creativity. It has helped me to imagine better and also appreciate how other competitors think and draw. As a web developer, I admire how the minimalistic game is designed an...
Read more

This Week I Learned - Week 4 2026

Image
This Week I Learned -  * "Studies show AI-generated text responses are now rated as more compassionate than those written by humans – even when those humans are trained responders from crisis hotlines. This isn't because AI is genuinely more compassionate, but rather a sobering indictment of how rarely we listen in a non-judgmental way. The irony that an algorithm powered by a large language model – the type of machine learning that underpins many AI chatbots – might be perceived as a better listener than an authentic human reveals important insights about our human listening shortcomings. It's when our agendas, backstories and emotional triggers run the show, that true deep listening becomes thwarted. Large language models don't have motivations or desires. They are programmed to be compliant so that people will continue to use them. They therefore exhibit perpetual patience – never suffering from empathy fatigue. While such a feat is not something we humans can or s...
Read more