AI workflows vs. AI agents When building applications with large language models (LLMs), it’s tempting to reach for complex frameworks or fully autonomous agents. But Anthropic’s recent reflections show that simpler, well-structured systems often outperform heavyweight setups. A key takeaway: success doesn’t come from sophistication, but from clarity and fit-for-purpose design. One major distinction Anthropic […]
Why AI currently gets dumber
Morpeus explains why AI models are currently getting progressively dumber.
China
Apps like DeepSeek, TikTok, CapCut, Shein and Temu are increasingly taking over the global app market. In countless cutting-edge tech areas, China is either leading the pack or right on the forefront. Its breakneck progress in the tech industry is reshaping the competition with the United States, and Europe is in danger of falling further […]
Client Side Tool Calling with the OpenAI WebRTC Realtime API
OpenAI has just rolled out a WebRTC interface for their real-time API, now with tool-calling capabilities—so you can invoke JavaScript functions directly on the client side. This unlocks a host of exciting new possibilities. In this Youtube video, Craig demonstrates how he set up a Cloudflare Worker to securely forward the WebRTC stream. He then […]
Speech to Text locally with Whisper
38C3 – Self Models of Loving Grace
Coffeehouses and today’s innovative thinking
This month my boy friend and I made a trip to Prague and we visited Cafe Slavia. This made me think (again) of how a vibrant environment former coffeehouses must have been. Today they only server as touristic hotspots. But throughout history, coffeehouses have been central to the creative and intellectual lives of writers, artists, […]
Building an Erlang-like compiler/VM
Here I share the log from my first successful run of my new compiler/VM written in C. It’s all based on the book about Clox (https://craftinginterpreters.com/) by Robert Nystrom. You can see the processing of three pseudo-parallel processes in Erlang style. Printed it out straight away to hang on the wall because it was really […]
AI and the degeneration of human intelligence (German)
Maybe these are real problems in the next decades…
Billion dollar mistake – null reference and types
(Video from 2009: “Null References The Billion Dollar Mistake”) History Programs in the days were written in machine code. The positive things was, that you could exactly analyze and diagnose, why things go wrong. Designing the language Principle 1: The result of a programmer’s mistake in the code should be predictable by looking at the […]