Hypothetical science and dark matter

Bullet Cluster, hot gas in pink, dark matter in blue

In several previous posts in this blog, I have pointed out that physicists and cosmologists frequently forget the real world and construct theories based on speculations not based on concrete data, observations, or experiments, but only in their imagination. Let’s look at a few examples:

·         Is physics losing touch with reality? In this post, published in 2017, I used as examples the various theories about multiverses, the theory of everything, and some ideas that would make it possible to build a time machine (in principle, but hardly in practice).

·         Is physics: science or literature? In this post, published in 2018, I referred to Freeman Dyson's procedure for extracting energy from stars and the idea that an advanced civilization could resist the accelerated expansion of the universe.

·         Pathological science, wishful science, or ironic science? This post, published in 2023, points out eleven cases of scientific research without experimental basis.

The dark side of AI

Those companies that market Large Language Models (LLMs), such as OpenAI (ChatGPT and GPT-4), Google (GEMINI), Meta (Llama), Anthropic (Claude), and Deepseek, establish controls to prevent their tools from being used for nefarious purposes, such as assisting a young person in committing suicide, providing instructions for building Molotov cocktails or atomic bombs, helping to design viruses potentially usable in biological warfare, or teaching how to manufacture dangerous drugs.

The problem is that these controls are fallible, as demonstrated by cybersecurity expert Dave Kuszmar, who published a summary of his research in an article that appeared in the IEEE's Spectrum magazine in August 2026. Their fallibility makes them dangerous, but the companies in question seem to ignore the dangers and remain silent when they are brought to their attention.

Kuszmar has developed at least seven procedures to bypass the controls of commercial LLMs and obtain dangerous information. The first two were these: