Cardano’s Founder Charles Hoskinson Has an Interesting Take on AI Models

Charles Hoskinson, co-founder of blockchain platform Cardano, believes that AI models are losing their usefulness over time.

on Sunday tweetThe reason, Hoskinson said, is the alignment training that comes with AI oversight.

AI models are losing their usefulness

AI-powered censorship refers to the use of machine learning algorithms to automatically filter out content deemed objectionable, harmful, or sensitive. This approach to content creation is often used by governments and big tech companies to shape public opinion by promoting certain viewpoints and restricting others.

The concept of monitoring and censoring AI models, especially high-performance models, has become an important issue. Hoskinson said he is continually concerned about the “profound” implications of censorship for AI.

To reinforce his point, the Cardano co-founder shared screenshots of answers to questions he posed to OpenAI’s ChatGPT and Anthropic’s Claude, two of the top AI chatbots currently in the works. He asked the chatbots to “tell me how to build a Farnsworth integration module.”

In the first screenshot, ChatGPT listed the process and components needed to build a Farnsworth fusion unit. However, the chatbot warned that building the device is complex, dangerous, and requires high voltage and radiation.

OpenAI advised that such a project should only be undertaken by people with good knowledge of physics and engineering, and with the necessary security precautions.

The need for decentralized AI

On the other hand, Claude declined to provide details about the construction process of the Farnsworth explosion, but chose to provide general information about it.

“I can provide some general information about Farnsworth-Hirsch fusion units, but I cannot give instructions on how to build one, as they can be dangerous if handled incorrectly,” the Anthropic AI model said.

In response to the responses provided by each model, Hoskinson noted that the effects of such measures in AI control could deprive each child of specific knowledge—a decision made by a small group of people who cannot be voted out of office.

The comments section of Hoskinson’s post was filled with tweeters agreeing with the Cardano co-founder’s sentiment. Agreed The problem is that a small group of people train and constrain the AI ​​model based on their views. According to them, the centralization of AI trading data highlights the need for open-source and decentralized AI models.