👍 Open vs. Closed Source: What's the Difference Between Sam Altman's and Elon Musk's AIs?

The AI industry is split into two camps in the debate over what source code should look like. This month, Meta unveiled the largest open-source LLM, Llama 3. Meta's Chief AI Scientist Yann LeCun criticizes Altman for ChatGPT being one of the most opaque projects. OpenAI defends its position, citing safety concerns. At the same time, Elon Musk promotes his chatbot, Grok, as the most open and transparent product on the market. But is it so?

🤨 What Is Open Source?

In 1998, the Open Source Initiative (OSI) was established to protect the idea of open source. According to their standards, such code must be fully available for use, modification, and distribution and must meet 10 criteria, including non-discrimination, open data, and free redistribution of modified software.

🙅‍♂️ Do Any Companies Meet All 10 OSI Criteria?

Most AI companies are only partially open. For example, Mistral shares the weights of its models — the numerical parameters that influence AI performance — but not the data or the training process itself. The same can be said for Elon Musk's Grok, which recently sued Altman for deviating from open-source principles. However, OSI points out that hiding the data on which the model was trained means it can't be considered open source. Thus, both ChatGPT and Grok, according to OSI, are closed-source solutions.

➕ What Are the Advantages of Open and Closed Source?

🛠 Proponents of closed source argue that it's easier to ensure safety, protecting people from scams and incorrect answers.

🧠 The main advantages of open source are that the technology is accessible to a broader audience, allowing developers to build their AI-based solutions while reducing AI-related risks thanks to a more informed community.

Which Camp Do You Support?

🤓 — Open Source
😎 — Closed Source

@hiaimediaen