I mentioned that OpenAI web crawler recently – the bot the company is now using to find content for its models, which web admins could block. The New York Times has blocked OpenAI’s web crawler, GPTBot, from using content from the publication to train its AI models. The change comes after the NYT updated its terms of service to prohibit using its content to train AI models and is considering legal action against OpenAI for intellectual property rights violations.
And speaking of OpenAI, the company has introduced fine-tuning to its GPT-3.5 Turbo model, allowing customers to customize the model and improve its reliability. Fine-tuning enables companies to make the model better follow instructions, improve formatting, and adjust the tone of the output. OpenAI claims that fine-tuned versions of GPT-3.5 can match or even outperform the base capabilities of GPT-4 on specific tasks. OpenAI also plans to retire the original GPT-3 base models on January 4, 2024.
Why do we care?
Talk with your customers about their feelings on their data being used by these models. There’s an easy value add to that conversation. The model changes key into the value of a provider – this is where you can bring expertise to bear with customers.