> Three months before ChatGPT debuted in November, Facebook’s parent company Meta released a similar chatbot. But unlike the phenomenon that ChatGPT instantly b

> Three months before ChatGPT debuted in November, Facebook’s parent company Meta released a similar chatbot. But unlike the phenomenon that ChatGPT instantly became, with more than a million users in its first five days, Meta’s Blenderbot was boring, said Meta’s chief artificial intelligence scientist, Yann LeCun. “The reason it was boring was because it was made safe,” LeCun said last week at a forum hosted by AI consulting company Collective[i]. He blamed the tepid public response on Meta being “overly careful about content moderation,” like directing the chatbot to change the subject if a user asked about religion. Here's the link to Blenderbot: https://blenderbot.ai/chat https://blenderbot.ai/chat It's not a question about it being made safe, at all. I wouldn't mind in the slightest if any AI system, at least in these early days, would refuse to answer anything where the prompt or response had anything remotely to do with religion or any hot-button issue. What I do mind, though, is that if you ask Blenderbot to write Python code to look up latitude and longitude on Google Maps, it will respond with a brief sentence about installing the absolutely irrelevant Google Plugin for Eclipse, "a java-based ide." If you ask Blenderbot to "make me an itinerary for vineyards in Tuscany," it will respond "Do you want to do a tour of the chianti region? It is in tuscany," and spoon-feed bits of trivia if you keep asking it for more. In contrast, ChatGPT, for those who haven't used it, will immediately write blog-ready content for either of these prompts. The issue with Blenderbot, and Microsoft Tay before it, is that it assumed that people wanted line-by-line milquetoast conversational chat, and that it needed to deliver value as a friendly conversational partner in that context - which would mean it would inevitably be judged by its ability to respond ethically to prompts. What people actually wanted, and what ChatGPT gave them, was a beginner-friendly interface to generate believable long-form content on non-controversial topics that previously would require a human expert or researcher. ChatGPT redefined what "chat" could be, and that's a product innovation that the big players missed, not simply a willingness to take reputational risk.