OpenAI is bringing ChatGPT to landlines; all you have to do is dial 1-800-CHATGPT. You might be thinking, "Who uses landlines in 2024?" Surprisingly, around 25% of U.S. adults still have landlines (though if we’re talking about actual users, that number drops to just 5%), with seniors being the most likely group to use them.
OpenAI's chief product officer, Kevin Weil, says this is the "next step" in making "artificial general intelligence beneficial to all of humanity." If you're in the U.S., you can get 15 minutes of free calling each month, though standard carrier fees may apply.
So, what's ChatGPT sound like over the phone? According to TechCrunch, it's identical to Advanced Voice Mode. For context, OpenAI rolled out Advanced Voice to ChatGPT Plus and Team users back in September.
The feature offers more natural, real-time conversations, allows you to interrupt anytime, and senses and responds to your emotions. However, it faced controversy when users noticed that one of the voices sounded a lot like Hollywood superstar Scarlett Johansson.
Along with ChatGPT on landlines, it's also now available on WhatsApp. It works just like the regular ChatGPT app, letting you have back-and-forth conversations. It'll be curious to see how many people adopt it, especially since WhatsApp already has its own LLM from Meta.
Big tech companies are pouring huge amounts of money into training their AI models, and they consume a ton of power. When it comes to user-facing large language models, OpenAI is a clear leader. It remains to be seen if other competitors will follow this "let's bring LLMs to landlines" move.
As for the cost, training a large language model like GPT-4 is estimated to consume between 51,773 and 62,319 megawatt hours (MWh) of electricity, which is about 40 times more than GPT-3, which used just under 1,300 MWh. This massive energy consumption has raised concerns about AI's environmental impact.
In addition to energy, training these models requires substantial financial investment. The costs include hardware depreciation and energy consumption over the duration of model training.
0 Comments - Add comment