Rolling out: ChatGPT's advanced voice mode

Howdy, wizards.

Check out friend. What in the black mirror episode is this, you ask? It’s the simplest UX for an AI-wearable ever, yet I wouldn’t be surprised if this one became a hit with young people. At least more successful than any Humane pins or rabbits. Indeed, the bar is very low. Can be pre-ordered now for 99$.

Dario’s Picks

The most important news stories in AI this week

1. OpenAI has started rolling out the new advanced Voice Mode. A small number of ChatGPT Plus users are getting the Alpha version of advanced Voice Mode, which OpenAI first demoed nearly 3 months ago. It offers more real-time conversations with a more natural-sounding voice that senses and responds to your emotions. It also allows you to interrupt the AI at any point with your voice. The new feature is launching to Plus subscribers on a rolling basis with plan of giving everyone access “by the fall”. OpenAI has been working behind the scenes to test the new mode in 45 languages, and implement guardrails to avoid abuse: it only speaks in four preset voices (none of which sound like ScarJo) and violent and copyrighted content is blocked.

The new feature seemingly works as advertised – here's some demos from Alpha users.

PS the video and screensharing capabilities that were demoed together with the new voice mode will launch “at a later date”.

‍

Why it matters‎ ‎ The new advanced voice mode is nearing human-level response times and expressiveness, an upgrade of several notches versus anything that’s currently on the market. It’s likely to open up practical use cases where real-time voice conversations with AI is handy, as well as social applications that benefit from natural sounding voices such as translations and language learning, storytelling, roleplaying, therapy aid, and much more.‍

‍

2. Meta's AI Studio now lets you create and discover characters. Meta just launched a platform to create and explore AI characters, leveraging their social apps for distribution. Characters can be made on any topic or, for content creators, as an extension of you (based on your Instagram profile and more). It’s rolling out first in the US. Look out for a deep dive on AI studio in your inbox tomorrow, where I’ll cover this in full detail.‍

‍

‎ Why it matters‎ ‎ Meta just launched an ecosystem akin to custom GPTs, but with a lot of UI improvements and better distribution. This could be big not only for getting regular folks using AI (it’ll be available right in their DMs) but also building their own AI apps/characters. It’s also an interesting and potentially profitable opportunity for content creators to leverage their influence and expand their reach.‍

‍

3. Perplexity gets a cease-and-desist letter from Condé Nast. It's only a month ago since Perplexity got caught red-handed ripping off Forbes' content inside its new "Pages" feature, with minimal attribution. Wired also found that Perplexity wasn't respecting robots.txt, a file website owners use to block bots from scraping their content.

‍

‎ Why it matters‎ ‎ Perplexity has been under a lot of heat lately, with good reason. To be fair, it's far from the only company being accused of plagiarism and copyright infringement in these AI times. However, the response matters – Perplexity initial shrugged off the plagiarism accusations as "rough edges" of a new product, and didn't lay out a meaningful plan to make things better, infuriating journalists worldwide.