OpenAI Advanced Voice is finally here! Full testing & review
TLDRThe video reviews OpenAI's new Advanced Voice feature, recently released and available through a paid subscription. The host demonstrates how to set up the feature, explores different voice options, and tests its capabilities by asking it to assist with language pronunciation, tell stories, and perform various tasks. While the voice feature is useful for educational and interactive purposes, it has limitations due to censorship, preventing it from performing certain sound effects like meowing or beatboxing. Despite these restrictions, the feature excels in tasks like guided meditation, language assistance, and storytelling.
Takeaways
- ๐ OpenAI's advanced voice feature is now available, and the user has begun testing its capabilities.
- ๐ฒ To access the feature, you must download the ChatGPT app and subscribe to the paid plan.
- ๐ค Users can select from various voice options, such as Maple, to enhance their experience.
- ๐ฃ๏ธ The voice feature can help with language learning, including pronunciation, but it may have limitations with certain languages like Chinese.
- ๐ It can tell bedtime stories with sound effects, such as wind and waterfalls, making it suitable for children's storytelling.
- โ ๏ธ The feature has strict guidelines preventing it from making certain sounds or engaging in inappropriate conversations (e.g., moaning, flirting).
- ๐ง The feature cannot analyze or tune sounds, such as musical notes or guitar tuning, but it can still guide through the process verbally.
- ๐ง Guided meditation is available, helping users focus on their breathing and relaxation.
- ๐ค There's a level of censorship in place, as the feature does not perform certain actions (e.g., beatboxing, singing) that early testers reported were possible.
- ๐ฐ The video highlights that this tool can be useful for practical tasks like language learning, storytelling, and guided exercises despite its limitations.
Q & A
What is the main topic of the video?
-The video is a review and testing of OpenAI's new advanced voice feature in the ChatGPT app.
How can users access the advanced voice feature?
-Users need to download the ChatGPT app from the Play Store or App Store and subscribe to a paid plan to access the advanced voice feature.
What is the first step to activate the advanced voice mode?
-After subscribing, users need to press a white button in the bottom right corner of the app, and they will be prompted to activate the voice feature.
What type of tasks can the advanced voice feature perform?
-The advanced voice feature can help with language pronunciation, tell bedtime stories with sound effects, assist with guided meditation, and conduct mock interviews.
What are some voice options users can choose from?
-Users can choose from various voices, each with a unique tone, such as friendly, professional, or conversational styles.
Does the advanced voice feature have limitations?
-Yes, it cannot produce certain sounds like meowing, singing, or beatboxing directly. It also cannot analyze or respond to live audio, such as tuning a guitar.
Can the advanced voice feature analyze sounds?
-No, the advanced voice feature cannot analyze sounds. It suggested using a tuning app when the user tried to tune their guitar using the feature.
What is the userโs impression of the featureโs censorship?
-The user believes there is an extra level of censorship, as the feature was unable to perform certain tasks, such as making certain sounds or responding to flirtatious comments.
What examples of storytelling does the advanced voice feature offer?
-The advanced voice feature can tell creative stories, like one about a fox named Finn, using sound effects like rustling wind and crashing waterfalls.
What are some of the potential uses for the advanced voice feature?
-The feature can be used for language learning, telling bedtime stories, guided meditations, mock interviews, and friendly conversation.
Outlines
๐ฑ Accessing and Setting Up OpenAI's Advanced Voice Feature
The narrator introduces OpenAI's advanced voice feature, which they recently gained access to. They explain that the feature requires the official ChatGPT app, available on both Play Store and App Store. The feature is only available to users on a paid plan. Once subscribed, users see a white button in the bottom right corner, allowing them to activate the advanced voice mode. After permitting audio recording, users can select from various voice options, demonstrated by the narrator as they choose 'Maple.' The narrator begins testing the feature by asking for help with French pronunciation.
๐ฃ๏ธ Testing the Voice Feature: Pronunciation, Languages, and Fun Requests
The narrator starts experimenting with OpenAI's voice feature by testing its ability to help with French and Chinese pronunciation. The AI corrects their French pronunciation but isn't as effective in Chinese. They then shift to humorous and creative requests, such as narrating stories in different accents (British, Scottish, Indian, Australian), setting secret codes, and answering philosophical questions using code words like 'chocolate' for 'yes' and 'banana' for 'no.' Throughout the tests, the AI adheres to strict guidelines, avoiding certain topics and maintaining its programmed limitations.
๐ผ Digital Marketing Mock Interview and Sponsored Segment
The narrator engages in a mock interview with the AI to prepare for a digital marketing role. They discuss their background in SEO, paid ads, and growth hacking, emphasizing creativity in digital marketing. The AI provides feedback, suggesting they frame challenges as learning experiences. The paragraph also includes a sponsored segment introducing 'AI Portrait,' an AI-based service for generating professional headshots quickly and affordably. The narrator promotes its benefits compared to traditional photo shoots.
๐ธ Guitar Tuning and Exploring Voice Feature Limitations
The narrator attempts to use the AI to help tune their guitar by playing a string for analysis. However, the AI cannot interpret sound and suggests using a tuning app instead. The conversation shifts toward more personal and playful requests, such as asking the AI to demonstrate emotions, beatbox, or perform sound effects, all of which the AI declines. Despite these limitations, the AI maintains a helpful demeanor and offers alternatives where appropriate.
๐๏ธ Bedtime Stories, Guided Meditation, and Final Thoughts
In the final test, the narrator asks the AI to tell a bedtime story with sound effects and perform guided meditation. The AI successfully delivers a story with sound effects embedded within the narrative but declines requests for sound effects on demand. The narrator reflects on the limitations and censorship in the feature, comparing their experience with early testers who reported more sound capabilities. They conclude by encouraging viewers to share their experiences with the feature and promote their free weekly AI newsletter.
Mindmap
Keywords
๐กAdvanced Voice Feature
๐กSubscription Plan
๐กVoice Options
๐กPronunciation Assistance
๐กStorytelling with Sound Effects
๐กGuided Meditation
๐กVoice Censorship
๐กTuning a Guitar
๐กVoice-Activated Commands
๐กAI Limitations
Highlights
OpenAI's advanced voice feature is now available, but it requires a paid plan.
Users need to download the ChatGPT app from official stores to access the voice feature.
The setup includes choosing from various voice options after enabling advanced voice mode.
The AI can assist with language pronunciation, offering feedback for French and Chinese.
While the AI could correct pronunciation in French, it struggled with detailed Chinese tone corrections.
It can narrate stories with sound effects, making it useful for bedtime stories.
The AI was asked to perform accents (British, Scottish, Indian, Australian), which it handled well.
It follows strict guidelines and cannot perform certain sound-related tasks like singing or meowing.
The voice assistant cannot analyze external sounds, like guitar tuning, but suggests alternatives.
Although it cannot produce musical or moaning sounds, it can simulate sound effects within storytelling.
The AI adheres to censorship, unable to perform intimate or inappropriate tasks despite creative testing.
The system helps with mock interviews and gives practical tips for answering job interview questions.
The AI cannot identify pitches or analyze audio directly from a user, relying on tools instead.
Guided meditations are one of the practical applications of the advanced voice mode.
There is strict censorship, but testers found it useful for learning, storytelling, and language practice.