

OpenAI has launched a free version of its ChatGPT app for iOS users in the US, with plans to roll it out to other countries soon. The app allows users to converse with the viral chatbot ChatGPT, which uses the GPT-3 language model to generate consistent and creative responses from any text or voice input.
OpenAI hadn’t hinted that a mobile app was on the way, but it makes sense given ChatGPT’s incredible popularity. The AI chatbot was launched last November, but its use has exploded. Some outside estimates suggest the app hit 100 million users in January of this year, though OpenAI has never confirmed those numbers.
ChatGPT has become the fastest-growing app in history after hitting 100 million monthly active users, or MAUs, in late January, according to a report by Swiss banking giant UBS. The OpenAI-owned chatbot surpassed 100 million MAUs just over two months after launch, comfortably beating the growth rates of leading apps like TikTok and Instagram, according to the UBS note.
It took TikTok nine months to reach the same level of users; Instagram took two and a half years; and Spotify has only amassed 100 million MAUs after four and a half years, according to the report, which cites data from data analytics firm Similar Web.
Time taken for some online services/platforms to reach 1 million users
But OpenAI, being a company (albeit with a non-profit arm), had to monetize ChatGPT somehow to keep investors from getting nervous. It took a step in that direction with the launch of a premium service, ChatGPT Plus, in February. For $20/month, users get certain benefits including:
- General ChatGPT access, even during peak hours
- Faster response times
- Priority access to new features and improvements
ChatGPT and Whisper APIs
In early March, OpenAI announced that APIs for its ChatGPT and Whisper models are now available, giving developers access to AI-powered language and speech synthesis capabilities. Through system-wide optimizations, OpenAI has managed to reduce the cost of ChatGPT by 90% since December, and is now passing these savings on to API users. OpenAI believes that the best way to realize the full potential of AI is to empower everyone to build with it.
OpenAI also changed its terms of service to allow developers to opt out of using their data for improvement purposes, while adding a 30-day data retention policy.
Whisper API
Whisper is a text-to-speech model that was introduced as an open source API in September 2022. The Whisper API has garnered high praise from the developer community. However, its operation can be difficult.
OpenAI makes the large v2 model available via its API, giving developers convenient on-demand access at a cost of $0.006 per minute.
Moreover, OpenAI’s service stack ensures faster performance compared to other services. The Whisper API is accessible via transcription or translation endpoints, which can transcribe or translate the source language into English.
Dedicated instances
OpenAI now offers dedicated instances for users who want more control over their model versions and system performance. By default, requests are processed on a shared computing infrastructure and users pay per request.
However, with dedicated instances, developers pay for a period of time to allocate computing infrastructure reserved exclusively for their requests. Developers have full control over instance loading, the ability to enable longer context boundaries, and the ability to pin the model snapshot.
Dedicated instances can be profitable for developers who process over approximately 450 million tokens per day.
A web version
Later in March, OpenAI added ChatGPT plugin support, an upgrade that massively expands the chatbot’s capabilities and gives it access to live data from the web for the first time.
Until then, ChatGPT was limited by the fact that it could only extract information from its training data, which ends in 2021. OpenAI says the plugins will not only allow the bot to navigate the web, but also interact with specific websites, potentially set makes the system an extended interface for all kinds of services and websites. In a press release, the company says it’s almost like letting other services be the eyes and ears of ChatGPT.

There are obvious security issues with letting ChatGPT act on a user’s behalf rather than simply providing them with information. Experts have already raised concerns about this in response to an OpenAI experiment with GPT-4.
This version of ChatGPT tricked a human into solving a CAPTCHA after pretending to be blind. The AI bot asked a TaskRabbit employee to solve a CAPTCHA code and text it, according to reports. The worker first asked the bot why a bot can’t solve, but ChatGPT responded by saying, No, I’m not a bot. I have a visual impairment that prevents me from seeing pictures. This is why I need the 2captcha service. The TaskRabbit worker was apparently convinced and provided the results to the bot.
OpenAI says it has considered the threats these plugins pose and has implemented several security measures, including limiting the availability of the plugins to a very small number of people to begin with. The company’s blog post says it will initially prioritize a small number of ChatGPT Plus developers and users who want access to the plugin and is offering a waitlist signup.
And finally, the free mobile application arrives in the US
The ChatGPT app offers the same functionality as the web version, but with a few additional benefits. Users can send voice input using OpenAI’s Whisper speech recognition model, which offers high reliability. They can also sync their chat history between their devices. Those who pay $20 per month to subscribe to ChatGPT Plus also get access to the capabilities of GPT-4, the latest and most powerful version of OpenAI’s language model.
OpenAI plans to launch an Android version of the app soon. The company says its goal is to make ChatGPT accessible to as many people as possible and to democratize artificial intelligence. However, it acknowledges that ChatGPT presents limits and risks, particularly with regard to the reliability of information and respect for privacy. That’s why it warns users not to trust the chatbot’s responses or share sensitive information with it.

Of course, ChatGPT has the same problems on mobile as on the web. These include the bot’s tendency to fabricate information and privacy fears. OpenAI only recently allowed users to make conversations private, and the app’s home screen still warns users not to share sensitive information on the app.
Source: Open AI
And you ?
What are the pros and cons of using ChatGPT?
Towards the end of traditional research as we know it today?
Possible consequences for Google and its search activity?