ChatGPT app is soon launching on Android, but research says the generative AI is slowing down

ChatGPT is coming up with an app for Android users soon. This comes nearly two months after OpenAI launched an iOS app for ChatGPT. Android users can "pre-register" on Google Play Store and get a notification when the app is available. 

OpenAI has not given a timeline for the app's launch so far. Reports say the app will be rolled out for Android users in the US initially, followed by other countries. 

Announcing ChatGPT for Android! The app will be rolling out to users next week, and you can pre-order in the Google Play Store starting today: https://t.co/NfBDYZR5GI

— OpenAI (@OpenAI) July 21, 2023

What's new on the app?

• While mobile users can still access ChatGPT via the web, an app will provide a better user interface experience. 
• The Android ChatGPT app is expected to look more or less like the iOS app except for voice assistance differences. 
• The app is also expected to get most of the web-based ChatGPT's features including the ability to sync conversations across devices. 
• If the popularity of the iOS ChatGPT app is anything to go by, then there will be at least half a million downloads of the app in the first week itself. 
• iPhone users downloaded the ChatGPT app about half a million times in the very first week.

[ALSO READ: Make ChatGPT work for you, not take your job]

However, there are some concerns cropping up about ChatGPT:

• According to one study, ChatGPT's latest version is getting dumber, slower, and less accurate than the previous versions. 
• Researchers at Stanford and UC Berkeley tested different versions of ChatGPT from March to June 2023 on four primary tasks - solving math problems, answering sensitive questions, visual reasoning and generating code. 
• They found that ChatGPT-4 was able to identify prime numbers with 97.6% accuracy in March, but by June, it got only 12 questions right, with an accuracy rate of a staggering low of 2.4%. 

GPT-4 is getting worse over time, not better.

Many people have reported noticing a significant degradation in the quality of the model responses, but so far, it was all anecdotal.

But now we know.

At least one study shows how the June version of GPT-4 is objectively worse than… pic.twitter.com/whhELYY6M4

— Santiago (@svpino) July 19, 2023

• Twitter users said that ChatGPT's poor performance could be due to OpenAI's cost cutting efforts. There is a possibility that the company isn't deploying its best LLMs to the public. 
• On the other hand, there are also contradictory studies showing that ChatGPT can match the intelligence of the top 1%. 

Can the latest AI match human creativity? New research from #UM researcher Erik Guzik and @UMontanaBiz suggests it's closer than we think.

👉 https://t.co/OG7SLPddUG pic.twitter.com/RDCGm0uJfm

— University of Montana (@umontana) July 6, 2023

• This observation was made by researchers at the University of Montana who tested ChatGPT in the Torrance Tests of Creative Thinking, often used for assessing creativity. 

[ALSO READ: 5 ChatGPT-like AI chatbots you can use]