On ChatGPT’s first anniversary, its mobile apps have topped 110M installs and nearly $30M in revenue
ChatGPT, since its launch a year ago, has significantly impacted the AI chatbot market. Its mobile app releases on iOS in May 2023 and Android in July 2023 have been pivotal, leading to over 110 million combined installs and nearly 30 million in consumer spending. The ChatGPT Plus subscription, priced at19.99 per month, offers benefits like access during peak times, faster responses, and early feature access, contributing to this success. Although ChatGPT is popular, it’s not the top revenue-generating chatbot app. Apps like Ask AI, which offers lower-priced premium options, have higher consumer spending. ChatGPT’s mobile apps have the highest download rates among generative AI apps, with significant user bases in India and the U.S. However, in terms of usage frequency, ChatGPT ranks third behind apps like Character AI and Chai. The firm data ai anticipates further growth in ChatGPT’s installs and revenue by the end of 2023.
Elon Musk, in an interview, strongly criticizes OpenAI, claiming it is deceptive about its use of copyrighted data for AI training. He asserts that contrary to OpenAI’s claims, the AI is undoubtedly trained on copyrighted material, including articles, quotes, and videos uploaded by various people. Musk discusses the complexities of copyright in the digital age, particularly how extensive collections of text and media are being used in AI training, often justified under fair use. He expresses concerns about the future legal implications and the pace of technological advancement, noting that by the time these issues are legally resolved, AI technology will have significantly advanced. Musk also shares his personal journey from being deeply worried about AI dangers to a kind of fatalistic acceptance, which has helped him cope with these concerns.
Stable Diffusion XL Turbo can generate AI images as fast as you can type
Stability AI recently launched Stable Diffusion XL Turbo (SDXL Turbo), a new AI image-synthesis model boasting rapid, near “real-time” image generation capabilities. This model can quickly transform images from sources like webcams. SDXL Turbo’s innovation is its ability to produce image outputs in a single step, significantly less than the 20–50 steps required by its predecessor. This efficiency is credited to the Adversarial Diffusion Distillation (ADD) technique, combining score distillation and adversarial loss to enhance realism.
Although SDXL Turbo images are less detailed compared to higher step count SDXL images, they offer impressive speed. When tested on an Nvidia RTX 3060, it generated a 3-step 1024×1024 image in about 4 seconds. The model shows potential for real-time AI video filters or video game graphics, although it still faces coherency challenges.
SDXL Turbo is currently under a non-commercial research license. Stability AI is open to commercial applications but has faced internal management issues and potential sale talks. Additionally, Stability AI has introduced Stable Video Diffusion for animating still images and offers SDXL Turbo demonstrations on Clipdrop and an unofficial demo on Hugging Face. Despite challenges and ethical concerns, advancements in AI image synthesis continue rapidly.
How to create multiple custom instruction profiles on ChatGPT
Hope you enjoyed today’s newsletter
⚡️ Join over 200,000 people using the Superpower ChatGPT extension on Chrome and Firefox.