For the past year, OpenAI has led the industry as the mostly undisputed champion of Generative AI. This form of AI is great at understanding context and generating text, visual content, and even audio. Google’s Bard chatbot has received an upgrade, Gemini, that is set to compete with OpenAI’s ChatGPT.
## The Generative AI BoomWe’re excited to announce 𝗚𝗲𝗺𝗶𝗻𝗶: @Google’s largest and most capable AI model.
— Google DeepMind (@GoogleDeepMind) December 6, 2023
Built to be natively multimodal, it can understand and operate across text, code, audio, image and video - and achieves state-of-the-art performance across many tasks. 🧵 https://t.co/mwHZTDTBuG pic.twitter.com/zfLlCGuzmV
Since ChatGPT was unveiled in late 2022, we have witnessed a boom in generative AI services that has outpaced previous tech. It enabled chatbots to become more advanced than ever, to the benefit of online industries like e-commerce. It also supports user-generated content creation, a growing industry housed within social media and other digital spaces. Increasingly, our work and entertainment is moving online, as seen with streaming and iGaming platforms. In the future, we expect users to continue to find productive and leisure activities online, where they can watch streamed shows, buy from retail sites, or play slot games with jackpot prizes. In online industries like iGaming or streaming, there are hundreds of options that are available on-demand, giving them unique appeal for those who have an internet connection.
Generative AI can enhance most online industries. It not only makes content creation online easier for new entrants into tech and UGC markets, but it’s also an effective tool for established businesses, assuming it has been leveraged in an unobjectionable way. For example, chatbots have been used in e-commerce spaces for a long time, but ChatGPT made them much more powerful. As an industry, e-commerce is filled with individual merchants from small Etsy stores to corporate giants like Amazon. Gen AI can support both while leveling the playing field.
The Details Of Google Gemini
The Gemini AI model is, in reality, three different versions that are intended to run simultaneously across different devices. These three Gemini variants are called Ultra, Pro, and Nano. The names reflect the scale that these models are supposed to work at, with Pro accessible via Google Bard while Nano is available for smartphones running Android OS. Ultra, the largest and most powerful, isn’t expected to have a public release until early 2024. When it does, it will be integrated with the higher tier of Google’s Bard chatbot, which is called Bard Advanced.
Gemini was announced with everything but a dry corporate announcement. Instead, Google made strong claims about how advanced this new AI model was, including how it could outcompete ChatGPT and that it was rated at 90% in industry-standard Massive Multitask Language Understanding (MMLU) tests. The test pits AI models against 57 tasks that span a range of academic disciplines, where Gemini scored better than the average human expert. Multimodal was the word of the day during Google’s Gemini announcement. If ChatGPT is mostly text-based, then that’s one mode that the AI operates in. GPT-4 pushes some boundaries of the text mode, mainly by allowing users to upload documents and analyse them. However, Gemini promises multimodal functionality, meaning it’s a one-stop shop for creating text, images, audio, video, and even code. The promise of a mobile-only version is particularly exciting for those who work on Android. Gemini Nano will be integrated into future Pixel phone releases, allowing devs to “build AI apps and features that work offline.” Operating gen AI on mobile is also an effective cost-cutting measure, as most current AI models take up cloud server space which is becoming increasingly costly to operate.
No comments yet