OpenAI has been dominating the headlines. Now, Google is getting into the mix. What does it mean for consumers and marketers alike?
Last week, Google unveiled Gemini, its largest and most sophisticated large language model. Much has already been written about the delays, comparisons to OpenAI, and of course, what it all means for consumers, workers, and “AI watchers.” We sat down with Jim Fleming, Attentive’s Director of Machine Learning, to break down some key takeaways from this development, and what it means for marketers. Hint: It’s likely a signal of more rapid innovation.
How does Gemini compare to other AI models?
Right now, there aren’t a ton of features and capabilities that make Gemini stand out from previous AI models. The Gemini model is multimodal, which means that it supports different types of information, including text, code, audio, image, and video. GPT-4 has all of the same capabilities. It’s also important to note the impressive demo video Google released is heavily edited and not an accurate representation of the Gemini user experience. Allegedly, Gemini Ultra—the most interesting model from a machine learning perspective—outperforms GPT-4, but these assertions should be taken with a grain of salt.
Prompting can heavily impact Large language model (LLM) performance, and it’s hard to tell how it stacks up in terms of day-to-day performance without consistent and comparable benchmarks (which won’t be available until Google releases the model). The big thing to note is that GPT-4 is available today and Ultra won’t be out until sometime next year (and that’s already after it’s been delayed once). To date, no one has come close to GPT-4, so the verdict here is, “We’ll just have to wait and see.”
How will the competition between OpenAI and Gemini impact AI innovation and research?
The biggest takeaway from this announcement is that other companies are catching up to OpenAI, which will ultimately increase competition in the space. This is a great thing. Why? More competition typically means more innovation (and better products).
OpenAI has been the undisputed leader in generative AI language offerings. And just last month, they released custom versions of ChatGPT that enable any company—or individual—to build highly tailored GPTs for any application they can dream of.
However, the pressure is on—especially when it comes to sustainable pricing. Traditionally, being the single most powerful player in the space would mean OpenAI could charge anything they wanted for their tools. But they haven't because they’ve understandably wanted to maintain their early market lead. In fact, they’ve been subsidizing prices for a significant loss, with Microsoft footing the bill. But now that other players (like Google) are releasing competitive models, we may see pressure on pricing as companies will select the best models and build superior tools for their customers' specific needs.
In terms of actual capabilities and adoption, we can expect Google to “catch up,” but OpenAI will likely remain the leader, mostly due to the velocity they’ve already got going for them. Although Google and DeepMind are leaders in the AI space, they’ve only recently begun investing in general-purpose language models for end-user applications like Bard.
What are the real-world applications of Gemini?
Much like we’ve seen Microsoft incorporating AI into their user-facing applications like Bing and Office 365, we’ll start to see deeper integrations of Gemini within the Google suite of products. The user-facing version of Gemini will be visible through Bard, Google’s equivalent to Bing Chat or ChatGPT.
We’ll surely see it integrated in other Google products, too. Today, you’re probably familiar with seeing autocomplete and autoreply functionality within Google Docs and Gmail. These are great productivity enhancements that'll continue to get more sophisticated with the infusion of AI capabilities.
What does this all mean for marketers?
While it’s still too early to say how innovations like this will impact marketing, one thing is for sure: New players entering the space means building highly specialized AI solutions will be more accessible than ever.
And while general-purpose AI tools are here to stay, their interfaces will be more tightly integrated with existing tools like Google Docs, Gmail, Bing Chat—all of which can have significant implications on workplace productivity. The more AI can play a sidekick role for us in day-to-day tasks, the more we’ll be able to allocate our time to more strategic, and creative work.
It’s exciting to think about more capable models getting released into the wild, and subsequently, the emergence of smaller models with similar performance profiles as larger ones, too. Both for consumers and marketers, we’ll see more personalized content, integrations, and efficiency. For marketers, there will be even more opportunities to supersize your skillset and increase productivity.
As we continue to invest in AI ourselves, we’ll be keeping close tabs on new developments here (and will continue to share how they could impact the marketing tools you use).
To learn more about what you can do with Attentive AI, join the waitlist. And if you’re interested in helping us build the most performant AI-powered SMS and email tools, join our team.
