Amidst the frenzy that’s the generative AI market, main gamers are fiercely vying for the shiniest product. For its half, Google, historically a extra measured participant on this race, unveiled a teaser video for his or her Gemini massive language mannequin this week. Nonetheless, issues took a controversial flip when experiences revealed the video was not really an actual time illustration of the AI in motion.
Within the demo video released by Google, the showcased AI mannequin reveals its multimodal capabilities, demonstrating a capability to deftly decipher and deal with info gleaned from dwell video and audio. It’s a formidable achievement for Google, notably within the fierce enviornment of competitors in opposition to the likes of OpenAI, the place it has lagged behind. Nonetheless, as reported by Bloomberg, the showcased demo was crafted by “using still image frames from the footage, and prompting via text,” somewhat than the real-time and vocal and video processing it appeared to realize.
On stage at Fortune‘s Brainstorm AI conference in San Francisco on Monday, vice president and general manager of Google Assistant and Bard Sissie Hsiao spoke about the contentious demo video, focusing on the benchmarks Gemini reached as a model, and how it’ll propel Google’s chatbot Bard.
“The video is completely real. All the prompts and the model responses are real,” Hsiao stated. “We did shorten parts for brevity, which we put in the video as information on making the video,” she famous.
The demo video shows the brand new AI mannequin’s multimodal capabilities, figuring out a squiggly line, then the curves of recent traces, culminating within the creation of the drawing of a duck. All through this course of, the mannequin constantly acknowledges every factor, providing duck-related details and solutions in real-time.
Hsiao highlighted the milestones conquered by Gemini, showcasing its skills in benchmarks that put AI fashions to the check, spanning highschool physics, skilled authorized quandaries, and ethical situations. In response to the Verge, Gemini Extremely beat OpenAI’s GPT-4 in 30 out of 32 benchmarks—an achievement value boasting about, though Gemini Extremely is not going to be launched till subsequent yr. For now, Bard makes use of the much less superior Gemini Professional, which is roughly akin to GPT 3.5.
Hsiao stated these Gemini fashions will proceed to enhance Google search in addition to the Google Bard chatbot, which she stated is “the most preferred free chat bot now in the market.”
