Site icon Aragon Research

Google Strikes Back at Microsoft and OpenAI with Gemini Services and its New AI Hypercomputer

By Jim Lundy

 

Google Strikes Back at Microsoft and OpenAI with Gemini Services and its New AI Hypercomputer

Wednesday, December 13th, Google made a series of announcements as a follow-up to its Gemini LM announcement last week. The number of new capabilities in both hardware and software signifies a new race that has emerged where updates from major providers are happening every quarter. This blog summarizes some of the high-level updates provided by Google.

Google Does a Counterattack Against Microsoft and OpenAI

There is a sparring match going on between Google and Microsoft and it’s still early in the match. Microsoft put Google on notice early this year and would have made a $10 million investment in open AI. Opening itself also put Google on notice with its GP T 3.5 and 4.0 releases.

Last week and this week Google made its counterattack by not only announcing its new Gemini LM but by enhancing the services that leverage this new large language model including Bard and all of vertex AI. However, Google also demonstrated that it is not new to this game particularly when it comes to GPU-based hardware.

The Google AI HyperComputer Powered by Google TPU v5p

In a move that hasn’t got as much press coverage as it should, Google’s new TPU 5P, which when stacked together can form a brand new AI hypercomputer that will offer many enterprises great AI capabilities at lower costs. This represents a continued innovation in AI infrastructure that puts others on notice. Note that this design is a SuperComputer that does more – hence the Hypercomputer term.

TPU Designs that Started in 2015

One of the proof points that has allowed Google to win on AI Infrastructure is that Google started work on its TPUs in 2012. This level of performance looks to rival that of Nvidia.

Gemini LLM and Multimodal (Content AI) Capabilities

Gemini LLM is the follow-on to PaLM2 and it looks very promising. Gemini Pro is already powering Bard but what is new is what is being called Multimodal. That simply means that an LLM can look at Documents (text), Images, videos, and voice. Aragon refers to those content types as Content AI.

So Gemini is Multimodal and that will make it much better at helping people in many situations.

Better Images with Imagen 2

Google is up to the task of generating photorealistic images and it got even better with the new Imagen 2. Key new capabilities:

Google also announced that it will indemnify people who use the Imagen2 service. This is significant and should make marketers more comfortable.

There were also many deep dives into new functionality, including easier ways to build with the Vertex AI Platform.

The 500 New Feature Battle Cry

In summing up all that Google has been working on, it cited the fact that there were 500 new features available in both Vertex AI and Duet AI. While the focus of these announcements was On Vertex and the Gemini LLM, we expect more news on Google Duet soon.

Bottom Line

The Bottom Line is that we are in a race to make LLMs Smarter and now with MultiModal, the ability to process and understand all forms of content. We expect more work to come on Images and videos but with Gemini Imagen2, Google now has more AI use cases to discuss with Enterprises. In the AI wars, Google has fought back effectively, courting developers more so than others. 


Get Ready for 2024 with Aragon’s 2024 Q1 Research Agenda!

Wednesday, January 17th, 2024 at 10 AM PT | 1 PM ET

Aragon Research’s 2024 Q1 Agenda

Aragon Research provides the strategic insights and advice you need to help your business navigate disruption and outperform your goals. Our research is designed to help you understand the technologies that will impact your business–using a number of trusted research methodologies that have been proven to help organizations like yours get to business outcomes faster.

On Wednesday, January 17th, 2024, join Aragon Research CEO and Lead Analyst, Jim Lundy for a complimentary webinar as they walk you through Aragon’s Q1 2024 research agenda.

This webinar will cover:

Register Here

Exit mobile version