EN

KR

UPP News: What's new in the AI field? (Part 2)

UPP News: What's new in the AI field? (Part 2)

Date

December 8th, 2023

Reading Time

10 mins

Google’s next-gen AI model Gemini outperforms GPT-4

"After more than eight months of research and development, Google has unveiled its most powerful A.I. so far.

For more than a year, Google has raced to build technology that could match ChatGPT, the eye-opening chatbot offered by the San Francisco artificial intelligence start-up OpenAI.

Bard is now powered by Google's new Gemini model, which, according to Google's benchmarking, not only matches but surpasses OpenAI's technology in various aspects.

Gemini is what scientists call a large language model, or L.L.M., a sophisticated mathematical system capable of acquiring skills through the analysis of extensive data. By identifying patterns in all that text, an L.L.M. learns to generate text on its own. That means it can write term papers, generate computer code and even carry on a conversation.

Bard is now running Gemini Pro, the middle tier of the Gemini series. Ultra is the biggest and slowest but the most capable, Nano is small and fast and meant for on-device tasks, and Pro sits right in the middle.

“I think people are just going to find that the product got a lot better,” Sundar Pichai, Google's CEO says. “It understands their intent better, it’s answering better. It’s more factual, higher quality. If you’re trying to code it’s better!”

Currently, Bard is still just a chatbot: you type, it types back. But a new version of Bard is coming soon that could be much more. Next year, Google is planning to launch a preview of “Bard Advanced,” powered by Gemini Ultra, which is the most powerful and capable version of Google’s new large language model. Gemini Ultra is also the multimodal version of the model, meaning it can accept and create images, audio, and video in addition to just text.

Pichai says he thinks of this launch both as a big moment for Bard and as the very beginning of the Gemini era. But if Google’s benchmarking is right, the new model might make Bard as good a chatbot as ChatGPT.

VinAI launches PhoGPT: A Vietnamese challenger to ChatGPT

"At the AI Day 2023 event on December 5, VinAI, a member of Vingroup, unveiled the groundbreaking PhoGPT, an open-source, large language model designed specifically for Vietnamese users.

"PhoGPT aims to mirror the capabilities of ChatGPT, but with a focus on Vietnamese language and culture," said Dr. Bui Hai Hung, CEO of VinAI.

According to the representative of the unit, the project aims to develop models similar to ChatGPT for the Vietnamese language and culture. PhoGPT can understand and generate written content in Vietnamese. The model is also trained from scratch using Vietnamese datasets, without relying on other world models.

"This model, built entirely on Vietnamese datasets, represents a major step towards Vietnam mastering advanced core technologies independently,"" the CEO added.

The PhoGPT development team is also continuing to improve the model and will expand the project to target other languages in the Southeast Asia region.

In comparative evaluations, PhoGPT-7B5-Instruct has shown remarkable performance, ranking just below the closed-source ChatGPT (GPT-3.5-turbo) in most metrics.

During the AI Day event, Hung showcased PhoGPT's versatile functionalities, including its ability to compose poetry and answer questions. He underscored the model's open-source nature, stating, "PhoGPT is available for diverse applications, including commercial purposes. It stands as a flexible platform for the burgeoning AI application development sector in Vietnam.

Bing’s GPT-4-powered Deep Search takes its time with AI questions

Deep Search, powered by GPT-4, introduces a new capability for users to generate and explore various topics. This feature enhances the Bing search experience by expanding on user queries and providing comprehensive information on related subjects from the web.

When you input a query into Bing, the GPT-4-driven Deep Search goes beyond a standard search, presenting a panel on the search results page with diverse topics related to your question. Users can then select the topic that aligns best with their inquiry, and Deep Search refines its search results by ranking them based on how well they correspond to the comprehensive description.

Microsoft emphasizes that conducting a Deep Search might take up to 30 seconds, acknowledging the potential for a longer processing time. However, for those who prefer a quicker search experience, Bing's standard search feature remains readily available.

Currently, Deep Search is in the experimental phase and is accessible to a limited group of users. Despite its experimental status, this innovative feature holds promise in streamlining the search process, especially when formulating prompts for AI can be a task in itself, potentially hindering productivity.

Bringing the Fun Back to Surveys: Trove AI's Conversational Approach

Gone are the days of monotonous, lengthy surveys with low response rates and poor data quality. Trove AI emerges as a revolutionary force, aiming to revitalize the survey experience through engaging conversation. Utilizing the power of GPT-4 and their own refined models, Trove AI crafts surveys that feel more like natural dialogue, making them enjoyable and attractive to participants.

Within a short six weeks of beta testing, Trove AI has already amassed a dedicated user base of over 1,000, primarily composed of small and medium-sized businesses. This early success underscores the growing demand for a more engaging survey solution. Fuelled by pre-seed funding from prominent figures like Zayd Enam and Tim Shi, Trove AI is poised for further growth and impact.

While established giants like Qualtrics and SurveyMonkey dominate the survey landscape, Trove AI stands out with its unique AI-powered approach. By prioritizing conversation over traditional question formats, Trove AI aims to unlock deeper insights and more authentic responses from participants.

Looking ahead, Trove AI envisions itself becoming a comprehensive platform for managing customer and employee experiences, catering to companies of all sizes. Their focus on SaaS (80%) and AI (20%) allows them to strike a balance between user-friendly accessibility and cutting-edge technology. Additionally, Trove AI is actively reducing its reliance on third-party APIs, fostering greater independence and control over their platform.

The rise of Trove AI signifies a paradigm shift in the survey landscape. Conversational AI promises to revolutionize the way we gather feedback, making it more engaging, informative, and ultimately, fun. This shift marks an exciting development for businesses of all sizes, as they gain access to richer data and deeper customer understanding, ultimately paving the way for better decision-making and improved experiences.

Newsletter

DISCOVER MORE

LET’S TALK...

Content delivered to your inbox

ENTER YOUR EMAIL

YOU WANT TO...

Subscribe
KSA Cloud
ISO 9001:2015
ISO 27001:2022

Hanoi, Vietnam

Web3 Tower, No. 15, Alley 4, Duy Tan, Cau Giay, Hanoi, Vietnam

© 2025 UPP Global Technology JSC

Look up for solutions? Look for UPP!

PRIVACY POLICY