A week in AI history: A showdown between Google vs. OpenAI

1. What is happening?

OpenAI organized an impromptu "Spring Update" event on Monday, during which they unveiled their latest AI model, GPT-4o, through a remarkable live demonstration. Utilizing the ChatGPT app on the iPhone, the model demonstrated capabilities such as interpreting live camera feeds, aiding in mathematical problem-solving, and facilitating real-time translation between English and Italian speakers.

In a similar move, Google also took action yesterday during its yearly developer conference, primarily centered around generative AI. In addition to showcasing several technological advancements, the company presented its perspective on search in the era of AI: Liz Reid, Google's search chief, announced: “Google will do the Googling for you,”

2. GPT-4.o and 'human-like' thrills world AI enthusiasts

2.1. What is GPT-4.o?

ChatGPT-4.o - where 'o' stands for omni - voice, text, and vision are unified into a singular model, enhancing its speed compared to its predecessor. The company stated that this new model operates twice as fast and is notably more efficient.

2.2. How GPT-4.o is "more human than ever?"

The OpenAi demo mainly featured the company's employees asking questions to voice ChatGPT, which responded with jokes and human-like banter

• Real-time interaction: GPT-4.o allows for back-and-forth conversation without waiting for the model to finish its statement
• Harmonized speed synthesis: GPT-4.o can generate different voices and even harmonize them for a more natural dialogue experience
• Sophisticated conservations: GPT-4.o offers normal conversational interactions, translations, and more, all with the sophistication expected from GPT-4.o level technology

3. How does Google catch up with GPT-4.o?

Following OpenAI's unveiling of ChatGPT4.o, Google swiftly rolled out a series of AI tools. Let's take a look at what they do?

3.1. Project Astra. What is that?

Google's Project Astra, unveiled at the Google I/O event by CEO @Demis Hassabis, promises a groundbreaking leap in AI technology. Designed to be a universal AI assistant, Astra excels in understanding context and responding instantly to user requests. In an impressive demo, the AI, equipped with a camera and microphone, demonstrated its prowess by identifying a speaker's location and detailing its components, as well as explaining lines of code on a screen. Its most striking feature? Astra's ability to remember and recall information, accurately locating misplaced glasses with just a brief scan of the room. This innovation marks a significant step towards making AI an integral part of everyday life, showcasing Google's continued leadership in AI development.

3.2. AI - Overview: Google search.

AI - Overview can bring many benefits for consumers, specifically as follows:

• Get quick answers with AI Overviews

At times, you may seek prompt responses without the luxury of assembling all the necessary information yourself. AI Overviews in search will handle the task for you.

Through our trial in Search Labs, AI Overviews have been utilized billions of times by individuals. They appreciate the ability to obtain a brief summary of a topic along with relevant links for further exploration. Our research indicates that with AI Overviews, there is an increase in search usage and higher satisfaction with search outcomes among users

• Ask your most complex questions

AI Overviews, powered by our Gemini model's advanced reasoning, can handle complex queries without the need for multiple searches. Instead of breaking down your question into several searches, you can now ask detailed questions in one go. For instance, you could search for the best yoga or pilates studios in Boston, including details on intro offers and walking distance from Beacon Hill, all in a single query.

3.3. Google Workspace.

So, what Google Workspace can help the users?

• Assisting writing, even while on the move

Duet AI is already functioning in the background within Workspace, aiding users with writing tasks, whether it involves refining existing content or initiating tasks in both Gmail and Docs. Now, this functionality is being expanded to Gmail mobile, allowing users to compose complete responses with just a few words as a prompt. Following the mobile launch, contextual assistance will be introduced, facilitating the creation of professional responses that automatically include relevant details.

• Creating unique images from text, directly within Google Slides

Integrating Duet AI into Slides aims to streamline the creation of captivating visuals for presentations. These image models have the power to generate entirely new depictions with minimal text input. For instance, if you're planning a campaign to promote safaris to Parisians, you can now easily create original visuals that reflect your creative vision, saving time and effort in the process.

3.4. Google Photo introduces an AI search feature

Google Photos is getting an AI boost with Ask Photos, an experimental feature powered by Google's Gemini AI model. This summer, users can search their photos using natural language queries, making content discovery more intuitive. This advancement was unveiled at Google's annual I/O 2024 conference. For instance, instead of searching for specific items like the "Eiffel Tower," users can ask the AI for complex requests, like finding the best photo from each National Park visited. The AI analyzes photo quality based on factors like lighting and blur, integrating geolocation or timestamps to retrieve relevant images from U.S. National Parks

3.5. VEO

At the Google I/O 2024 conference, the company introduced Veo, developed by DeepMind. Veo creates high-quality 1080p videos in various cinematic styles, extending over a minute. It uses advanced natural language understanding and image semantics to match the user's creative vision. Veo can comprehend film terms like "time-lapse" or "aerial shot." Google is working on integrating Veo's features into VideoFX for YouTube Shorts, but no release date has been announced

4. Which one will win this game?

In conclusion, the landscape of AI development is rapidly evolving, with OpenAI and Google emerging as key players in driving innovation forward. OpenAI's commitment to creating AI that is more human-like than ever before, while also striving to make it accessible to all, demonstrates a dedication to democratizing this transformative technology. On the other hand, Google's recent advancements, particularly in their Core Search program and Google Workspace, showcase their efforts to compete in various segments of the AI market and solidify their position as a leading player. Despite their different strengths, OpenAI and Google share a common vision for AI: making it seamless, intuitive, and human-like across text, voice, visual, and video platforms. As they shape AI's future, they are driving towards seamless integration into daily life, enhancing experiences and interactions in unprecedented ways.

The game is heating up with MS Conference next week, starting from May 21st. Let's guess who will win?

Follow Edtronaut and subscribe to AI & the Future of Work newsletter. ??

#Edtronut #AIinAction #OpenAI #GPT4o #google #GoogleIO2024 #FutureofWork