This Week's Most Read: DeepMind's New AI Agent System

Here are this week's most popular stories on AI Business:

1. DeepMind's New AI can Learn Tasks Directly from Humans

Google DeepMind has developed an AI agent system that can learn tasks from a human instructor. After enough time, the AI agent can not only imitate the actions of the human instructor but also recall the observed behavior.

In a paper published in Nature, researchers outline a process called cultural transmission to train the AI model without using any pre-collected human data.

This is a new form of imitative learning that the DeepMind researchers contend allows for more efficient transmission of skills to an AI. Think of it like watching a video tutorial – you watch, learn and apply the teachings as well as remember the video’s lessons later on.

This few-shot imitation process played out in a simulated environment called GoalCycle3D. There, the AI agent learned from a human demonstrator without having interacted with a human before. The results saw the agent able to perform the tasks it was shown and recall demonstrations “long after the expert has departed.”

DeepMind said the approach “paves the way for cultural evolution to play an algorithmic role in the development of artificial general intelligence.”

Read the full story

Related:AI Startup Roundup: OpenAI Buying Chips from Sam Altman-backed Startup

To get all of the latest AI news and insights, subscribe to the email newsletter.

2. Get Ready: AI is Coming for Your Job (But It Might Be a Good Thing)

In the latest column from Seth Dobrin, the founder and CEO of Qantm AI and former global chief AI officer at IBM, explore how with preparation, organizations and workers can harness AI to unlock new potential rather than simply react to disruption.

“Rapid change means today's workers have only about half a generation to reskill and transition to the AI-powered economy. Educational systems, government policymakers, corporate training programs and social safety net structures will all need to undertake rapid adaptations at an unprecedented speed to address the workforce disruption driven by advances in generative AI over the next decade.

“The condensed timeline makes urgent, coordinated action critical. Unlike past transformations centered on manual labor, the breadth of impact spanning white-collar and professional roles means no workforce segment can ignore this trend.”

Read the full column

3. Google Launches Gemini Next-Generation Large Language Model

Google has finally unveiled its fabled next-generation large language model, Gemini, claiming it’s the largest and “most capable” model the company has ever built.

Related:FICO's EVP of Software on Building an AI-enabled Platform

After months of silence, the company dropped a blog post showcasing the model, revealing Gemini comes in three sizes - Ultra, Pro and Nano, enabling it to run on mobile devices up to data centers.

Google announced that Gemini will start powering solutions today – with its Bard chatbot powered by Gemini Pro, the initial version of the model, but only in English. The company claims Gemini Pro will improve Bard’s reasoning and understanding abilities.

In early 2024, Google plans to launch Bard Advanced. No further details were given, but Google said it would provide access to its “advanced models and capabilities — starting with Gemini Ultra.”

Starting Dec. 13, developers and enterprise customers can access Gemini Pro via the Gemini API in Vertex AI or Google AI Studio, its free web-based developer tool.

Read the full story

4. Create Unique Sounds with Meta's New Audiobox AI

Facebook parent Meta has unveiled its latest audio generation AI model called Audiobox that lets users turn text into sound.

Simply describe what you want to hear and the text-to-audio model will turn it into the sound you requested. The successor to its Voicebox audio generation model, Audiobox takes prompts in natural language.

For example, you can type ‘a beaver munching on a slice of pineapple’ or ‘a young woman talking inside a church’ and the model will generate the desired sound.

You can listen to some AudioBox audio samples on Meta’s research website.

Audiobox can also handle audio inputs – so users can combine a voice input plus a text prompt to better synthesize the audio. This lets users specify the style of speech and sound effects they want generated – a feature not found in the prior version of the model. “When a voice input and text prompt are used together, the voice input anchors the timbre, and the text prompt can be used to change other aspects,” according to Meta.

Read the full story

5. OpenAI Buying Chips from Sam Altman-backed Startup

In this week’s AI Startup Roundup, it emerged OpenAI has a $51 million deal in place to purchase AI chips from a startup that CEO Sam Altman has personally invested in.

In 2019, OpenAI penned a deal with Rain AI, a San Francisco startup building chips designed to replicate the way the human brain processes information. Altman led a Rain seed round a year earlier.

Last month, Altman was fired and re-hired within five days for not being “consistently candid” with OpenAI’s nonprofit board. No other details were released and the issue under internal investigation.

Under the deal, Rain would provide OpenAI with its neuromorphic processing units, or NPUs. Wired saw investor documents from Rain that said the startup is hoping to ship its first batch of hardware in October 2024.

Read the full story