Category: Top News

IBL News | New York

A charter school in Arizona, Unbound Academy, received official approval to offer next year a fully online experience from through to eighth grades, giving students two hours of academic instruction taught entirely by AI.

The Arizona State Board for Charter Schools greenlighted the initiative. Unbound Academy proposed in its application an “AI-driven adaptive learning technology that condenses academic instruction into a two-hour window.”

Unbound’s approach leans on edtech platforms like Khan Academy and IXL, with students engaging with “interactive, AI-powered platforms that continuously adjust to their individual learning pace and style.” Guides, rather than traditional teachers, will provide “targeted interventions and coaching” for each student.

IBL News | New York

Salesforce Inc. said it signed deals to install AI agents at more than 200 companies, including Accenture, Adecco Group, FedEx Corp., and IBM.

Experts predict that agents will become more widespread in the first quarter of 2025, reshaping workforces at companies.

Agents are AI tools that can handle multistep, repetitive tasks like onboarding clients, approving expenses and other administrative tasks, and answering customer service requests, all with minimal human supervision.

“This is really the rise of digital labor,” Salesforce CEO Marc Benioff said.

“Agents are the next giant breakthrough,” said OpenAI’s CEO Sam Altman. “An agent acts like a really smart senior coworker with whom you can collaborate on a project,” he added.

Salesforce launched its agent product, “Agentforce,” two months ago, while OpenAI is set to unveil an agent at a research preview in January.

• Bloomberg: Big Tech’s New AI Obsession: Agents That Do Your Work for You

IBL News | New York

OpenAI released Projects, a feature that allows the organization of any file and data with labeled folders. It will roll out for ChatGPT Plus, Pro, and Teams subscribers.

Until now, keeping track of conversations with ChatGPT often meant dealing with a cluttered sidebar and a lot of scrolling.

Within a Project, ChatGPT remembers specific instructions and relevant data. Users can set specific instructions for ChatGPT while working within that particular Project space.

They can also add existing chats to a Project, start new ones, and even pull in data from files mid-conversation.

The Projects feature rivals Claude, Anthropic’s AI chatbot.

Projects are beneficial for building websites. Users can upload design files and content ideas and specify their preferred coding languages. ChatGPT can then help generate code for the site and improve any design or content. Unlike the usual ChatGPT coding help, the AI will remember what you’ve already worked on together because it happens within the same Project space.

IBL News | New York

Coursera released its 2025 Job Skills Report based on data from its five million enterprise learners.

This report highlights how the rise of GenAI is transforming the learning priorities of employees, students, and job seekers.

The year-to-year increase in generative AI demand among all employees, students, and job seekers has been 866%.

The learning company reported that there have been six million engagements with Coursera Coach, its AI-based learning assistant, to practice job skills with interactive exercises.

Coursera said that, based on trends obtained through 7000+ business, government, and higher ed institutions, these are the fastest-growing skills in Business, Data Science, and Tech.

Explore the fastest-growing skills in the Job Skills Report 2025. Based on data drawn from Coursera’s 5M enterprise learners, you can identify the skill and learning trends that employees, students, and job seekers are prioritizing for career success in 2025.

Leverage data to… pic.twitter.com/b6XZGzanqY

— Coursera (@coursera) December 19, 2024

IBL News | New York

Google released its latest generative AI video generator model, Veo, in a private preview for businesses on the Vertex platform.

Launched in early December, after first being unveiled in May, Veo competes with OpenAI’s Sora.

It generates 1080p resolution videos in visual and cinematic styles from text—or image-based prompts. The length of the videos hasn’t been specified.

Google said Veo and Imagen 3 carry built-in safeguards to prevent them from generating harmful content or violating copyright protections.

Veo includes an invisible digital watermark that Google says can “decrease misinformation and misattribution concerns.”

AI-generated video content appeared in ads like Coca-Cola’s recent holiday campaign.

IBL News | New York

xAI, Elon Musk’s AI company, began to roll out an upgraded version of its Grok 2 chatbot model to all users on X this month.

The company claimed in a blog post that the enhanced Grok is “three times faster, offers improved accuracy, instruction-following, and multi-lingual capabilities.”

According to xAI, the Grok button” to X has been designed to help users discover “relevant context, understand real-time events, and dive deeper into trending discussions.”

Free users can only ask Grok 10 questions every two hours. Subscribers to X’s Premium and Premium+ plans get higher usage limits.

Pricing has been reduced from $5 per million input tokens (~750,000 words) or $15 per million output tokens to $2 per million input tokens and $10 per million output tokens.

xAI plans to bring the image generation model Aurora to the API.

IBL News | New York

An undisclosed internal agreement signed in 2023 between Microsoft and OpenAI defines what is achieving AGI. It’s based on OpenAI generating $100 billion in profits as a company. By this financial definition, OpenAI is many years away from reaching it.

It means that Microsoft won’t lose access to OpenAI’s technology soon. This disengagement between the two companies will only occur when OpenAI achieves AGI, per their agreement. It could be a decade or more.

A report from The Information disclosed these terms established in an agreement signed last year.

OpenAI is reportedly set to lose billions of dollars in 2024. Moreover, the San Francisco-based startup told investors it won’t turn a profit until 2029.

IBL News | New York

In an effort to beat OpenAI’s Sora, Google’s DeepMind research lab announced Veo 2 this month.

This is a video-generation AI model that, according to the company, can create two-minute-plus clips in resolutions up to 4k (4096 x 2160 pixels), reportedly four times the resolution and six times the duration of Sora.

VideoFX, where Veo 2 is now exclusively available, is behind a waitlist. However, Google will make Veo 2 available via its Vertex AI developer platform as the model becomes ready for use at scale.

Like Veo, Veo 2 can generate videos given a text prompt or and a reference image.

DeepMind said that Veo 2 can produce more realistic and cinematographic models with human expressions, fluid dynamics (like coffee being poured into a mug), and properties of light (such as shadows and reflections).

These are some examples shared by Google:

 

IBL News | New York

Meta announced its text-only AI model Llama 3.3 70B.

Ahmad Al-Dahle, VP of generative AI at Meta, said in a post on X that Llama 3.3 70B delivers the performance of the largest Llama model, Llama 3.1 405B, at lower cost, with improvements in areas like math, general knowledge, instruction following, and app use.

Al-Dahle published a chart showing Llama 3.3 70B outperforming Google’s Gemini 1.5 Pro, OpenAI’s GPT-4o, and Amazon’s newly released Nova Pro on several industry benchmarks, including MMLU, which evaluates a model’s ability to understand language.

Llama 3.3 70B is available for download from Hugging Face and the official Llama website.  

Meta’s latest move is to dominate the AI field with “open source models, which can be equally used and commercialized for various applications.

Meta AI’s CEO, Mark Zuckerberg, said the company has nearly 600 million monthly active users on Llama models.

However, training generative AI models is a costly business.

The company announced this month that it would build a $10 billion AI data center in Louisiana — the largest AI data center Meta has ever built.

Meta has procured a cluster of more than 100,000 Nvidia GPUs for model development, rivaling the resources of competition like xAI.

Introducing Llama 3.3 – a new 70B model that delivers the performance of our 405B model but is easier & more cost-efficient to run. By leveraging the latest advancements in post-training techniques including online preference optimization, this model improves core performance at… pic.twitter.com/6oQ7b3Yuzc

— Ahmad Al-Dahle (@Ahmad_Al_Dahle) December 6, 2024

IBL News | New York

OpenAI introduced on December the full version of its flagship reasoning model, o1, for its paid customers. This model was initially released as a limited preview in September.

The company also launched ChatGPT Pro, a new $200 monthly subscription tier with unlimited access to OpenAI o1, GPT-4o, and Advanced Voice mode.

The company also said it includes a version of o1 exclusive to Pro users “that uses more compute to provide the best possible answer to the hardest problems (called o1 pro mode).”

OpenAI continues to offer a Plus tier for $20 a month.

It also announced a ChatGPT Pro Grant Program that awards ten grants of ChatGPT Pro to medical researchers at leading institutions, with plans for additional grants across various disciplines.