This Week in AI: Big new from Apple, Meta & DeepMind

TOP HEADLINES

• Brainstorming AI workers could take over workplaces in 2025, Capgemini reports.

• AI is already taking jobs in the video game industry.

• OpenAI reassigns the leader of its Preparedness team to another “bigger” project.

• Meta's Q2 earnings showed a 22% revenue increase due to AI investments, while Amazon projected lower-than-expected profits due to increased AI development spending.

• How Elicit evaluates LLM outputs - The tradeoffs between trust, scale, and flexibility.

• The European AI Act officially came into force on August 1, 2024, setting clear requirements and a risk-based approach for AI systems in the EU.

• Meta announced that India has become its largest market for AI usage, particularly on WhatsApp, with 500 million users.

LATEST TUTORIALS

We are on track to release over 200 tutorials over the next 90 days! Here are the latest tutorials we released.

• How to Transform Long-Form Content into Engaging Social Media Posts with AI

• Creating Powerful Product Comparison Sales Documents with ChatGPT

• Creating Effective Product Roadmaps with AI

• Creating an Effective Creative Brief Using ChatGPT

• Introducing Claude: A Powerful AI Assistant That Can Transform Your Productivity

• Elevate Your Interactions with Claude

→ SEE ALL TUTORIALS

COMMUNITY UPDATES

• Meet in Person at the Chatbot Conference in San Francisco

• Join us on Discord: Jobs, Events & Community

• Submit your AI Art Here

AI Image of the Day

IN THE NEWS

Apple AI Wont be Ready in September

Apple's upcoming AI features, known as Apple Intelligence, will not be available when the next iPhone is released in September.

Instead, anticipate these features to be released by October.

This delay allows Apple to take additional time to address any bugs and ensure a seamless user experience. However, it also means that the highly anticipated feature of the next iPhones will not be available from the start.

Developers will have the opportunity to preview Apple Intelligence with the upcoming iOS 18.1 and iPadOS 18.1 betas, allowing them to test its features in advance.

Here's what you need to know:

• AI features will be available in October, separate from the initial releases of iOS 18 and iPadOS 18.

• Get access through iOS 18.1 and iPadOS 18.1.

• The full AI rollout is set to continue throughout late 2024 and 2025, with exciting enhancements planned for Siri.

Meta Releases Lama 3.1

How is Llama 3.1 different from 3? A lot went into the name change. Meta's model has 405 billion parameters, up from 70 billion.

Suddenly powerful, Llama beat GPT-4o and Claude 3.5 in benchmarks after the update. First open-source “frontier model” release boosts AI performance. Llama's 8b and 70b work better.

What distinguishes Meta's model? Alphabet, Anthropic, and OpenAI want sole LLM control. Meta thinks customers and businesses want customizable models. It could result in millions of cloud-based LLMs with specific uses.

Its open source model lets developers change its code, which is the biggest difference. Optimized models can outperform base models with enough changes. Independent developers are thrilled to get the latest model for free.

Meta is creating an ecosystem for Nvidia, Amazon, smaller companies, and bootstrapped developers to collaborate on new projects.

Zuckerberg expects Llama's performance boost and new features like a 'Imagine Me' selfie tool to make it more popular than OpenAI by year's end. ChatGPT's website visitors have dropped in recent months, so that's possible. GPT-5 may derail Zuckerberg's prediction.

TRENDING TOOLS

• Websim.AI: With simple prompt, AI can generate a website, an application or even a game!

• GumLoop: Let’s you create an army of AI workers that can handle different parts of a project.

• Create Your own AI Clone: Meta now allows you to create AI Characters and Digital Clones of yourself on Facebook and Instagram.

• Gen-3 by Runway: Runway AI Generated videos just set the world on fire. If you haven’t see it, check it out below.

WIERD WORLD OF AI

ELON SHARES FAKE VIDEO ON X

A video with an AI-generated voice of Vice President Kamala Harris has sparked concerns as the US election draws near.

The video, shared by Elon Musk on his social media platform X, imitates Harris's voice and attributes statements to her that she never made, without clearly indicating that it was intended as a parody.

Utilizing visuals from an authentic Harris campaign ad, the video cleverly replaces the voice-over with an AI impersonation, putting words in her mouth.

Harris's campaign strongly criticized the video, highlighting the significance of honesty.

Was this parody or propaganda?

Even though the original user stated that the video was a parody, Musk's post, which has been viewed over 123M times, failed to provide any clarification on this matter.

Specialists have verified that the audio in the video was created by artificial intelligence, showcasing the impressive capabilities of generative AI and deepfakes.

Some worry that the video's comedic nature could mislead many viewers.

In other cases, AI deepfakes have been employed to sway voters through the spread of false information and comedic content.

Can we please put an end to the AI drama for the remainder of the campaign?

AROUND THE WEB

@stefan.speaks

One of these videos is a real person the other is an AI deep fake. Can you tell which one is real? #aivideo #aivideos #texttovideo #ai #pe... See more

Do you Want to Reach a Bigger Audience?

🚀 Advertise in our newsletter to reach 120,000+ founders, executives, designers and engineers.

Let us know how we are doing

Remember to leave some feedback for us! We love hearing what you thought.

❤️ We need your feedback to become even better.