Apple Car coming 2028

Welcome to Daily Zaps — Here’s what we got for you today: 🙌

Apple Car coming 2028

Apple's plan to make an electric self-driving car has been pushed back to 2028, which is later than they first thought. Initially, Apple aimed to develop a fully autonomous (SAE Level 5) vehicle but has scaled back to a more achievable "Level 2+" autonomy, an informal category that suggests enhanced features compared to standard Level 2 systems.

This revised plan also includes the possibility of future upgrades to Level 4 autonomy. The shift in strategy reflects the challenges Apple faces in transitioning from consumer electronics to the complex automotive sector. Despite these setbacks, the company's vision for the car includes a sleek design, advanced safety systems, and a unique user interface, with aspirations to compete in certain aspects with established players like Tesla.

Google’s text-to-video generation tool

Google, along with researchers from Weizmann Institute and Tel Aviv University, has created Lumiere, a new model for making realistic videos. Unlike current leaders in AI video like Runway and Pika, Lumiere uses a special method that makes videos look more real and smooth. You can tell it what kind of video you want using just text, or turn a photo into a moving video.

Lumiere is better than other AI video tools because it can handle the whole video at once, making everything flow better. It's trained on a huge number of videos and does a great job, but it's not ready for everyone to try yet. It also can't do videos with different scenes or transitions, but that's something they're still working on.

From Our Sponsor

Xembly – Your company’s new AI executive assistant

• Need to schedule a meeting with 7+ people?

• Need to take meeting notes while staying engaged?

• Need to manage your deliverables and deadlines?

Meet Xena by Xembly, your company’s new AI executive assistant that schedules meetings, takes notes, and manages tasks for you — giving you 400 hours extra a year.

Boost your team’s productivity with Xembly today (free for a limited time only).

Microsoft’s new “GenAI” team

Microsoft is creating a new AI team, GenAI, to develop artificial intelligence that's smaller and uses less computing power than OpenAI's models. This team, formed by moving top AI developers from Microsoft's research group, aims to reduce costs by building small language models (SLMs) that work like OpenAI's GPT-4 but require less power.

These SLMs, like the Phi model Microsoft developed, can run on mobile devices and handle tasks similar to GPT-4. Microsoft plans to integrate these efficient AI models into products like Office and Azure, offering them to customers and app developers. Microsoft's CEO Satya Nadella highlighted the importance of these small AI models in controlling the company's future in AI development.

A.I. or human: can you tell?

Recent studies have shown that people often find AI-generated images, particularly of white faces, to be more realistic than actual photographs. This phenomenon, known as hyper-realism, occurs because AI tools like Dall-E and Midjourney were predominantly trained on images of white people, leading to highly realistic depictions of white faces.

In contrast, the accuracy drops when identifying nonwhite AI-generated faces. Interestingly, the studies revealed that participants were often overly confident in their ability to distinguish real faces from AI-generated ones, but this confidence was frequently misplaced.

Take the quiz and see how you do.