Superpower Daily
Posts
BREAKING: “The most ambitious product Apple has ever created”

BREAKING: “The most ambitious product Apple has ever created”

Open-source alternative to ChatGPT with web search

Saeed Ezzati
June 06, 2023

In today’s email:

🚨 Today, Apple starts a new chapter. NEW Mixed Reality Headset
🔥 open-source alternative to ChatGPT with search capability
📚 Artificial Intelligence for Beginners - A Curriculum by Microsoft
🛠 Various AI-related tools and platforms, including Albus, Luminous, Artifact, Dumme, Brainworm, localGPT, and more.

Highlights💡

SAM-HQ - Segment Anything in High Quality [GitHub], [Paper]

SAM-HQ - Segment Anything in High Quality [GitHub], [Paper]

The Segment Anything Model (SAM) is a significant advancement in segmentation models, enabling robust zero-shot abilities and adaptable prompting.
Despite being trained with 1.1 billion masks, SAM often fails to provide high-quality mask prediction, especially for intricate objects.
The newly proposed HQ-SAM model improves SAM's ability to accurately segment any object while preserving its original features like promptable design, efficiency, and zero-shot generalizability.
HQ-SAM maintains the pre-trained model weights of SAM, introducing only minimal additional parameters and computation, ensuring efficiency.
A unique component of HQ-SAM is a learnable High-Quality Output Token, which is integrated into SAM's mask decoder for predicting high-quality masks.
The High-Quality Output Token is fused with early and final ViT features for enhanced mask detail before applying it on mask-decoder features.
To train the newly introduced parameters, a dataset composed of 44K fine-grained masks from various sources is used. Training on this dataset with HQ-SAM takes only 4 hours on 8 GPUs.
HQ-SAM's effectiveness has been tested across nine diverse segmentation datasets covering different downstream tasks. It performed successfully in seven out of these datasets following a zero-shot transfer protocol.

HuggingChat, the 100% open-source alternative to ChatGPT by HuggingFace just added a web search feature [Link], [GitHub]

Apple Vision Pro and everything else from Apple WWDC23 announcement [Link]

“One of the coolest results involved predicting a user was going to click on something before they actually did. That was a ton of work and something I’m proud of. Your pupil reacts before you click in part because you expect something will happen after you click. So you can create biofeedback with a user's brain by monitoring their eye behavior and redesigning the UI in real time to create more of this anticipatory pupil response. It’s a crude brain-computer interface via the eyes, but very cool”

Throughout the whole event, Apple mentioned AI exactly ZERO times. They used “machine learning” instead. Here’s a quick recap of all the new things announced:

And you can watch the full (2 hours) announcement here, and learn more about the new Apple Vision Pro product here

Watch people’s reactions live at the Apple event when they announced the Vision Pro price at $3499 🤯

You won’t find a better pre-sale indicator than this sound from the audience
— Ramp Capital Guy (@RampCapitalLLC)
9:36 PM • Jun 5, 2023

What does Palmer Luckey, the founder of Oculus VR and designer of the Oculus Rift, think about Apple Vision Pro

My official thoughts on Apple Vision Pro head mounted display, all press outlets please use this as my comment:
— Palmer Luckey (@PalmerLuckey)
1:55 AM • Jun 6, 2023

Microsoft has launched a FREE course on AI for beginners 🔥 [Link]

It’s a 12-week, 24-lesson curriculum all about Artificial Intelligence.

In this curriculum, you will learn:

Different approaches to Artificial Intelligence, including the "good old" symbolic approach with Knowledge Representation and Reasoning (GOFAI).
Neural Networks and Deep Learning, which are at the core of modern AI. We will illustrate the concepts behind these important topics using code in two of the most popular frameworks - TensorFlow and PyTorch.
Neural Architectures for working with images and text. We will cover recent models but may lack a little bit on the state-of-the-art.
Less popular AI approaches, such as Genetic Algorithms and Multi-Agent Systems.