Crazy week in AI: Meta, OpenAI, Google and Figure raising up...

?? Meta Expands AI Studio for U.S. Creators to Build Custom Chatbots

Meta is rolling out AI Studio, a platform enabling users to create, share, and discover AI chatbots without any technical skills required. This innovative tool lets anyone build personalized AI characters designed to entertain, advise, and engage.

Key Features of AI Studio

• Custom AI Characters: Users can create AI characters for various purposes, including generating memes, giving travel advice, and answering common questions.
• Creator Integration: Creators can build AI extensions of themselves to interact with their audience, enhancing fan engagement.
• Wide Availability: AI Studio is accessible across Meta platforms, including Instagram, Messenger, WhatsApp, and the web.

Getting Started with AI Studio AI Studio is now available in the US. It allows users to create AI characters via Meta Website or directly within the Instagram app. Users can customize their AI's name, personality, tone, avatar, and tagline using a variety of templates or by starting from scratch.

Examples of AI Characters:

• Eat Like You Live There!: Chef Marc Murphy's AI offers personalized local dining tips.
• What Lens Bro: Angel Barclay’s AI provides photography advice.
• Flip Pawsitive Affirmation Dog: Rocky Kanaka's AI gives positive affirmations.
• Sammy The Stress Ball: Created by Assistants vs. Agents, this AI helps manage stress.

Enhancing Fan Engagement Creators can use AI Studio to set up AI versions of themselves, quickly responding to common questions and story replies. These AI assistants can share facts, links to favorite brands, and past videos, reaching more fans efficiently.

Transparency and Control Creators have full control over their AI’s responses through the professional dashboard in the Instagram app. Responses from these AIs are clearly labeled to maintain transparency with fans.

Creators like Chris Ashley, Violet Benson, Don Allen, and Kane Kallaway have already launched their AIs. Fans can interact with these AI characters by tapping the Message button with the icon on their Instagram profiles (accessible via mobile only).

Safety and Responsibility Meta has implemented policies and protections to ensure the safe and responsible use of AI, making AI interactions fun and helpful.

With AI Studio, Meta is taking the first steps towards enabling everyone to harness AI's creative capabilities. This launch marks just the beginning of an exciting journey in AI innovation.

??Google ?????????????????? ?????? ?????????????? ???? ?????? ?? ?????????????? (????) ?????????????????? ?????????????? ???? ?????????? ??, ?????????? ?????????????????????? ??????-??.?? ?????? ?????????? ?? ??????

Firstly launched earlier this year, Gemma is a family of lightweight models designed specifically for developers and researchers.

Built on the tech behind Google Gemini, the 2B model leverages distillation from larger models to deliver impressive results.

?????? ?????? ???????? ????????? Gemma 2 2B is accessible even on Google Colab's free tier, bringing cutting-edge AI to everyone!

Google also introduced ShieldGemma, a powerful suite of safety classifiers. This system is designed to detect and remove harmful content, focusing on hate speech, harassment, sexually explicit, and dangerous content, ensuring a safer AI experience for all.

OpenAI is rolling out Advanced Voice Mode to a select group of ChatGPT Plus users, enabling natural, real-time conversations with emotional responsiveness.

This feature will expand to all Plus users by fall, with video and screen sharing coming later.

Safety measures include testing with 100+ external red teamers across 45 languages, limiting the model to four preset voices, and blocking violent or copyrighted content.

A detailed report on GPT-4o’s capabilities and safety will be released in early August.

?? Intel Accelerates AI Everywhere ??

Top Highlights:

1?? Launch of Intel? Xeon? 6 processors with Efficient-cores for high-density, scale-out workloads.

2?? Introduction of Lunar Lake client processor architecture for AI PCs, delivering up to 40% lower SoC power.

3?? Intel? Gaudi? 3 accelerators deliver significant performance improvements and cost savings for AI training and inference.??Learn more here https://intel.ly/3X2FzH1

??Reddit Takes a Stand! “No more free data for AI usage, says CEO Steve Huffman.

Reddit has updated its policies to prevent AI developers from scraping its user data without consent or payment. CEO Steve Huffman announced that Reddit will block companies like Microsoft from using data from its platform until they agree to pay and explain how the content will be used.

Huffman emphasized that Reddit will not tolerate the unlicensed use of its data to train AI models. He accused Microsoft of using Reddit’s data to train its AI and summarize content in Bing results without permission. Huffman also stated that Reddit’s data has been sold through the Bing API to other search engines.

Blocking Data Scraping

In response to these issues, Reddit has updated a web standard to block automated data scraping. This includes rate-limiting and blocking unknown bots and crawlers.

What do these techniques mean?

• Rate-Limiting: Reddit will control the number of requests from a single entity to prevent excessive data scraping.
• Blocking Bots: Unknown bots and crawlers will be blocked from extracting data.

AI Companies' Response

Microsoft, in response to Reddit’s accusations, claimed that it respects the robots.txt protocol and ceased crawling Reddit on July 1, 2024. Huffman also mentioned other AI companies like Anthropic and Perplexity, accusing them of using internet content for free and refusing to negotiate payment for Reddit’s data. In contrast, Reddit has secured a $60 million deal with Google and a similar agreement with OpenAI in May.

Tools and Protocols

Reddit plans to use the Robots Exclusion Protocol (robots.txt) to identify and block web crawlers extracting data without permission. This move aims to prevent AI firms from using data to create summaries without giving credit or asking for permission.

This move from Reddit’s is a strong statement to highlight the need for transparency and fair compensation in the use of online content.

At Generative AI, we’re at the forefront of this innovation, building an AI-native learning platform that prioritizes speed, accuracy, and ultra-personalization.

Join us as we shape the future of AI-driven education.

??A Friendship Day Gift from AI: Meet Friend - An AI Pendant cum Companion

Say ‘HI’ to "Friend," an AI companion designed to provide more than just digital assistance. Initially named Tab, Friend has been in development for several years and aims to offer a unique form of interaction with AI, and offer a means of companionship in loneliness.

Originally conceived as a productivity tool, Friend’s purpose has shifted significantly. Founder Schiffmann has moved away from traditional task-oriented AIs, finding them restrictive. He critiques similar products like Microsoft’s Recall and Humane’s AI Pin, emphasising that competing with giants like Apple or OpenAI in building highly functional AIs is not his goal.

What Makes Friend AI Unique?

• Friend has got your back: Friend is designed to be a supportive and validating companion. It can brainstorm ideas, discuss relationships, and provide encouragement.
• It’s not just an app, but a fun pendant: The device is a small, glowing orb that can be worn as a pendant or clipped onto clothing. It features a built-in microphone for ambient recording or direct conversation, with future plans to include a camera. Communication is primarily text-based through the Friend app.

The makers of Friend AI highlight the success of AI platforms like Character AI and Replika, which have fostered meaningful relationships between users and their AI counterparts. Unlike these session-based services, Friend is designed to be a constant presence, offering companionship without the need for a phone or keyboard.

Friend represents a shift in how we perceive and interact with AI. By prioritizing companionship over productivity, Friend hopes to create a product that offers genuine support and interaction.?

However, there is an undeniable threat of privacy for users and those users interact with!

As the development of Friend progresses, it’s also interesting to wonder what the humour setting on this one will be! ??

NVIDIA Supercharges Enterprise AI with Llama 3.1 and AI Foundry!

Introducing NVIDIA AI Foundry:

A comprehensive service for building custom "supermodels" using Llama 3.1 and NVIDIA's cutting-edge tech.

New NVIDIA NIM inference microservices for Llama 3.1: Deploy models with up to 2.5x higher throughput! Accenture leads the charge: First to use AI Foundry to build custom Llama 3.1 models for clients. Learn more here

?? UK Hospitals Begin Live Trial of AI for Prostate Cancer Detection ??

Three hospital systems across England have launched a live clinical trial of advanced AI technology designed to detect and grade prostate cancer. The study, named ARTICULATE PRO, is being led by the University of Oxford in collaboration with Paige, a pioneer in clinical AI applications for cancer diagnosis.

??? Participating Hospitals and Trial Focus

The participating hospitals include:

- North Bristol Trust Southmead Hospital

- University Hospitals Coventry and Warwickshire

- Oxford University NHS Foundation Trust

These institutions are integrating Paige's AI technology, the Paige Prostate Suite, into their standard care procedures. The trial aims to evaluate the potential of AI in enhancing patient outcomes amidst rising prostate cancer cases.

??? Clinical Impact and Expert Insights

Professor Clare Verrill, OUH Cellular Pathology Consultant and Principal Investigator of ARTICULATE PRO, emphasized the trial's patient-centric focus:

> "The central focus of ARTICULATE PRO is patients. We are striving towards our goal to safely and effectively ensure they benefit the most from powerful AI technology."

The Paige Prostate Suite is designed to assist pathologists in detecting, grading, and measuring tumors in prostate biopsies and tissue samples. This AI technology is being evaluated for its impact on clinical decision-making, pathology service delivery, and resource utilization in real-world clinical settings.

Dr. Jon Oxley, Uropathologist and Bristol lead of ARTICULATE PRO, expressed optimism about the technology's potential:

> "Paige’s AI applications have achieved a level of validation and performance that allows safe and effective live clinical use. Using Paige Prostate Suite alongside our standard of care has the promise to increase efficiency and improve reproducibility of results for patients."

??? Broad Implementation and Research Goals

The study's implementation across hospitals with different digital pathology scanners and information systems enables a diverse assessment of Paige's AI technology. This diversity is crucial in understanding how the AI can benefit patients, histopathologists, and hospital systems in prostate cancer diagnosis.

Dr. Bidisa Sinha, Uropathologist at University Hospitals Coventry and Warwickshire, highlighted:

> "We believe AI can help to improve the accuracy and consistency of grading cancer and assist in detecting small areas of cancer which are easy to miss. This is world-leading research being carried out at UHCW. We are proud to be a global leader in the field of digital and computational pathology."

??? Funding and Future Implications

The ARTICULATE PRO study is funded by the Accelerated Access Collaborative (AAC) Artificial Intelligence in Health and Care Award, under the Department of Health and Social Care. As prostate cancer rates continue to rise, the integration of AI in diagnostics could lead to earlier detection, more accurate grading, and ultimately improved patient outcomes.

The results of this trial could pave the way for broader adoption of AI in cancer diagnosis across the UK and beyond, marking a significant advancement in medical technology and patient care.

??Meta Releases Segment Anything Model (SAM) 2

Meta’s Segment Anything Model 2 (SAM 2) sets a new benchmark in real-time video object tracking and segmentation, advancing video AI capabilities significantly. SAM 2 expands Meta's image segmentation to video, overcoming issues like fast movements and occlusions.

Details:

• Extended Capabilities: Extends Meta's previous image segmentation to video, handling challenges like fast movement and object occlusion.
• User-Friendly: Create cutouts in a few clicks, with a free demo available.
• Open Source: Meta is open-sourcing SAM 2 and releasing a large, annotated database of 50,000 videos.
• Versatile Applications: Useful in video editing, mixed reality, and scientific research.

Prompt Flexibility: Two ways to run SAM 2:

1. Automatic Mask Generator: Segments all objects in an image or video and generates corresponding masks.
2. Point Prompt: Allows specific segmentation using a point or series of points, applicable to both images and videos.

SAM 2 simplifies intricate video editing, making tasks like object removal or replacement as easy as a click. This, along with Llama 3.1, exemplifies Meta's commitment to pioneering AI advancements. By making these tools open and free, Meta encourages widespread AI use and innovation across various industries. This democratization of AI technology paves the way for groundbreaking applications in science, medicine, and beyond.

??AI Reprograms Cancer Cells To Turn Against Themselves?

Scientists at USC, backed by the NIH, have used AI to reprogram glioblastoma cells into immune-activating dendritic cells (DCs) in mouse models, potentially revolutionizing treatment for this aggressive brain cancer.

The details:

• Glioblastoma is the deadliest adult brain cancer, with less than 10% of patients surviving five years post-diagnosis.
• The AI identified genes that can convert glioblastoma cells into DCs, which sample cancer antigens and activate immune cells to attack the tumor.
• In mouse models, combining this approach with immune checkpoint therapy boosted survival chances by up to 75%.
• The study, led by the Keck School of Medicine at USC, also identified human genes for potential clinical trials.

Turning cancer cells against themselves offers a novel way to fight tumors from within. If the 75% increased survival in mice translates to humans, this could make glioblastoma treatment much more effective. This research also opens doors for similar approaches in other hard-to-treat cancers, leveraging AI's power to reprogram cells, providing new hope for effective treatments.

PROMPT OF THE WEEK

Commercial Food Photography

#insert food type, commercial photography, high resolution, studio lighting, expert styling and prop placement, #insert mood/tone/style, focus on the indulgent texture --s 50 --q 2 --v 4