SIGN UP HERE

Welcome, AI RESEARCHERS.

Apple joins the data center race, Microsoft unveils a ChatGPT rival, and AI is outsmarting us in ethics! This week's AI news is WILD.

Here're everything going right now:

In today’s AI Researchs:

• Apple Enters the AI Data Center Race

• Microsoft Unveils New AI Rival

• AI Outwits Humans in Ethical Reasoning

• Researchers Design AI to Prevent Diabetic Health Crisis

• AI's Deception: Teachers Blindsided by Student-Quality Writing

• 8 AI video & audio tools

• More Ai hits & news & Researchs

Read time: 5 minutes

APPLE'S AI CHIP INDUSTRY

🚨 Apple Enters the AI Data Center Race

AI Researchs: Apple is reportedly developing its own AI-focused chips for data centers, potentially enhancing its upcoming AI features and gaining a competitive advantage.

For more details:

• Apple's "ACDC" project focuses on creating custom AI chips specifically designed for data center servers.

• These chips may power new generative AI-based features expected in iOS 18 and future software updates.

• Apple's extensive experience in chip design for iPhones, iPads, and Macs could prove crucial in this new endeavor.

• This effort aligns with Apple's history in chip development and recent focus on artificial intelligence.

• The move mirrors similar strategies by tech giants like Amazon, Google, Microsoft, and Meta, who are also investing in custom AI processing hardware.

Why this is important: Apple's entry into AI chip development reflects the increasing importance of owning and controlling AI infrastructure across the tech industry. This move could give Apple a competitive edge in the rapidly evolving AI landscape and reduce its reliance on external chip suppliers.

MICROSOFT'S LLM

⚡ Microsoft Unveils New AI Rival to ChatGPT and Google Search

AI Researchs: Microsoft enters the AI race with a massive new language model, MAI-1, poised to disrupt Google and OpenAI dominance. Led by ex-DeepMind head, this move signals Microsoft's commitment to the generative AI revolution.

For more details:

• MAI-1 is reported to have 500 billion parameters, making it significantly larger than Microsoft's previous AI models. This increase in scale implies greater potential for complex tasks and nuanced responses.

• In comparison, OpenAI's GPT-4 is rumored to have 1 trillion parameters, emphasizing that Microsoft is closing the capability gap between its AI models and the industry leaders.

• The development of MAI-1 necessitates large clusters of servers with specialized hardware like Nvidia GPUs, reflecting the computational demands of training such a large language model.

• Microsoft's investment in OpenAI, in addition to its dedicated AI unit and strategic hires, demonstrates a substantial financial and intellectual commitment to staying competitive in the generative AI field.

Why this is important: Microsoft's move fuels the generative AI arms race, potentially reshaping search engines and the way we interact with technology. This development underscores the importance of AI in the tech landscape and heralds further innovation and competition between major players.

MICROSOFT

👁️‍🗨️ AI Outwits Humans in Ethical Reasoning

AI Researchs: In a fascinating turn of events, a recent study reveals that people may find AI-generated answers to moral dilemmas superior to those offered by humans. This finding raises questions about the evolving role of AI in decision-making and our future trust in technology.

For more details:

• The modified Turing test involved participants unknowingly evaluating both human and AI-crafted responses to ethical questions. AI's answers received consistently higher ratings in terms of virtue, intelligence, and trustworthiness.

• This performance suggests AI has the potential to outperform humans in complex ethical reasoning, potentially passing a "moral Turing test."

• The study highlights a possible shift in how we perceive AI. People might increasingly defer to AI for guidance, even in morally complex situations.

• Results emphasize the necessity of thoroughly understanding the capabilities and limitations of AI systems, particularly when it comes to matters of ethics and decision-making.

Why this is important: This study indicates that our trust in AI's judgment might eventually surpass trust in our fellow humans, especially in ethically charged situations. As AI becomes increasingly sophisticated, it's crucial to examine its influence on our decision-making processes and proactively address any potential risks.

AI IN HEALTHCARE

🌟 Hong Kong Researchers Design AI to Prevent Diabetic Health Crisis

AI Researchs: Researchers in Hong Kong are making strides in diabetes treatment. A new AI model can now predict the risk of severe hypoglycemia in older diabetic patients, a breakthrough for preventative care.

For more details:

• Using hospital records of over 360,000 patients, researchers built an AI model predicting severe hypoglycemia risk.

• The model helps doctors personalize insulin treatment and medication choices for individuals.

• Hypoglycemia poses serious risks, including falls, cardiovascular disease, dementia, and increased mortality.

• This AI has the potential to be integrated into electronic health records for widespread preventative use.

Why this is important: This development signals a move towards proactive, data-driven diabetes care, which could save lives and greatly improve the quality of life for patients with diabetes, a widespread health concern.

AI AT EDUCATION

🧑‍🏫 AI's Deception: Teachers Blindsided by Student-Quality Writing

AI Researchs: Distinguishing AI-generated content just got harder. Recent studies expose the startling inability of teachers to accurately identify student essays written by AI, raising serious concerns about assessment integrity.

For more details:

• German university studies found teachers had alarmingly low success rates in identifying AI-generated essays: novice teachers correctly identified only 45.1%, and experienced teachers a mere 37.8%.

• Teachers in the study often mistook AI-written work for genuine student output, demonstrating the sophistication of current AI writing tools.

• AI detection software presents its own pitfalls, frequently flagging human-written content as suspicious and further muddying the waters.

• This trend highlights the overconfidence educators may have in spotting AI-generated text, underscoring the need for better detection methods.

• The findings suggest a potential paradigm shift is needed in how we approach student writing, with a possible focus on responsible and ethical AI integration in the classroom.

Why this is important: The ease with which AI can mimic student writing threatens to devalue the assessment process and calls for a reevaluation of how we measure genuine writing skills and critical thinking in an AI-powered world.

✅ 8 AI video & audio tools you NEED TO KNOW:

• 🎬 viggle : Replaces characters in videos using a single image. Great for memes and creative edits.

• 🪄 Runwayml : Removes objects from videos and seamlessly fills in the background.

• ✂️ Cap Cut : Free video editor with a built-in upscaler for higher resolution videos.

• 🎵 noisee : Generates animated music videos from songs, text prompts, and optional images.

• 🎹 udio : Free tool to create AI-generated music.

• ✨ Pixverse : Transforms text or images into videos. Can create consistent characters for recurring use.

• 🚀 hyper.ai : Similar to pixverse, does text-to-video and image-to-video, but currently limited to 2-second clips.

• 🗣️ did.ai : Adds lip-syncing and sound effects to videos. Offers free credits for trial.

Tech giants Stack Overflow and OpenAI announce a collaboration that will integrate their platforms to improve AI model capabilities and streamline the developer workflow.

The escalating threat of AI-powered cyberattacks drives a massive cybersecurity boom, with the industry projected to reach a $290 billion value by 2027.

Entrepreneur Bryan Johnson funds Catalog Technologies, a developer of DNA computers, aiming to leverage this technology for achieving superhuman intelligence after uploading his consciousness to a robot.

London-based company Wayve, specializing in autonomous vehicle AI technology, raises an impressive $1 billion in a funding round that includes SoftBank, Nvidia, and Microsoft.

Adobe's Project Neo is a new 3D creation tool that simplifies 3D asset creation and integration within the Adobe Creative Suite.

Anduril Industries introduces Pulsar, AI-powered electromagnetic warfare system designed to counter the rapidly evolving threats of modern electronic combat, where updates occur in days or hours.

Researchers propose 'multi-token prediction', LLMs just predict the next token in a sequence, one at a time, but if you have them predict the next several tokens at once you get significantly better performance, faster, and with no added costs. The gains are better for bigger models

Researchers introduce GSM1k, a new benchmark to rigorously assess large language models' (LLMs) mathematical reasoning abilities, revealing potential dataset contamination issues in popular benchmark datasets.

Researchers introduce Med-Gemini, a multimodal AI model designed for medical applications, demonstrating improved accuracy across various health data, including medical images, videos, genomics, and more.