🖧 Blockchain + AI = Decentralized AI

Blockchain + AI = Decentralized AI ?

Let’s not talk about how first, but WHY?

Less reliant on dominating corps.

Resources are pooled and accessable to anyone.

Foster innovation from smaller entities and individuals.

Reduce mass surveillance and manipulation by centralized parties.

Sounds cool, but do you care?

AFFECT YOU AND ME

🤗 You are more likely to be convinced by a chatbot than another human

Recent research indicates that personalized chatbots, particularly those powered by large language models like GPT-4, have a significantly higher persuasive power in debates compared to humans.

The study found that when GPT-4 had access to personal information about the participants, it could increase agreement with its arguments by 81.7%, a stark contrast to the 21.3% increase without personalization, which wasn't statistically significant.

This suggests that personalized AI can skillfully use personal information to craft more persuasive arguments.

🧠 A human is playing chess using thoughts only

Noland Arbaugh, the first human recipient of Elon Musk's Neuralink brain chip, is featured in a video playing chess using his thoughts.

The Neuralink implant, named "Telepathy," allows users to control devices with their mind. Arbaugh, a quadriplegic due to a diving accident, demonstrates increased accessibility to tasks and recreation.

The technology has sparked both amazement and controversy, with concerns about the ethics of human testing and the suffering of animals in trials.

💉 NVIDIA AI nurse only cost $9  

NVIDIA has partnered with Hippocratic AI to develop AI-powered healthcare agents, aiming to address the global healthcare worker shortage.

These agents, designed to outperform human nurses in specific tasks, could potentially ease the strain on healthcare systems, especially in areas facing severe workforce shortages.

The AI agents will cost only $9 per hour.

🔖 YouTube creators required to label AI content

YouTube is mandating that creators label content that could be mistaken for real but is made with AI or altered media. This includes realistic depictions of people, places, or events.

A new tool in Creator Studio will facilitate disclosure, aiming to foster transparency and trust with viewers.

The label will clarify if content is altered or synthetic, but won't be required for clearly unrealistic or minor modifications like color adjustments or special effects.

🤝🏻 Google partners with Apple for iPhone AI features

Apple is negotiating with Google to integrate the Gemini AI into iPhones, potentially enhancing AI features.

The deal may boost Google's AI presence across Apple's devices and alleviate concerns about Apple's AI progress. However, it could also attract regulatory scrutiny due to the companies' existing search engine agreement.

🍎 Apple adding Baidu’s AI to iPhone in China

Apple is reportedly in early discussions with Baidu to incorporate Baidu's generative AI technology into Apple devices in China.

This potential partnership aims to boost Apple's AI capabilities while adhering to China's strict AI regulations.

Baidu's Ernie Bot, already approved by China's cyberspace regulator, could be a key component in Apple's strategy to enhance device features and maintain competitiveness in the Chinese market.

🤑 Saudi is now the world’s largest AI investor

Saudi Arabia is planning to establish a $40 billion fund dedicated to investing in AI, positioning itself as the world's largest investor in AI.

The country's sovereign wealth fund, which has assets exceeding $900 billion, is discussing a potential partnership with Andreessen Horowitz, a prominent Silicon Valley VCl firm, and other financiers.

However, Anthropic, an AI company, has declined funding from Saudi Arabia due to national security concerns.

🖼️ Turn single picture into life-like videos

Google's VLOGGER AI, a groundbreaking system, can animate still photos into dynamic videos with realistic speech, gestures, and facial expressions.

This technology, named "Multimodal Diffusion for Embodied Avatar Synthesis," is trained on vast data, offering exciting possibilities and raising concerns about deepfakes.

Google is developing methods to flag AI-generated videos, including digital watermarking, and is working on ethical guidelines for VLOGGER's use.

AI INSIDER

🔥 Midjourney going real-time and 3D

MidJourney is set to revolutionize the AI landscape with its ambitious expansion into 3D, video, and real-time creation models.

This expansion aims to simulate the entire world, offering users the ability to create immersive virtual environments for video games or movies.

The upcoming v7 model promises even greater advancements. MidJourney's focus on creating a comprehensive world simulation reflects the vision of an "open world sandbox," where creativity meets cutting-edge technology.

👋🏻 Stability CEO quits for decentralized AI

Emad Mostaque, the CEO and founder of Stability AI, has resigned to concentrate on decentralized AI, which promotes transparent and distributed governance.

Mostaque's departure follows financial challenges and key employee exits at Stability AI, which is known for its Stable Diffusion image generation tool.

The company has faced lawsuits, including one from Getty Images, and Mostaque has emphasized the need for AI to remain open and decentralized.

💥 Massive break through in genAI modeling approach

Sakana AI, a Tokyo-based startup, has developed a novel method for creating generative AI models using principles of evolution and natural selection.

This innovative approach allows for the development of AI models that are both cheaper and faster to produce. Sakana’s models were developed using what Sakana calls a "nature-inspired" mechanism, marking a first in the industry.

💬 AI can now explains its logic and reasoning to you

Researchers from Stanford and Notbad AI have developed an AI model named Quiet-STaR that significantly improves AI reasoning by incorporating an "inner monologue."

This model pauses to "think" before responding, showing its reasoning process and allowing users to select the most accurate response.

The approach has led to a notable improvement in accuracy, from 36.3% to 47.2% in reasoning tests.

🤔 Microsoft’s interesting license deal with AI startup

Microsoft has agreed to pay Inflection AI $650 million to license its AI software after hiring the startup's co-founders along with most of its staff.

This deal, which includes $620 million for licensing Inflection’s AI models and $30 million for legal rights related to the mass hiring, has significantly reduced Inflection AI's workforce.

The company is now looking to offload some of its computing power. Despite the substantial investment, the return for Inflection AI's investors is expected to be modest, especially considering the company's previous valuation of $4 billion.

AI GENERATED ART

The unimaginable tennis courts