Hi, AI! | media

📣 Hello! Welcome to our Sunday digest, featuring the most exciting AI news from the 26th week of 2024. TO WATCH 🎱 Elon Musk gave an interview about the future of the Internet, AI, and humanity at the recent Cannes Lions Festival… and he shared on X an impressive video showcasing the successful landing of two Falcon Heavy stages 🔼 SAVE THIS — IT'S HELPFUL 🎙️ Hedra: Generate Videos with Expressive AI Characters. Hedra AI will animate your image or generate a talking character from scratch based on a text prompt. TO READ 🐶 AI Dog Translator. Researchers from the University of Michigan have found a way to determine the context, breed, gender, and age of a dog by its bark. ❓ When Will the Data for Training LLMs Run Out? Humanity might face the strangest shortage in history in the next two years — running out of human-created texts. 🌊 AI as a Monster Wave Detector! Researchers have used AI to discover extreme waves' causes and created a model that can predict how and when they can occur. 👾 AI Buddy for Playing Minecraft. Most teenagers play video games to spend time with other players. In response to this trend, Altera is developing an AI buddy for playing Minecraft. 💊 AI to Search for New Antibiotics. Using machine learning, scientists have identified nearly a million compounds that could become the basis for new antimicrobial drugs. 🇸🇦 AI Superpowers: Saudi Arabia. $100+ billion for AI development, biggest startup space, chip and supercomputer manufacturing. ➡️ We are launching a channel for our readers in Arabic: Hi, AI! | الذكاء الاصطناعي #AIweek | @hiaimediaen

👾 AI Buddy for Playing Minecraft According to a Pew Research Center study, 72% of 1,423 surveyed teenagers ages 13-17 use video games to socialize with other players. In response to this trend, the startup Altera is developing an AI buddy for playing Minecraft. Who are the AI buddies? 🔥 AI buddies are AI models that play Minecraft alongside regular players. Unlike bots, they do not require prompts and act independently, resulting in a more natural game experience. Players can't "program" them, just as they can't program a real friend. 🌷 Altera was founded in December 2023 and has already raised $11.1 million in investments. Founders include former MIT neurologist Robert Yang. Altera aims to create digital "people" who will collaborate with humans to solve complex tasks, understand emotions, and establish real connections beyond current AI assistants. What makes Minecraft unique? 🌷Minecraft is the best-selling game in the world, created in 2009 by Mojang. The game has no set goals or storyline; everything depends on the player's improvisation. This creates an enormous scope for creativity and self-expression. 🌳 The game is built on mining and placing blocks — differing in texture, physics, and other characteristics — and crafting objects from ingredients. The minimalist cubic graphics evoke nostalgia for retro games and set Minecraft apart from other projects. 📱 Minecraft has become a cultural phenomenon: it is featured in movies, theatrical performances are based on it, and Netflix recently announced an animated Minecraft series for the game's 15th anniversary. ▶ To play Minecraft with Altera's AI buddy, you need to join the waitlist. The company plans to open access this summer. Sources: TechCrunch, VentureBeat, Game Industry Biz More on the topic: 🎮 7 Games About a Future Where Humans Live Alongside AI 🤖 How Did NPC AI Emerge in Video Games: Non-Player Characters with Their Opinion #news @hiaimediaen

🎱 Elon Musk Talks about the Future At the recent Cannes Lions International Festival of Creativity, Elon Musk gave an extensive interview about the future of the Internet, artificial intelligence, and humanity itself. 📱 Highlights ➡️ About AI prospects. One cannot be 100% optimistic or completely pessimistic. One of the godfathers of AI, Geoffrey Hinton, thinks there's a 10-20% probability of something terrible happening, like the extinction of the world's population. But I think the glass is 80% full. We're at the most interesting time in history. You'll see quite radical changes even next year and very radical changes within five years. ➡️ About life with superintelligence. I think the most likely outcome is one of abundance, where goods and services are available to anyone on Earth. It wouldn't be universal basic income; it would be universal high income. Work will be optional. But if AI can do everything that you can do, but better, I think there will perhaps be a crisis of meaning. In the positive scenario, the AI will be doing its be to make you happy. If some superintelligence is trying to make you happy, it'll probably succeed. ➡️ About Neuralink. The long-term goal of Neuralink is to achieve human-AI symbiosis. We think you could restore full body mobility by transferring the signals from the motor cortex in the brain past where the spinal injury is. We've already demonstrated this in animals, and it's looking very good. ➡️ About the future of the Internet. About the future of the Internet There's going to be a significant disruption in Internet search. If AI can provide you with a better answer than a bunch of links, then you'll prefer that over Google. There’ll be a new model of news. You've got real-time aggregation of the collective wisdom of tens of millions of people — the information from people who are at the scene and who are experts in the field. ➡️ About humanoid robots. A fully functional humanoid robot will be doing a wide range of tasks: walking your dog, taking care of your house, babysitting the kids, teaching the kids, cooking dinner, and playing the piano. I think everyone will want one. There'll be at least one for every person and then a whole bunch more in the industry making things, so my guess is 20 billion. People will start to regard their personal Optimus robot as a sort of friend, like R2-D2 and C-3PO in Star Wars. ➡️ About multiplanetary travel. If we're a multiplanetary civilization, our probable lifespan is much greater. We can explore the galaxy and perhaps we'll find many long-dead one-planet civilizations out there. Expanding the scope and scale of consciousness will help us better understand what questions to ask about the universe. 🎁 Today is Ilon Musk's birthday - 53 years old. Congratulations! 📱 Interview here. More on the topic: Elon Musk vs Jeff Bezos: Who Will Be the First on the Moon? Interview with Sam Altman on the Future of AI #news @hiaimediaen

🇸🇦 AI Superpowers: Saudi Arabia While the US and China remain the key players in the global AI race, other countries are picking up the pace of investments as well. We will be covering them, starting with Saudi Arabia. The Kingdom of Saudi Arabia (KSA) is looking to diversify its oil-driven economy by developing tourism, sports, and technology. Initiatives include building data centers, developing AI chips, and cloud technology, opening offices of international tech companies, and launching AI startups. 💸 Investments ▶ Earlier this year, Saudi Arabia created a $100 billion fund to invest in AI and other technology. ▶ Amazon announced at a conference in Riyadh a $5.3 billion investment in data centers and AI technology in KSA. ▶ The Saudi government is in talks with the Silicon Valley venture capital firm Andreessen Horowitz about investing $40 billion in AI development. If the deal goes through, the country will become the world's largest AI investor. 🤝 Projects ✅ Last year, the Garage opened in the Saudi capital. Boasting 28,000 square meters, it is the largest startup space in the Middle East, housing more than 300 companies. ✅ The Saudi government has launched GAIA, a generative AI startup accelerator similar to Y Combinator. The $1 billion project will attract young AI entrepreneurs and create more than 300 companies over three years. Each startup will receive $40,000 and a potential investment of $100,000, provided it operates in Riyadh and registers the company in the kingdom. ✅ King Abdullah University of Science and Technology (KAUST) has become one of the world's academic centers for AI research, attracting scientists from the US, China, and other countries. They are building the Shaheen 3 supercomputer and discussing a $100 million deal with Nvidia (pending US approval). ⬆️ Prospects 📊 According to the Kingdom's Vision 2030 program, Saudi Arabia is set to become a world leader in AI by 2030. 📊 The desert along the Red Sea will be home to cities with a fully AI-powered infrastructure. The first smart city, Neom, is being built in the northwest of the country. Source @hiaimediaen

📱 Did you know you can connect your AI assistant to Telegram Business? Stay connected with clients, students, employees, and candidates 24/7, even while on vacation! Deliver personalized and accurate responses precisely when needed. Let the assistant engage users in dialogue and clarify all necessary details to resolve issues quickly. 👨‍💻 Connecting takes just a few clicks, and our specialist will help train the assistant. All you need is an account on our AI assistant management platform, Brainfab. Test the Brainfab's CEO AI assistant: @andreykovtun. Additionally, we share practical case studies and ideas for applying AI to automate real businesses on our channel, covering everything from law firms to development agencies: @aidoit. 🔖 Get a lifetime subscription to the full functionality of the Brainfab platform for a one-time payment of $250, available for teams of up to 10 people. Limited offer! ▶️ Watch a video presentation. ➡️ Register an account, and we'll set everything up! #Sponsored

🌊 AI as a Monster Wave Detector! Rogue waves, as tall as a ten-story building, often appear in the ocean. They can break apart or sink ships and damage oil platforms. Legends of such "monster waves" have been around since ancient times. In 2015, it was possible to film this rare phenomenon: huge waves crashed on the Borgholm Dolphin platform in the North Sea. Still, until recently, scientists have not been able to fully explain the origin of this phenomenon or predict its occurrence. Researchers from the University of Copenhagen have used artificial intelligence to discover these extreme waves’ causes and created an AI model that can predict how and when they can occur. 🔓 What are the findings?  In their model, the researchers combined available data about ocean and sea movements, depth and bathymetric information. The buoys located at 158 different points along the US coast and overseas territories collected the wave data 24 hours a day. Scientists ended up with a dataset containing 700 years’ worth of wave height and sea state information from more than a billion waves. They also registered more than 100,000 unusually high waves that could be defined as rogue ones. The analysis showed that abnormal waves can appear at almost any time. They arise from the interaction between currents and winds, when two wave systems cross and quickly reinforce one another. In this case, high crests and deep troughs emerge, making waves at least twice as high as the others.  First identified in the 18th century, this phenomenon has now been validated by AI. With the creation of a new mathematical model, AI has achieved unprecedented accuracy in predicting rogue waves, far surpassing earlier forecasting techniques. 🚢 Why is it useful? There are around 50,000 cargo ships sailing around the world at any given time. Using this publicly available algorithm, it is possible to receive warnings of dangerous waves and send this information to weather services.

As shipping companies plan their routes well in advance, they can use our algorithm to get a risk assessment of whether there is a chance of encountering dangerous rogue waves along the way. Based on this, they can choose alternative routes. Dion Häfner, co-author of the study

#news @hiaimediaen

❓ When Will the Data for Training LLMs run out? In the next 2 years, humanity might face the strangest shortage in history — running out of human-created texts. This will lead to language models (LLMs) depleting their training data, causing a scaling crisis. Researchers studying AI's impact on our world have come to this conclusion.

Number of the day 300 trillion tokens — the amount of text created by humanity that is currently available for training AI models.

0️⃣ "Data Drought" 2026–2032 — researchers consider this period the most likely timeframe for the complete depletion of text data for training LLMs. It could happen even sooner if models are heavily overtrained due to the AI race and the scaling of popular LLMs. Three Main Conclusions from Researchers 1️⃣ Textual data will become the bottleneck in developing more advanced LLMs. 2️⃣ Synthetic data from AI is still insufficiently studied. They are useful in narrow fields like mathematics and programming. Some believe such data can be dangerous as AI might make mistakes when creating them. 3️⃣ Private data, such as personal messages, are unlikely to be used on a large scale due to legal issues. 🔠 Solutions to the Crisis Researchers propose several solutions for developing LLMs: ➡️ Synthetic data. ➡️ Training on other types of data. ➡️ Increasing data efficiency. 💲 Who Can I Sell My Data to Companies are already offering internet users monetary rewards for their data, which can be used to train AI models. Here are some of them: ➡️ TIKI — for access to users' mobile devices. They are interested in user behavior within apps partnered with TIKI. ➡️ Caden — for access to personal accounts on Netflix and Amazon. Earnings range from $5 to $50 per month. ➡️ Invisible offers access to paid news articles in exchange for demographic and behavioral data, including information on vaccinations and users' political affiliations. The company plans to trade this data for digital subscriptions costing between $4 and $15 per month. @hiaimediaen

🐶 AI Dog Translator We continue to monitor AI translators from animal to human language. Researchers from the University of Michigan have found a way to determine the context, breed, gender, and age of a dog by its bark. 🧪 How did they do this? In collaboration with Mexico’s National Institute of Astrophysics, Optics, and Electronics, scientists discovered that AI models initially trained on human speech may also be applied to animals. Following this discovery, the researchers collected recordings of barks from 74 dogs of various breeds, ages, and genders, including aggressive, playful, and anxious barks. They chose the Wav2Vec2 model initially trained on human speech data for the analysis. This is the first time such methods have been used to decode animal communication. Wav2Vec2 successfully handled the task and outperformed other models trained, especially on dog bark data, by 70%. 😉 What happens next? Scientists believe studying patterns in human speech using AI can also help understand other animals. Source: University of Michigan Would you like to find out what your pet is saying? ❤️ — Absolutely! 🙉 — I already know it's asking for food More on the topic: 🐳 What do sperm whales talk about? 🐶 Lost in Translation: Will AI soon help us talk to animals? #news @hiaimediaen

🎙️ Hedra: Generate Video with Expressive AI Characters Hedra AI is a new model for creating and animating videos with virtual characters. It will animate your image or generate a talking character from scratch based on a text prompt. Hedra is currently in beta testing, and you can try the service for free. Key features: ➡️ Voice Generation. Synthesizes voice. ➡️ Image animation. Brings static images to life by adding emotion and movement to them. ➡️ Lip sync. Synchronizes characters' lip movements with voiceover. How to use it? 1️⃣ Go to website 2️⃣ Write a text voiceover 3️⃣ Select a voiceover 4️⃣ Upload a character photo or describe it to generate a new avatar 5️⃣ Click Generate video ✅ In the first 48 hours, the service has created 100,000 videos! For more instructions at #manual More on the topic 👾 VASA — Hyper-Realistic Talking Avatar ⚡️ ElevenLabs Dubbing Studio — now you can edit dubbed videos #startup @hiaimediaen

💊 AI to Search for New Antibiotics Using machine learning, scientists at the University of Pennsylvania have identified nearly a million compounds that could become the basis for new antimicrobial drugs. Antibiotic-resistant infections kill more than a million people a year. The WHO predicts that by 2050, the number of deaths could rise to 10 million. Previously, scientists studied the effectiveness of antimicrobials using traditional methods, such as testing water and soil samples. It took years to find and create drugs. Artificial intelligence can significantly speed up the process. 💥 How Does it Work? Scientists used AI to analyze huge datasets, containing the genomes of tens of thousands of microorganisms. Looking for DNA fragments that might have antimicrobial activity, researchers selected around 900,000 antimicrobial peptides. At least 79% of these fell into the “candidates for antibiotics” category, meaning they have the potential to kill at least one pathogen. In the laboratory, scientists synthesized 100 of these peptides and tested them on 11 dangerous bacterial pathogens, including antibiotic-resistant strains of Escherichia coli and Staphylococcus aureus.

AI in antibiotic discovery is now a reality and has significantly accelerated our ability to discover new candidate drugs. What once took years can now be achieved in hours using computers. César de la Fuente, PhD, study co-author

The results of the study are published in the journal Cell. 🆕 In the history of medicine, this is the most promising effort to find new drugs. In 2020, scientists at the Massachusetts Institute of Technology discovered, with the help of AI, the first new antibiotics in 30 years. More on the topic: Drug prescription by AI How AI Is Pioneering New Drug Discoveries #news @hiaimediaen