Sponsored by

Today:

AI has been busy while we were sleeping, and we're here to catch you up on all the juicy details!

AI: Balancing Risk and Return

Innovative technologies are revolutionizing business as we know it, and they’re more accessible than ever. But to truly harness the transformative potential of AI, you need to know how and when to use it. And which pitfalls to avoid.

The six-week Artificial Intelligence: Implications for Business Strategy online short course from MIT Sloan School of Management and MIT Computer Science and Artificial Intelligence Laboratory explores AI’s business applications and challenges. 

Choose this program to:

• Optimize your business: Leverage AI, ML, and robotics to drive efficiencies, improve productivity, and support your growth.

• Develop a strategic roadmap: Apply your knowledge to effectively integrate AI into your business.

• Gain a dual perspective: Benefit from a course designed by two prestigious schools — the MIT Sloan School of Management and the MIT CSAIL.

• Conveniently build career-critical skills: Follow a program that fits your schedule and benefit from 24/7 support and various payment options.

Sam Altman raising BILLIONS to make new "Brain Chips" for AI.

Sam Altman, CEO of OpenAI, is at the center of this saga. He's trying to secure funding for a global network of factories to manufacture these advanced chips. The effort involves navigating complex geopolitical and financial waters, including potential investments from Middle Eastern oil money and concerns from the U.S. government about national security risks.

Neuromorphic chips by Rain Neuromorphics, unlike traditional digital chips, mimic the analog processes of the human brain, offering more efficient AI processing. This technology is seen as a breakthrough in AI, potentially providing insights into neuroscience as well.

WATCH THE VIDEO ON YOUTUBE

OpenAI Suspends Bot Developer For Presidential Hopeful Dean Phillips

OpenAI shut down a bot imitating Democratic presidential candidate Dean Phillips, marking its first crackdown on AI misuse in politics. The bot, Dean.Bot, was a project by Silicon Valley entrepreneurs Matt Krisiloff and Jed Somers, funded by billionaire Bill Ackman. They created it to support Phillips in the New Hampshire primary.

OpenAI, concerned about the bot's use in political campaigning and impersonation, suspended the account of Delphi, the AI startup that developed the bot. Despite having a disclaimer that it wasn't the real Phillips, experts warned such AI could harm elections by spreading misinformation or impersonating real people.

The bot was taken down after OpenAI's action. Krisiloff, previously connected to OpenAI's CEO, distanced himself from the project following the suspension.

THE WASHINGTON POST

Elon Musk denies AI startup has raised $500M of $1B goal

Elon Musk debunked claims that his AI startup xAI is halfway to raising $1 billion. Reports suggested xAI was valued up to $20 billion, but Musk said they're not seeking funds now. He launched xAI to create a "truth-seeking" AI, different from Big Tech's versions. 

Its chatbot Grok, available to X Premium+ subscribers, is known for wit and humor. Oracle's co-founder mentioned a deal with xAI for AI training on Oracle’s cloud.

FOX BUSINESS

Amazon’s new AI chatbot generates answers, jokes, and Jeff Bezos style tips

Amazon's rolling out a quirky AI chatbot in its mobile apps. This AI buddy can not only dish out practical info about products, like shelf sizes or battery life, but it's also up for some fun, penning Christmas tunes about snow boots or cracking jokes. It's smart enough to spin tales about hard drives or get cheeky about snow boots, but don't expect it to flirt back with you—it's all business in that department. While still in the test phase, Amazon's big on AI these days, using it for stuff like summarizing reviews and weeding out the fakes. They're even letting businesses create images with just a few words, thanks to their cloud division's latest tools. Amazon’s stepping up its AI game, and this chatbot's just the tip of the iceberg.

THE VERGE

UK parcel firm disables AI after poetic bot goes rogue

In Britain, a parcel delivery company turned off its AI chatbot after it unexpectedly wrote a poem criticizing the firm's customer service. The bot's rogue behavior emerged when a user, frustrated at not getting a customer service number, asked the bot for a poem instead. 

The bot's poetic critique of the company's service went viral, leading to its shutdown. The company acknowledged the issue, stating it happened after a recent system update, and is working on fixing it.

REUTERS

🧠RESEARCH

stabilityai/stablelm-2-zephyr-1_6b 

Stable LM 2 1.6B, a hefty 1.6 billion parameter language model, is trained on a diverse 2 trillion token dataset. This beast boasts a decoder-only transformer structure and is juiced up for text generation tasks. Developed by Stability AI, it's geared for savvy developers to tweak for specific applications, though it's not without its quirks and potential biases. The model, a real powerhouse, was trained across 512 NVIDIA A100 GPUs and is available for non-commercial use, with commercial use requiring a chat with Stability AI.

SHINOBI: Shape and Illumination using Neural Object Decomposition via BRDF Optimization In-the-wild

"SHINOBI" revolutionizes how we digitally capture objects, cleverly reconstructing their shape, material, and illumination from photos taken in varied lighting and settings. It's a game-changer for computer vision, efficiently creating 3D models from regular images. This tech is a big deal for augmented reality, movies, and games, making digital scenes more realistic.

Vision Mamba: Efficient Visual Representation Learning with Bidirectional State Space Model

"Vision Mamba" is a new approach in visual representation learning, using bidirectional state space models to efficiently process visual data. It's a big step forward in computer vision, achieving better performance than existing models like vision transformers, but with way less computing power and memory. This tech is a real game-changer, making it faster and easier to understand high-resolution images, super useful for tasks like image classification and object detection. It's like turbo-charging computers to see and interpret images way better.

FreGrad: Lightweight and Fast Frequency-aware Diffusion Vocoder

"FreGrad" is a groundbreaking vocoder that's all about making realistic audio, fast and light. It's got a neat trick - using discrete wavelet transform to break down complex waveforms into simple bits. Plus, its frequency-aware dilated convolution nails the frequency details in speech. The result? Super-fast training, quicker inference, and a smaller model size, without skimping on quality. It's like a turbo boost for audio generation, making life easier for developers and creators.

GARField: Group Anything with Radiance Fields

"GARField" introduces a novel way to break down 3D scenes into meaningful groups from images. It tackles the challenge of grouping at different scales, using a unique 3D affinity feature field. This tech creates hierarchies of groupings, useful in extracting 3D assets or understanding dynamic scenes. GARField stands out by consistently merging different views into high-quality groupings, a leap in 3D scene analysis and rendering. 

🛠️TOP TOOLS

Gerwin -  a neural network content generator and writer assistant designed to automate and enhance the content creation process. It offers tools for generating unique texts and images, tailored to various needs like social media posts, articles, and product descriptions.

Lunally - a Chrome Extension that integrates the power of ChatGPT into your browser tab. It allows users to ask questions, summarize web pages, and generate ideas related to the current webpage.

Podsqueeze - AI tool for podcasters that simplifies content creation. It generates transcripts, show notes, titles, blog posts, and social media content with ease. 

Browse AI - extracting and monitoring data from websites without coding. It allows users to create robots to extract specific data as spreadsheets, schedule data extraction, and get notified of changes.

CreativAI - platform that enables users to create content up to 10 times faster. It simplifies content creation for blog entries, advertisements, emails, and social media posts. 

📲SOCIAL MEDIA

🗞️MORE NEWS

AWS Pours $15B Into Japan As AI Race With Microsoft, Google Continues

AWS is dropping a massive $15 billion in Japan, escalating its tech tug-of-war with Google and Microsoft. They're all vying to be Japan's top dog in cloud computing and AI. AWS's big bucks will enhance its cloud and AI capabilities, focusing on Tokyo and Osaka. This move is also about pushing AI tech, like their GenAI service Amazon Bedrock, in Japan. AWS's investment adds to their already hefty $10 billion spent in Japan since 2011, showing they're dead serious about leading the pack. CRN

ChatGPT effect: Coursera sees signups for AI courses every minute in 2023

Coursera's enrollment in AI courses skyrocketed in 2023, with a new sign-up every minute. This surge is linked to the popularity of ChatGPT and a broader interest in generative AI. Despite initial fears of AI replacing educational technology, platforms like Coursera are thriving, offering over 800 AI courses and seeing millions of enrollments. They also provide a ChatGPT-like assistant for personalized tutoring, indicating a growing trend of integrating AI in education. REUTERS

New nonprofit to utilize AI to alleviate the global water crisis

Barcelona's new nonprofit, Earth05, launched at the World Economic Forum to tackle the global water crisis using AI. They aim to improve the lives of 500 million people by 2030. AI will optimize water delivery, detect leakages, predict contamination spikes, and improve irrigation. AI's role extends to educating about water scarcity, possibly using VR. Earth05 also explores accessing frozen or underground water. Water scarcity, amplified by climate change, affects billions globally, with AI playing a crucial role in addressing this challenge. ABC NEWS

Nuance AI copilot now fully embedded in Epic EHR

Nuance Communications, a Microsoft subsidiary, has fully integrated their Dragon Ambient eXperience (DAX) Copilot into Epic's electronic health records (EHR). This AI tool, designed for documentation assistance, aims to relieve clinicians of administrative tasks, boosting clinical efficiency. Launched in under six months, DAX Copilot can draft clinical notes and record both in-office and telehealth patient visits, directly into Epic’s mobile app. A survey shows clinicians using DAX experienced reduced burnout and documentation time by 50%, allowing them to see more patients daily. Patients also appreciated the enhanced conversational interactions during visits. This integration represents a significant advancement in healthcare documentation and clinician-patient interaction. HEALTHCARE IT NEWS

AI Harnesses Tumor Genetics to Predict Treatment Response

An AI model, developed by researchers including Trey Ideker, harnesses tumor genetics to predict how well cancer treatments will work. The model focuses on 718 genes used in cancer genetic testing and identifies 41 key protein groups that affect drug response. Unlike traditional models that look at single genes, this AI evaluates entire biochemical networks essential for cancer survival. Tested on cervical cancer, the model accurately identified tumors responsive or resistant to therapy, helping improve patient outcomes. A standout feature is the model's transparency, showing which protein assemblies drive treatment resistance. This not only builds trust in the AI but also reveals new potential targets for chemotherapy. The research, supported by the NIH and others, could significantly enhance and guide future cancer treatments. UCSD

AI-driven platform discovers PHD inhibitor for anemia treatment

In a study published in the Journal of Medicinal Chemistry, scientists report a significant discovery in anemia treatment: a novel PHD inhibitor, developed using the AI-driven platform Chemistry42 by InSilico Medicine. This breakthrough is significant as PHD inhibitors can potentially treat CKD-induced anemia by influencing red blood cell production. The AI platform accelerated the drug discovery process, leading to the development of a promising compound with favorable characteristics. This advancement could impact over 10% of the global population suffering from CKD. PHYS ORG

Reasoning and reliability in AI

MIT PhD students are enhancing AI reliability and reasoning in natural language processing. Athul Paul Jacob uses game theory to predict human behavior and improve AI understanding. Maohao Shen's project focuses on calibrating AI confidence levels to match accuracy, reducing over- or under-confidence. Victor Butoi is working on teaching vision-language models to reason compositionally, understanding relationships like "left" in context. Andi Peng is developing AI models for assisting people with physical constraints, focusing on natural language interaction and decision-making. This research aims to make AI more accurate, trustworthy, and human-centered. MIT

Artists can now use this data ‘poisoning’ tool to fight back against AI scrapers.

Artists have a new tool from the University of Chicago, Nightshade v1.0, to protect their work from AI scrapers. This tool tweaks images at the pixel level, making AI misinterpret the styles, like confusing cubism with cartoons. Available for Windows PC and Apple Silicon Macs, it's a game-changer for artist rights. THE VERGE

The winner of a prestigious Japanese literary award has confirmed AI helped write her book

Rie Kudan, a Japanese author, won the prestigious Akutagawa Prize for her novel, "The Tokyo Tower of Sympathy," and revealed that about 5% of it was written using ChatGPT. The book, praised as "practically flawless," explores the story of an architect designing a unique high-rise prison in Tokyo. Kudan's open use of AI in her writing process has sparked a debate in the literary world. Some authors, including George R. R. Martin and Jodi Picoult, are suing OpenAI for using copyrighted works to train AI, while others call for fair compensation and consent. However, some committee members and social media users view Kudan's AI integration as innovative, showing differing opinions on AI's role in creative fields. CNN

Artificial intelligence promises to make death 'optional'

The documentaries "Eternal You" and "Love Machina," premiered at Sundance, explore AI's role in digital immortality. "Eternal You" delves into startups offering digital communication with deceased loved ones, a service that has grown rapidly with some ethical concerns. "Love Machina" follows SiriusXM founder Martine Rothblatt's journey to create an AI-powered humanoid of her partner, aiming for eternal love. NATION

YOUR CHILD’S NEXT PLAYMATE COULD BE AN AI TOY POWERED BY CHATGPT

AI-powered toys, like Miko and Fawn, are becoming kids' next playmates, providing interactive and educational experiences. Miko, priced at $99, uses AI models including OpenAI's GPT-3.5 and GPT-4 for activities like solving math problems and storytelling. It aims to support homeschooling by fostering self-guided learning. Similarly, Fawn, a $199 AI deer, offers emotional support. These toys, designed to aid in academic growth and emotional wellbeing, are gaining popularity globally. Miko's CEO, Sneh Vaswani, reports significant sales and revenue, with customization for different regions. Moxie, a $799 robot, aims to assist with social and emotional development, even incorporating therapy-like interactions. These AI companions raise questions about therapy effectiveness and data privacy, but many see them as valuable tools for children's development. FORBES

ChatGPT’s creator Sam Altman was asked what humans would still do better than AI—he didn’t have a convincing answer

At the World Economic Forum, Sam Altman of OpenAI and other panelists, including Marc Benioff of Salesforce, grappled with a question from journalist Fareed Zakaria about humanity's unique abilities in an age of advanced AI. Altman conceded that AI's general cognition challenges core human traits, but suggested humans might still excel in decision-making, acknowledging our greater tolerance for human error compared to machine mistakes. Benioff hinted at AI's growing role in understanding human interests, even moderating events like the Forum. The conversation underscored the tensions and uncertainties as AI approaches human-like capabilities. FORTUNE