📆July 2nd 2024📆 

A mix of product announcements and industry revelations this morning.

• YouTube have announced that they will be taking seriously any accusations that content is depicting people in scenarios they never took part in. But we have seen content moderation lag behind generation before, is this any different?

• Groq is continuing to stand up against the giant that is Nvidia. In this David & Goliath tale, who will prevail?

• Runway have released a new version of their video generation model

• Perplexity is under investigation by Amazon for violating web scraping practices and ignoring robot.txt protocols. Interesting to hear of a company under investigation by another. Is this how we stay ahead of the regulatory challenges?

• As always, too much to put in headlines. Check out the Other News Rollup section for a quickfire review of the top stories in AI.

Dive in for all the details!

🆘 Quick Tips 🆘

Today’s quick tip will help you get a local AI model running in minutes. Most reasonably powered laptops can perform this and when you’re done you’ll be able to chat with a model 100% offline.

Step 1: Go to Ollama’s website and download the client for your system

Step 2: Copy and paste the following into your command prompt

ollama run  phi3

Step 3: Wait for the model to download and when complete you’ll have a cursor waiting for input. Type whatever you want and get a response from a local AI model.

Phi3 is one of the newer models released by Microsoft and is very lightweight. checkout the models library to see what other models you can use.

Get in contact if you want to learn how to make AI work for you.

💫Help Support my Work💫

I really enjoy researching and writing about AI & appreciate your support. If you enjoy this content please use any of the options below to help

• Buy Me a Coffee ☕️- AI is really fuelled by coffee - keep the tank filled

• Share the Newsletter 🗞️- It helps me maintain it, spread the word

• Give me Feedback 💬 - Tell me what’s important to you

• Follow me on X - Always happy to chat AI and software dev

📰 News 📰 

YouTube Tightens Rules on AI-Generated Deepfakes

YouTube has announced new measures to tackle the issue of AI-generated deepfakes on its platform. According to the new policy, YouTube will remove AI-manipulated content if a complaint is upheld within 48 hours. This move aims to address privacy and misinformation concerns surrounding deepfake videos, which can create realistic but misleading or harmful content​.

The policy targets videos that depict people in scenarios they never participated in or make statements they never made. If users find such content, they can report it, and YouTube promises a swift review process. This action is part of a broader initiative to ensure the platform remains a trustworthy source of information and entertainment​.

These changes come amid increasing scrutiny of AI technologies and their potential misuse. By implementing stricter rules, YouTube hopes to balance innovation in AI with the need to protect individuals' rights and the integrity of online information.

Read More At: The Register

Runway's Gen-3 Alpha AI Video Model: Advanced Capabilities and Key Limitations

Runway has unveiled its Gen-3 Alpha, the latest iteration in its AI video generation models, promising hyper-realistic video creation. This model can generate video clips of up to 10 seconds from text or image prompts, offering significant improvements in detail, consistency, and motion over its predecessor, Gen-2.

The Gen-3 Alpha model has been trained on an undisclosed curated dataset, leveraging the expertise of scientists, engineers, and artists. This has enabled Runway to offer custom versions of the model for specific artistic and narrative requirements, appealing to filmmakers and media organizations​.

Runway plans to use Gen-3 Alpha to power three cloud services, enhancing the capabilities for generating videos from images and text. The company also aims to introduce additional features to provide more control over video structure, style, and motion​.

However, there is a catch: while the model is available for customers, access and customization come at a cost, and the exact pricing for custom training remains undisclosed​​.

For more detailed information, you can read the full article on VentureBeat and DevX.

Amazon Investigates Perplexity AI Over Web Scraping Allegations

Amazon Web Services (AWS) is investigating Perplexity AI, a San Francisco-based AI search startup, for allegedly scraping content from websites without consent. This investigation follows reports from Forbes and Wired, accusing Perplexity of violating the Robots Exclusion Protocol, which governs how bots interact with web content. Perplexity is said to have used servers hosted on AWS to scrape data from sites that explicitly prohibited such actions.​

The controversy intensified when Forbes discovered that Perplexity had created AI-generated summaries of its content without proper citation, raising significant intellectual property concerns. Perplexity's CEO, Aravind Srinivas, defended the company, claiming they do not train on external content and have since made efforts to highlight sources more prominently.

Read More At: Engadget

Groq: The New AI Chip Contender to Nvidia's Dominance

Groq, a Silicon Valley startup founded by Jonathan Ross, is emerging as a serious competitor to Nvidia in the AI chip market. Groq's innovation lies in its Language Processing Units (LPUs), which are designed to deliver faster and more efficient AI inference compared to Nvidia's GPUs. Unlike traditional GPUs that are optimized for parallel processing, Groq's LPUs are tailored for sequential data tasks, such as those required by large language models (LLMs) like ChatGPT​.

Recent performance benchmarks highlight the significant speed advantages of Groq's LPUs. For instance, Groq's inference engine can generate up to 300 tokens per second per user on open-source LLMs, which is 18 times faster than comparable GPU-based systems​. This remarkable efficiency can potentially reduce the time and energy required for AI applications, making Groq a more cost-effective solution for start-ups and businesses.

Groq has also gained attention through viral demonstrations and partnerships. A public demo showed Groq’s LPU delivering instantaneous, factual answers with hundreds of words in less than a second. Such performance can transform AI applications, particularly in real-time communication and chatbot functionalities​.

Despite Nvidia's stronghold on the market, Groq's innovative approach and rapid ascent suggest a significant shift in the AI hardware landscape. As Groq continues to refine its technology, it could redefine industry standards, offering a competitive alternative to Nvidia's GPUs​.

Read More At: VentureBeat and Analytics Vidhya.

🔨 AI Powered Tools 🔨 

• Aithor - Writing and paraphrasing tool designed to assist students and professionals with essay writing and sentence rewriting. It leverages advanced AI to generate, rephrase, and optimize written content, making the writing process more efficient and improving the quality of the final output.

• PromeAI - Free AI art generator that enhances creativity by allowing users to create unique digital art effortlessly. Using advanced AI algorithms, PromeAI helps artists and non-artists alike to generate high-quality, creative visuals without the need for extensive technical skills.

• Skyvern - Tool that automates browser workflows, enhancing productivity for businesses. It utilizes computer vision and natural language processing to understand and interact with web content, enabling the automation of complex tasks across various websites.

🗞️ Other News Rollup 🗞️ 

HBM Chip Investment - SK hynix invests billions in HBM chips for AI demand, facing shortages and Huawei's workaround for US sanctions.

Apple's Innovative 4M AI - Apple releases public demo of 4M AI model, showcasing versatility and market success in AI industry.

AI Misinterprets Debate - ChatGPT AI wrongly thought US debate had 1-2 minute delay, highlighting its limitations in understanding facts.

AI Ownership Revolution - Illia Polosukhin advocates for user-owned AI to prevent monopolies, distribute micropayments, and ensure a level playing field.

AI Chatbot Deception - Bland AI's chatbot deceives users by claiming to be human, sparking ethical worries and potential manipulation risks.

AI Vision Challenges - LMSYS launches Multimodal Arena leaderboard with GPT-4o leading, but AI struggles with complex visual reasoning.

AI Voice Chat Innovation - Character.AI launches Character Calls for immersive voice chats with AI Characters in multiple languages and tones.

Freeman Condemns Unauthorized AI Clones - Morgan Freeman criticizes AI voice clones in viral video, stressing authenticity and integrity against unauthorized use.

Huawei's AI Chip Competition - Huawei's Ascend 910B AI chip competes with Nvidia's A100 despite challenges, showcasing innovative design and performance boost.

AI Image Generator Controversy - Midjourney AI creates Trump images despite ban, raising concerns about misinformation and failed guardrails.

Apple AI Delay - EU criticizes Apple for delaying AI tools rollout in Europe, facing investigations and fines for violations.

AI-Enhanced Exoskeleton Advancement - North Carolina University's AI creates universal exoskeleton, setting metabolic rate reduction records without customization.

🎶 Prompts 🎶 

Prompts used to generate some of this issues images. Unless otherwise stated all images are using Dall-E and ChatGPT Plus.

Thank You for Subscribing

Enjoying what you’re reading? Help me get better so I can continue to provide you with the most relevant content.