AI Weekly Rundown (July 6 to July 12)
Major AI announcements from Cloudflare, Microsoft, Anthropic, OpenAI, and more.
Hello Engineering Leaders and AI Enthusiasts!
Another eventful week in the AI realm. Lots of big news from huge enterprises.
In today’s edition:
🛡️ Cloudflare launches a one-click feature to block all AI scraping bots
🆕 SenseTime released SenseNova 5.5 at the 2024 World AI Conference
🚨 Waymo's Robotaxi gets busted by the cops
🖼️ LivePotrait animates images from video with precision
⏱️ Microsoft’s ‘MInference’ slashes LLM processing time by 90%
🚀 Groq’s LLM engine surpasses Nvidia GPU processing
🎬 Odyssey is building a ‘Hollywood-grade’ visual AI
📜 Anthropic adds a playground to craft high-quality prompts
🧠 Google’s digital reconstruction of the human brain with AI
🧬 OpenAI teams up with Los Alamos Lab to advance bioscience research
📊 China is leading global gen AI adoption: A New survey reveals
⌚ Samsung introduces new, advanced AI wearables at ‘Unpacked 2024’
🤖 Google’s Gemini 1.5 Pro gets a body: DeepMind’s office “helper” robot
🌐 OpenAI’s new scale to track the progress of its LLMs toward AGI
📢 Amazon announces a blitz of new AI updates for AWS
Let’s go!
SenseTime released SenseNova 5.5 at the 2024 World Artificial Intelligence Conference
Leading Chinese AI company SenseTime released an upgrade to its SenseNova large model. The new 5.5 version boasts China's first real-time multimodal model on par with GPT-4o, a cheaper IoT-ready edge model, and a rapidly growing customer base.
SenseNova 5.5 packs a 30% performance boost, matching GPT-4o in interactivity and key metrics. The suite includes SenseNova 5o for seamless human-like interaction and SenseChat Lite-5.5 for lightning-fast inference on edge devices.
With industry-specific models for finance, agriculture, and tourism, SenseTime claims significant efficiency improvements in these sectors, such as 5x improvement in agricultural analysis and 8x in travel planning efficiency.
Cloudflare launched a one-click feature to block all AI bots
Cloudflare just dropped a single-click tool to block all AI scrapers and crawlers. With demand for training data soaring and sneaky bots rising, this new feature helps users protect their precious content without hassle.
Bytespider, Amazonbot, ClaudeBot, and GPTBot are the most active AI crawlers on Cloudflare's network. Some bots spoof user agents to appear as real browsers, but Cloudflare's ML models still identify them. It uses global network signals to detect and block new scraping tools in real time. Customers can report misbehaving AI bots to Cloudflare for investigation.
Waymo's Robotaxi gets busted by the cops
A self-driving Waymo vehicle was pulled over by a police officer in Phoenix after running a red light. The vehicle briefly entered an oncoming traffic lane before entering a parking lot. Bodycam footage shows the officer finding no one in the self-driving Jaguar I-Pace. Dispatch records state the vehicle "freaked out," and the officer couldn't issue a citation to the computer.
Waymo initially refused to discuss the incident but later claimed inconsistent construction signage caused the vehicle to enter the wrong lane for 30 seconds. Federal regulators are investigating the safety of Waymo's self-driving software.
LivePotrait animates images from video with precision
LivePortrait is a new method for animating still portraits using video. Instead of using expensive diffusion models, LivePortrait builds on an efficient "implicit keypoint" approach. This allows it to generate high-quality animations quickly and with precise control.
The key innovations in LivePortrait are:
1) Scaling up the training data to 69 million frames, using a mix of video and images, to improve generalization.
2) Designing new motion transformation and optimization techniques to get better facial expressions and details like eye movements.
3) Adding new "stitching" and "retargeting" modules that allow the user to precisely control aspects of the animation, like the eyes and lips.
4) This allows the method to animate portraits across diverse realistic and artistic styles while maintaining high computational efficiency.
5) LivePortrait can generate 512x512 portrait animations in just 12.8ms on an RTX 4090 GPU.
Microsoft’s ‘MInference’ slashes LLM processing time by 90%
Microsoft has unveiled a new method called MInference that can reduce LLM processing time by up to 90% for inputs of one million tokens (equivalent to about 700 pages of text) while maintaining accuracy. MInference is designed to accelerate the "pre-filling" stage of LLM processing, which typically becomes a bottleneck when dealing with long text inputs.
Microsoft has released an interactive demo of MInference on the Hugging Face AI platform, allowing developers and researchers to test the technology directly in their web browsers. This hands-on approach aims to get the broader AI community involved in validating and refining the technology.
Groq’s LLM engine surpasses Nvidia GPU processing
Groq, a company that promises faster and more efficient AI processing, has unveiled a lightning-fast LLM engine. Their new LLM engine can handle queries at over 1,250 tokens per second, which is much faster than what GPU chips from companies like Nvidia can do. This allows Groq's engine to provide near-instant responses to user queries and tasks.
Groq's LLM engine has gained massive adoption, with its developer base rocketing past 280,000 in just 4 months. The company offers the engine for free, allowing developers to easily swap apps built on OpenAI's models to run on Groq's more efficient platform. Groq claims its technology uses about a third of the power of a GPU, making it a more energy-efficient option.
Odyssey is building a ‘Hollywood-grade’ visual AI
Odyssey, a young AI startup, is pioneering Hollywood-grade visual AI that will allow for both generation and direction of beautiful scenery, characters, lighting, and motion.
It aims to give users full, fine-tuned control over every element in their scenes– all the way to the low-level materials, lighting, motion, and more. Instead of training one model that restricts users to a single input and a single, non-editable output, Odyssey is training four powerful generative models to enable its capabilities. Odyssey’s creators claim the technology is what comes after text-to-video.
Anthropic adds a playground to craft high-quality prompts
Anthropic Console now offers a built-in prompt generator powered by Claude 3.5 Sonnet. You describe your task and Claude generates a high-quality prompt for you. You can also use Claude’s new test case generation feature to generate input variables for your prompt and run the prompt to see Claude’s response.
Moreover, with the new Evaluate feature you can do testing prompts against a range of real-world inputs directly in the Console instead of manually managing tests across spreadsheets or code. Anthropi chas also added a feature to compare the outputs of two or more prompts side by side.
Google’s digital reconstruction of human brain with AI
Google researchers have completed the largest-ever AI-assisted digital reconstruction of human brain. They unveiled the most detailed map of the human brain yet of just 1 cubic millimeter of brain tissue (size of half a grain of rice) but at high resolution to show individual neurons and their connections.
Now, the team is working to map a mouse’s brain because it looks exactly like a miniature version of a human brain. This may help solve mysteries about our minds that have eluded us since our beginnings.
Enjoying the weekly updates?
Refer your pals to subscribe to our newsletter and get exclusive access to 400+ game-changing AI tools.
When you use the referral link above or the “Share” button on any post, you'll get the credit for any new subscribers. All you need to do is send the link via text or email or share it on social media with friends.
OpenAI teams up with Los Alamos Lab to advance bioscience research
This first-of-its-kind partnership will assess how powerful models like GPT-4o can perform tasks in a physical lab setting using vision and voice by conducting biological safety evaluations. The evaluations will be conducted on standard laboratory experimental tasks, such as cell transformation, cell culture, and cell separation.
According to OpenAI, the upcoming partnership will extend its previous bioscience work into new dimensions, including the incorporation of ‘wet lab techniques’ and ‘multiple modalities”.
The partnership will quantify and assess how these models can upskill professionals in performing real-world biological tasks.
China dominates global gen AI adoption
According to a new survey of industries such as banking, insurance, healthcare, telecommunications, manufacturing, retail, and energy, China has emerged as a global leader in gen AI adoption.
Here are some noteworthy findings:
Among the 1,600 decision-makers, 83% of Chinese respondents stated that they use gen AI, higher than 16 other countries and regions participating in the survey.
A report by the United Nations WIPO highlighted that China had filed more than 38,000 patents between 2014 and 2023.
China has also established a domestic gen AI industry with the help of tech giants like ByteDance and startups like Zhipu.
Samsung reveals new AI wearables at ‘Unpacked 2024’
Samsung unveiled advanced AI wearables at the Unpacked 2024 event, including the Samsung Galaxy Ring, AI-infused foldable smartphones, Galaxy Watch 7, and Galaxy Watch Ultra.
Take a look at all of Samsung’s Unpacked 2024 in 12 minutes!
New Samsung Galaxy Ring features include:
A seven-day battery life, along with 24/7 health monitoring.
It also offers users a sleep score based on tracking metrics like movement, heart rate, and respiration.
It also tracks the sleep cycles of users based on their skin temperature.
New features of foldable AI smartphones include:
Sketch-to-image
Note Assist
Interpreter and Live Translate
Built-in integration for the Google Gemini app
AI-powered ProVisual Engine
The Galaxy Watch 7 and Galaxy Watch Ultra also boast features like AI-health monitoring, FDA-approved sleep apnea detection, diabetes tracking, and more, ushering Samsung into a new age of wearable revolution.
Google’s AI gets a body: DeepMind’s office “helper” robot
A tall, wheeled “helper” robot is now roaming the halls of Google's California office, thanks to its AI model. Powered with Gemini 1.5 Pro’s 1 million token context length, this robot assistant can use human instructions, video tours, and common sense reasoning to successfully navigate a space.
In a new research paper outlining the experiment, the researchers claim the robot proved to be up to 90% reliable at navigating, even with tricky commands such as “Where did I leave my coaster?” DeepMind’s algorithm, combined with the Gemini model, generates specific actions for the robot to take, such as turning, in response to commands and what it sees in front of it.
OpenAI’s new tracker for its LLMs’ progress toward AGI
OpenAI has created an internal scale to track its LLMs' progress toward artificial general intelligence (AGI).
Chatbots, like ChatGPT, are at Level 1. OpenAI claims it is nearing Level 2, which is defined as a system that can solve basic problems at the level of a person with a PhD.
Level 3 refers to AI agents capable of taking actions on a user’s behalf.
Level 4 involves AI that can create new innovations.
Level 5, the final step to achieving AGI, is AI that can perform the work of entire organizations of people.
This new grading scale is still under development.
AWS gets a blitz of new AI updates
At the AWS New York Summit, AWS announced a wide range of capabilities for customers to tailor generative AI to their needs and realize the benefits of generative AI faster.
Amazon Q Apps is now generally available. Users simply describe the application they want in a prompt and Amazon Q instantly generates it.
With new features in Amazon Bedrock, AWS is making it easier to leverage your data, supercharge agents, and quickly, securely, and responsibly deploy generative AI into production.
It also announced new partnerships with innovators like Scale AI to help you customize your applications quickly and easily.
That's all for now!
Subscribe to The AI Edge and gain exclusive access to content enjoyed by professionals from Moody’s, Vonage, Voya, WEHI, Cox, INSEAD, and other esteemed organizations.
Thanks for reading, and see you on Monday. 😊