- The Rundown AI
- Posts
- Chinese AI bests GPT-4T
Sign Up | Advertise | Tools | AI University
Welcome, AI enthusiasts.
The U.S. has been confident it’s winning the global battle for AI supremacy — but a Chinese model may have just shown the battle is far from over.
SenseTime’s new release just bested GPT-4T across many key benchmarks, proving the country’s AI progress is not to be underestimated. Let’s explore…
In today’s AI rundown:
Chinese AI model bests GPT-4 Turbo
Sanctuary releases next-gen Phoenix
Access Llama 3 on your Phone
Synthesia unveils ‘Expressive Avatars’
6 new AI tools & 4 new AI jobs
More AI & tech news
Read time: 3 minutes
LATEST DEVELOPMENTS
SENSETIME
Image source: SenseTime
The Rundown: Chinese tech firm SenseTime just launched SenseNova 5.0, a major update to its large model — featuring capabilities that beat GPT-4 Turbo across nearly all key benchmarks.
The details:
The ~600B parameter model features a 200k context window and was trained on over 10TB of largely synthetic data.
The model shows major advancements in knowledge, math, reasoning, and coding — outperforming top models like GPT-4T across major benchmarks.
According to the company, a text-to-video model featuring stylistic consistency and presets will also be available soon.
Why it matters: SenseNova 5.0’s insane scores across key benchmarks suggest that China’s AI might be much closer to competing with the U.S. than previously realized. But without public availability to test the validity of the models’ capabilities, there are still many unknowns.
TOGETHER WITH HUBSPOT
The Rundown: Discover how to harness the transformative power of AI in your business with HubSpot's Free AI Adoption Checklist — your roadmap to smarter workflows and unparalleled efficiency.
This comprehensive guide covers:
Identifying AI opportunities in work processes
Setting clear, measurable goals for success
Empowering your team with custom AI training
Ensuring robust AI data security
Download the free checklist now and start revolutionizing your workplace with the power of AI.
SANCTUARY AI
Image source: Sanctuary AI
The Rundown:Sanctuary AI just released the seventh-gen of its Phoenix humanoid robot, featuring substantial new enhancements to both the physical design, AI systems, and training process.
The details:
Gen 7 brings longer operational times, streamlined manufacturing, and significantly lower production costs.
The robot’s body has been overhauled with improved dexterity, durability, visual perception, and sensing capabilities.
Phoenix’s ‘Carbon’ AI system can now master new tasks in under a day, a major leap from earlier generations' weeks-long training times.
Why it matters: Humanoid robots continue to advance at lightning-fast speeds — with new models inching closer and closer to sci-fi capabilities. With rapidly accelerating AI models coming alongside improved hardware, it’s only a matter of time before these robots are integrated into the real world.
AI TRAINING
The Rundown: Llama 3 is Meta’s most powerful open LLM, released just last week. In this tutorial, you’ll learn how to access it on your phone for free using the HuggingChat app.
Step-by-step:
Download the HuggingChat app from the App Store or access it directly in the browser.
Sign in using a Hugging Face account or Apple ID.
Access the settings menu by tapping on the gear icon and select "meta-llama/Meta-Llama-3-70B-Instruct”
Start prompting by typing your question or input into the chat interface and hitting send!
PRESENTED BY BRILLIANT
The Rundown: Brilliant breaks down complex AI concepts into digestible, interactive lessons — empowering you to understand the technology currently revolutionizing our world.
Brilliant’s expert-designed courses feature:
Bite-sized lessons that make complex AI concepts accessible
Engaging challenges with real-world applications
Personalized learning path and real-time feedback
Start your 30-day free trial and join a global community of 10M learners. Plus, unlock 20% off, exclusive to Rundown readers.
SYNTHESIA
Image source: Synthesia
The Rundown: AI startup Synthesia just unveiled a new generation of “Expressive Avatars” that can convey a wide range of human emotions, bringing new levels of hyper-realism to the video outputs.
The details:
The avatars use a custom EXPRESS-1 model to understand emotional sentiment — adjusting tone, facial expressions, and body language.
The upgrade unlocks more emotionally driven use cases, such as promotional videos, customer support, and healthcare content.
The company said the ‘digital actors’ will also soon be capable of moving in 3D spaces and communicating with other avatars.
Why it matters: The line between real and digital has officially been completely blurred. Synthesia’s AI avatars can reshape how employees connect, interact, and learn — and we’re not far off from wild concepts like 3D digital assistants, AI tour guides, full AI avatar presentations, and more.
NEW TOOLS & JOBS
🗣️ PeddleBot - Perfect your sales pitch with realistic AI roleplays
🧠 Web3 Summary -Simplify crypto projects or web3 research
🎓 Kaiden - Create lesson plans 10x faster
✅ BeeDone - Get things done with gamified AI
🎨 Recraft - Create and edit images with consistent brand style
🌴 Layla - AI personal trip planner
🚀 AE Studio - Tech Lead & PM
📊 The Rundown - Head of Finance
🤖 Sanctuary AI - Senior Software Engineer
👨🔬 OpenAI - Data Scientist, Incident Response
QUICK HITS
Elon Musk is escalating his legal battle with OpenAI, subpoenaing ex-board member Helen Toner demanding documents and communications relating to the November drama surrounding Sam Altman.
A Huawei-led group of Chinese chipmakers is working to produce high-bandwidth AI chips by 2026, hoping to lessen reliance on Western production and circumvent export restrictions.
xAI is reportedly close to securing $6B in funding, which would boost the Elon Musk venture’s value to $18B.
Microsoft reported a significant revenue increase in its 2024 Q1 earnings, led by strong growth in AI and cloud computing alongside major AI investments.
Adobe introduced VideoGigaGAN, a new feature coming to its platforms capable of upscaling videos 8x with new levels of sharpness and minimal quality loss.
Sam Altman’s Worldcoin is reportedly aiming to collaborate with OpenAI, coming despite regulatory concerns and scrutiny surrounding its eye-scanning orbs.
THAT’S A WRAP
SPONSOR US
Get your product in front of over 550k+ AI enthusiasts
Our newsletter is read by thousands of tech professionals, investors, engineers, managers, and business owners around the world. Get in touch today.
FEEDBACK
How would you rate today's newsletter?Vote below to help us improve the newsletter for you. |
If you have specific feedback or anything interesting you’d like to share, please let us know by replying to this email.
Join the conversation