Elon Unveils World’s Most Powerful Supercomputer, Meta’s Llama 3.1 405B: Takes on GPT-4o

Welcome to the DSA Newsletter!?

In this DSA edition, we’re diving into some groundbreaking advancements and hot topics in the tech world:

• The World’s Most Powerful Supercomputer: Discover the cutting-edge machine Elon Musk’s team just unveiled.
• Senators Demand Answers on OpenAI’s Practices: What does this mean for the future of AI?
• Meta’s Llama 3.1 Takes on GPT-4: A head-to-head showdown in the AI landscape.
• Google’s AI-Powered Weather Model: Revolutionizing weather prediction as we know it.
• Adobe’s New AI Features for Photoshop: Redefining creativity with AI.
• MIT’s AI Identifies Breast Cancer Risk: A potential game-changer in healthcare.
• New AI Tools & 4 New AI Jobs: Stay ahead with the latest in AI innovations and career opportunities.
• More AI & Tech News: Your roundup of what’s happening in the tech world.

Are you excited about these AI developments? Do you think AI will continue to revolutionize our daily lives, or are there concerns we need to address first?

Sign up now to stay updated with all the latest AI and tech news!?

Follow Us for more updates and share this newsletter with your network to keep everyone in the loop. Share This Newsletter

The “world’s most powerful” supercomputer

The Brief: Elon Musk and xAI just announced the Memphis Supercluster?—?“the most powerful AI training cluster in the world“, also revealing that Grok 3.0 is planned to be released in December and should be the most powerful AI in the world.

The details:

• Musk tweeted that xAI just launched the “Memphis Supercluster,” using 100,000 Nvidia H100 GPUs, making it “the most powerful AI training cluster in the world.”
• The xAI founder also revealed that Grok 2.0 is done training and will be released soon.
• The supercluster aims to create the “world’s most powerful AI by every metric”, Grok 3.0, by December 2024.
• In a separate tweet yesterday, Musk also revealed that Tesla plans to have humanoid robots in “low production” for internal use next year.

Why it matters: Love him or hate him, the speed at which Elon and the team at xAI operate has been wild to witness. If estimates are accurate, xAI might be on track to create the most powerful AI systems in the world by year’s end?—?solidifying its position as one of the top competitors in the space and not just another AI startup.

Senators demand answers on OpenAI’s practices

The Brief: Five U.S. Senators have just sent a letter to OpenAI CEO Sam Altman, demanding details about the company’s efforts to ensure AI safety following reports of rushed safety testing for GPT-4 Omni.

The details:

• Senators question OpenAI’s safety protocols, citing reports that the company rushed safety testing of GPT-4 Omni to meet a May release date.
• The letter requests OpenAI to make its next foundation model available to U.S. Government agencies for deployment testing, review, analysis, and assessment.
• Lawmakers ask if OpenAI will commit 20% of computing resources to AI safety research, a promise made in July 2023 when announcing the now disbanded “Superalignment team”.

Why it matters: With allegations of rushed safety testing, potential retaliation against whistleblowers, and the disbanding of the “Superalignment team,” OpenAI is under intense scrutiny. This letter also marks a critical moment for the entire AI industry?—?with the potential to lead to stricter government oversight and new industry standards.

Meta’s Llama 3.1 takes on?GPT-4o

The Brief: In case you missed our exclusive deep dive with Mark Zuckerberg yesterday, Meta released Llama 3.1, including it’s long awaited 405B paramater model?—?the first open sourced frontier model that beats top closed models like GPT-4o across several benchmarks.

The details:

• The 405B parameter version of Llama 3.1 matches or exceeds top closed models on several benchmarks.
• Meta is offering open and free weights and code, with a license enabling fine-tuning, distillation into other models, and deployment anywhere.
• Llama 3.1 features a 128k context length, multi-lingual abilities, strong code generation performance, and complex reasoning capabilities.
• For exclusive insights on Llama 3.1, open source, AI agents, and more, read our full deep dive with Mark Zuckerberg here, or watch the full interview here.

Why it matters: Meta’s release of Llama 3.1 405b is a significant moment in AI history because it’s the first time an open-source AI model matches or outperforms top closed AI models like OpenAI’s GPT-4o. By offering a private, customizable alternative to closed AI systems, Meta is enabling anyone to create their own tailored AI.

Google’s AI-powered weather?model

The Brief: Google researchers have developed a new AI-powered weather and climate model called ‘NeuralGCM’ by combining methods of machine learning and neural networks with traditional physics-based modeling.

The details:

• NeuralGCM has proven more accurate than purely machine learning-based models for 1–10 day forecasts and top extended-range models.
• NeuralGCM is up to 100,000 times more efficient than other models for simulating the atmosphere.
• The model is open-source and can run relatively quickly on a laptop, unlike traditional models that require supercomputers.

Why it matters: At up to 100,000 times more efficient than traditional models?—?NeuralGCM could dramatically enhance our ability to simulate complex climate scenarios quickly and accurately. While still a ton of adoption challenges ahead, it’s a big leap forward for more informed climate action and resilience planning.

Adobe’s new AI features for Photoshop

The Brief: Adobe just unveiled major AI-powered updates to Illustrator and Photoshop, leveraging its Firefly AI model to accelerate creative workflows and introduce new generative design capabilities.

The details:

• Illustrator introduces Generative Shape Fill using Firefly Vector AI to add detailed vectors to shapes and create scalable patterns via text prompts.
• Text to Pattern in Illustrator creates scalable, customized vector patterns for designs like wallpapers.
• Photoshop’s new AI-powered Selection Brush Tool and Generate Image function are now generally available.
• Photoshop also gets an enhanced version of its popular Generative Fill for improved sharpness in large images.

Why it matters: These updates could dramatically increase designers’ productivity by automating tedious, time-consuming tasks. We’ve always preached that the best AI products are those embedded into everyday workflows?—?and Adobe is doing just that by putting powerful tech directly into designers’ everyday tools.

AI TRAINING

Transform images in seconds with Freepik?Expand

The Brief: Freepik Expand allows users to seamlessly adapt images when zooming or resizing by filling the gaps with AI.

Step-by-step:

1. Head over to Freepik and upload any image.
2. Choose an aspect ratio and adjust on the canvas.
3. Add a prompt for custom fills, hit “Expand” and select your favorite generated image.
4. Enhance your image with Freepik’s AI tools: Upscaler for sharper details, Retouch for edits, and Reimagine for variations.

Get more AI tutorials →

…

Create 8-second clips with new AI video?model

The Brief: Haiper’s new AI video generation model allows users to generate up to 8-second videos for free and even upscale them to HD resolution.

Step-by-step:

1. Visit the Haiper website and click on “Create video with text”.
2. Craft your video prompt by selecting pre-defined options or inputting your own description.
3. Submit your video generation request.
4. Log in to your account (or create one) to view your AI-generated masterpiece.
5. Enhance your video with HD upscaling if desired, then download and share your creation.

Get more AI tutorials →

…

Tame your AI’s unstructured data

The Brief: DataChain is a new open-source tool designed to address the challenges of managing unstructured data in AI projects?—?providing a framework for developers to handle complex datasets more efficiently.

DataChain offers features for:

• Sourcing, curating, and versioning data from cloud locations at scale
• Integrating metadata from JSON, parquet, and CSV sources
• Accelerating AI model inferences through parallelization
• Simplifying AI output management with Python integration

Ready to supercharge your data workflow? Try DataChain out here and star the repo today!

AI RESEARCH

MIT’s AI identifies breast cancer?risk

The Brief: Researchers from MIT and ETH Zurich have developed an AI model that can identify different stages of ductal carcinoma in situ (DCIS), a type of preinvasive breast tumor, using simple tissue images.

The details:

• The model analyzes chromatin images from 560 tissue samples (122 patients), identifying 8 distinct cell states across DCIS stages.
• It considers both cellular composition and spatial arrangement, revealing that tissue organization is crucial in predicting disease progression.
• Surprisingly, cell states associated with invasive cancer were detected even in seemingly normal tissue.

Why it matters: This AI model could democratize advanced breast cancer diagnostics, offering a cheaper, faster way to assess DCIS risk. While clinical validation is still needed, AI is likely going to work hand-in-hand with pathologists in the near future to catch cancer earlier and more accurately.

NEW TOOLS & JOBS

Trending AI?Tools

• Supermemory?—?An open-sourced personal second brain
• Food Mood?—?Get recipe inspiration by blending two cuisines
• Educato?—?Build an online teaching platform with AI
• Vozo Rewrite & Redub?—?Transform viral videos into new stories from prompts
• UltraPixel?—?Generate up to 4096x4096 px ultra high resolution images
• Rodin Gen-1 3D ControlNets?—?Generate 3D models from a structure reference

Browse more AI tools →

New AI Job Opportunities

• Meta?—?Research Scientist, Computer Vision (PhD)
• Metropolis?—?Director, Technical Sales
• Writer?—?Data Analyst
• Lambda?—?Machine Learning Researcher

Browse more AI jobs →

QUICK HITS

Meta AI unveiled a new ‘Imagine me’ selfie feature, allowing users to create stylized AI-generated selfies based on prompts, alongside expanded language support and integration with Meta’s Llama 3.1 405B model.

OpenAI announced GPT-4o fine-tuning?—?offering 2 million free daily training tokens until September 23 to help developers customize the model for specific applications.

The FTC launched an investigation into eight companies offering AI-powered “surveillance service pricing”, seeking information on how these technologies use consumer data to personalize pricing.

Harvey, an AI platform for legal services, secured a $100M Series C funding round led by Google Ventures, valuing the company at $1.5B, with plans to expand its AI capabilities in the professional services sector.

Cohere launched Rerank 3 Nimble, a new foundation model for enterprise search and RAG systems, offering 3x faster performance than its predecessor while maintaining accuracy and supporting 100+ languages.

Elon Musk proposed a $5 billion Tesla investment into xAI, subject to shareholder and board approval, potentially boosting Tesla’s AI leadership and FSD capabilities.

NVIDIA is reportedly working on a version of its new flagship AI chips for the Chinese market that would be compatible with the restrictions around current U.S. export controls.

Elon Musk revealed in a tweet that Tesla will have humanoid robots working internally next year at Tesla, with high production for other companies expected in 2026.

Cohere announced a new funding round of $500M from investors including Cisco, AMD, and Fujitsu, bringing its new valuation to a whopping $5.5B.

The Monetary Authority of Singapore committed $74.3M to support financial institutions in building capabilities towards quantum computing and artificial intelligence.

Scaled Foundations, a research company building ‘General Robot Intelligence’, announced GRID Beta, a free and cloud-based IDE to develop AI for robots.

Disney Research published a new paper on its famous bipedal robots, explaining the system that combines expressive motions and dynamic mobility to make the robots particularly lifelike.

SPONSOR US

Get your product in front of over 700k+ AI enthusiasts

Our newsletter is read by thousands of tech professionals, investors, engineers, managers, and business owners around the world. Get in touch today.

How would you rate today’s newsletter?

Comment below to help us improve the newsletter for you.