Ethan B. Holland

Over 62,700 manually organized AI links and counting

This week’s cover is inspired by Meta’s release of Segment Anything Model 2.  Segmentation is essentially object tracking.  SAM 2 is an open source model that can handle live video segmentation.   The MidJourney prompt is “wide shot, a herd of llamas wandering through a music festival --chaos 20 --ar 4:3 --style raw --personalize 2nvgnvq --v 6.1”.  The image was then run through Meta’s Segment Anything Model 1, which identifies objects in an image.  The font is Futura.

AI News #44: Week Ending 08/02/2024 with Executive Summary, Top 83 Links, and Helpful Visuals

This Week’s Covers

This week’s cover is inspired by Meta’s release of Segment Anything Model 2.  Segmentation is essentially object tracking.  SAM 2 is an open source model that can handle live video segmentation.   The MidJourney prompt is “wide shot, a herd of llamas wandering through a music festival –chaos 20 –ar 4:3 –style raw –personalize 2nvgnvq –v 6.1”.  The image was then run through Meta’s Segment Anything Model 1, which identifies objects in an image.  The font is Futura.

The covers this week reference Meta, by including a Llama in every image. Llama is Meta’s flagship open source model.  However, the images were all created using FLUX.1 [dev], an open source model launched this week by Black Forest Labs.  FLUX excels at complex prompts with multiple elements as well as text.  In fact, FLUX is so good, I don’t think I’ll make a cover again with Photoshop, as a test to keep everything with the generative AI world. Just look at the category covers for Meta, VR, and Business, below.

Executive Summaries: Week Ending 08/02/2024

Meta Introduces Advanced Object Identification and Tracking in Videos and Images

Meta has unveiled the Segment Anything Model 2 (SAM 2), a groundbreaking AI model that can instantly select and track objects in both images and videos in real time. It overcomes challenges such as fast object movement and visual obstructions, making it a powerful tool for video editing, mixed reality experiences, and autonomous systems (like robots, drones, and cars).  The software is open-source and can be downloaded for free and run locally on a laptop.

Here are some examples of my kids’ dance team and the Delmarva Sports Network’s play of the week: https://ethanbholland.com/2024/09/14/trying-metas-segment-anything-2-sam2-demo/ 

NVIDIA’s “Project GR00T” Accelerates Robotic Learning With VR and Simulations

The project combines VR tech like Apple Vision Pro to capture remote human operation of real robots, then multiplies that data by over 1000x in simulations.  Using Vision Pro, a human operator controls a humanoid robot in real-time, experiencing teleoperation as if they were inside the robot itself.  Next, NVIDIA’s simulation framework expands the collected data by varying the visual environment. For example, a robot performing a task in one physical kitchen can now appear to perform the same task in countless simulated kitchens with different layouts and objects. MimicGen further enhances the data by creating a wide range of new motion trajectories for the robot. It filters out unsuccessful attempts, generating a vast dataset from a single human demonstration. By using a synthetic data pipeline, Project GR00T effectively breaks the limitations of traditional teleoperation, which is restricted by the number of hours a robot can operate in the real world.  More from Jim Fan on Twitter.

Other robotics headlines: 

Zuckerberg says Meta will need 10x more computing power to train Llama 4 than Llama 3

Meta is bracing for a substantial investment in computing, with Mark Zuckerberg revealing that training Llama 4 will require 10x the computing power than Llama 3. This eye-popping increase highlights the massive resources needed to stay competitive. Despite the cost, Zuckerberg stressed the importance of building infrastructure now rather than risk falling behind. Meta’s largest model to date, Llama 3.1 with 405 billion parameters, is a testament to this trend.  Meta’s CFO Susan Li noted that the expanded AI efforts will raise capital expenditures in 2025, with significant investments in data centers and servers.  Almost every AI company is diving into spending on data centers… hoping the investment pays off later.  It’s worth listening to the interview of former OpenAI employee Leopod Ashenbrenner for context on just how crazy it will get if it continues – we’re talking about an entire state’s worth of electricity consumption.

Black Forest Labs Launches FLUX.1 Image Models, Rivals or Bests MidJourney 

It’s free.  It can generate solid text within images. It’s good at complicated requests.
Black Forest Labs, a newly launched AI company rooted in the generative AI research community (Stable Diffusion alums), unveiled its FLUX.1 suite of models, rivaling the best of text-to-image generation tools. FLUX.1 introduces three variants—[pro], [dev], and [schnell]—offering state-of-the-art image quality, diversity, and speed. In many tests, the models beat major competitors like Midjourney v6.0 and DALL·E 3 (HD) in visual quality and prompt adherence.  Dev and Schell are open source.

Salesforce Introduces Multimodal “Einstein Sales Agents” for AI-Driven Lead Generation

Salesforce has launched Einstein Sales Agents, an AI-powered digital sales assistant designed to prospect, nurture, and develop leads around the clock. These agents leverage a company’s CRM data and custom sales strategies to offer personalized, contextual engagement across channels. Built on the Einstein 1 Platform, they integrate seamlessly with Salesforce Sales Cloud, automating pipeline nurturing and handing off qualified leads to human sales reps. Backed by the Einstein Trust Layer (almost a satire at this point with the names, LOL), the multimodal (?!) agents support natural language, images, video, and audio.  This means the AI agents can engage in conversations not only through texting but also through spoken conversations and even video.  According to Salesforce, sales teams using AI saw an 83% revenue boost compared to 66% revenue increases without.  

Related post rom Reddit: “’I work in a call center, so do 17 million people around the world. 95% of costs associated with call centers are due to staffing, and its a 300 billion dollar industry. I’ve recently been made aware of a demo with a new call center AI agent and I immediately started looking for a different career. The bad part is, my job and everyone else’s in this industry is over, the good part is, you will never be put on hold again, the AI agent will likely already know who you are when you call, have predicted the issue you are most likely having, and be ready to offer the solution to that issue. Most people will not be able to tell they are talking to an AI Agent. 2 years max for the call centers to train these friendly, helpful AI’s on every possible scenario, implement them, and fire everyone. Ai will take all our jobs. Fine by me – working in a call center sucks.”

US Department of Commerce Supports Open-Source AI Models 

This is a big thing to note in an election year, and runs counter to a lot of tech bro claims.

The debate around open-source AI models has taken a significant turn as the U.S. Department of Commerce, through the NTIA, issued policy recommendations backing “open-weight” models. The report advocates for making key AI components widely accessible to developers, smaller companies, researchers, and nonprofits, promoting innovation beyond the closed-source commercial systems used by large corporations. This move reinforces the argument that open models benefit the public and should be encouraged, while the focus remains on monitoring the risks associated with AI, regardless of whether it’s open-source or proprietary.

OpenAI Launches Voice-Enabled ChatGPT for Limited Users (I got it!)

OpenAI has introduced a new voice-enabled ChatGPT bot, available today for select users and set to reach all ChatGPT Plus subscribers this fall. The chatbot, powered by the new GPT-4o model, features advanced voice recognition, real-time responses, and emotional nuance, positioning it as a major player among AI assistants like Siri and Alexa. Initially delayed to improve safety features, including filters to prevent unwanted content and deepfakes, the release marks OpenAI’s next step toward multimodal AI. Future updates may include video and screen sharing capabilities to enhance user interactions.

Google’s Small Gemma Models Deliver Incredible Performance And Enhanced Transparency

Google’s Gemma 2 series excels in efficiency, with the 9B and 27B parameter models performing better on tasks like text generation and understanding than larger models like GPT 3.5. Their standout feature is interpretability through Gemma Scope, which offers transparency by allowing developers to see how the models make decisions. This boosts trust and helps identify potential biases or issues, fostering responsible AI use. Additionally, the ShieldGemma classifier ensures safety by detecting harmful content, further enhancing the models’ responsible deployment.  The key takeaway here is smaller open source models are getting really good – so use in small drones, robots, toasters, etc… is getting closer and closer.

DeepMind Snuck A New AI Model To the #1 Spot With No Announcement 

Gemini 1.5 Pro Achieves New Heights in AI Performance.  Google has released the experimental version of Gemini 1.5 Pro (0801) for testing in Google AI Studio and via the Gemini API. This model has already claimed the #1 spot on multiple leaderboards, surpassing GPT-4 and Claude-3.5 in overall performance.   According to tests by Lmsys, Gemini excels in multilingual tasks, math, coding, and handling hard prompts. Despite its quiet launch, the model is making waves for its robust technical capabilities and wide-ranging proficiency.  

Apple Delays AI Features in iOS 18 Release

Apple’s new AI system, Apple Intelligence, will miss the initial iOS 18 launch in September, arriving later in October updates due to bug fixes. Developers will begin testing AI features, such as Siri improvements and integration with OpenAI’s ChatGPT, via iOS 18.1 betas this week. However, full AI functionality will roll out gradually through 2024 and 2025. Apple is banking on these AI enhancements to boost sales of the iPhone 16, as the device’s design remains largely unchanged.

Runway announces Gen-3 Alpha Image to Video

Runway’s update lets users start video generation from any image, either as a standalone frame or combined with a text prompt. The new Gen-3 Alpha Turbo model can produce videos 7x faster than before while maintaining quality and will be available to free users soon at a lower price. These examples do the best job explaining it.

Morgan Stanley Expands AI Tools with Custom In-House Solutions

Morgan Stanley recently deployed its second in-house generative AI tool, AI “Morgan Stanley Debrief”, which summarizes video meetings and drafts follow-up emails. Built in collaboration with OpenAI, this tool follows a previous AI assistant designed to streamline research access for financial advisers. Morgan Stanley’s focus on custom-built AI, rather than off-the-shelf solutions, reflects a growing trend in the financial sector, where firms are partnering with model builders like OpenAI to ensure “seamless integration with existing workflows” (Buzzword BINGO,.. but finance is one place this makes sense for security). 

Intel Announces Layoffs Amid Struggles to Cut Costs and Boost Growth

Intel is laying off over 15,000 employees, representing more than 15% of its workforce, as part of a $10 billion cost-cutting plan aimed at recovery by 2025. The layoffs come after the company reported a $1.6 billion loss in Q2 2024, significantly higher than the previous quarter. Most losses stem from its Foundry business, which has been investing heavily in new technologies, but its core PC and server businesses remain profitable. While Intel struggles to compete with AI chipmakers like Nvidia, it expects modest improvements with its upcoming AI chip, Lunar Lake. Investors remain wary, with Intel’s stock performance lagging, and the company plans further restructuring and spending reductions to stabilize its future.

TikTok’s $20 Million Monthly AI Spending Fuels Microsoft’s OpenAI Reselling Revenue

TikTok has been spending $20 million per month on OpenAI models through Microsoft, reportedly making up about 25% of Microsoft’s AI revenue. This heavy reliance on Microsoft is risky for both companies, as TikTok’s parent, ByteDance, is working to develop its own AI models. If successful, TikTok may no longer need OpenAI’s technology, potentially cutting Microsoft out. This situation mirrors Nvidia’s risk, as tech giants use its hardware to build their own AI chips. As Microsoft’s AI Cloud sales have underperformed, this reliance on key customers like TikTok is critical for their revenue.

Amazon Unveils New AI Chip with 50% Boost in Performance

Amazon has introduced a new AI chip designed to rival NVIDIA’s offerings, with tests showing a performance improvement of 40-50%. The chip, developed by Amazon’s Annapurna Labs, aims to reduce reliance on expensive NVIDIA processors for AWS AI cloud services, offering customers a more cost-effective solution for handling complex data tasks. While Amazon is relatively new to AI chip design, it leads in non-AI processing with its Graviton chip. AWS, which accounts for nearly a third of the cloud computing market, deployed its custom chips during Prime Day to manage a surge in platform activity.

Top 83 Links of The Week

Agents and Copilots

Airtable’s New Cobuilder Unlocks Instant No-Code App Creation – Airtable – https://www.airtable.com/newsroom/product-and-technology/airtables-new-cobuilder-unlocks-instant-no-code-app-creation

“introducing friend. not imaginary. order now at

Friend

https://www.friend.com

Wear This AI Friend Around Your Neck | WIRED

https://www.wired.com/story/friend-ai-pendant

“This is SUPER COOL. @github now will host all the Models directly. So with Codespaces, now there’s a zero-friction path to experiment with the model inference code before dropping it into your own project.

“A nice use case for HuggingChat Assistants: a chatbot for any GitHub repo in a minute” / X

“🤯🤯 You can now create a chatbot on ANY GitHub repo using the Llama 3.1 405B model with @huggingface assistants — FOR FREE! 💰 This is insane! 🚀 Link:

Amazon

Investors in Adept AI will be paid back after Amazon hires startup’s top talent | Semafor

https://www.semafor.com/article/08/02/2024/investors-in-adept-ai-will-be-paid-back-after-amazon-hires-startups-top-talent

Augmented and Virtual Reality (AR/VR) 

3D Modeling

“Unveiled at #SIGGRAPH2024: @WPP is the first to test & use new @NVIDIA NIM microservices to build gen 3D worlds for clients incl @CocaColaCo & @Ford 👇

Other AR/VR News

Activision Releases Call of Duty®: Warzone™ Caldera Data Set for Academic Use

https://blog.activision.com/activision/2024/activision-releases-call-of-duty-warzone-caldera-data-set

“In addition to the new model, we’re also releasing SA-V, a dataset that’s 4.5x larger + has ~53x more annotations than the largest existing video segmentation dataset. We hope this work will help accelerate new computer vision research ➡️

Autonomous Vehicles

Tesla

Elon Musk sets new date for Tesla robotaxi reveal, calls everything beyond autonomy ‘noise’ | TechCrunch

Elon Musk sets new date for Tesla robotaxi reveal, calls everything beyond autonomy ‘noise’

Tesla FSD Beta 10 vs Tesla FSD (Supervised) 12.5 – YouTube

“Tesla FSD Beta 10 vs FSD (Supervised) 12.5 I took 12.5 on the exact same route I recorded FSD Beta 10 handling in September 2021. Zero intervention for 12.5, 8 interventions and 21 nags for 10. Watch in 4K:

Business and Enterprise

“😂 this is so funny… Basically MSFT finance team cannot get their heads around Sam Altman’s doubling revenue every six months projections that drive the data center buildout… So they ended up justifying the buildout over a 15 year period instead of like a 5 year period.” / X

AI Is a Services Revolution – by Rex Woodbury

https://www.digitalnative.tech/p/ai-is-a-services-revolution

Chips, Hardware, and Infrastructure

Education

Heeyo built an AI chatbot to be a billion kids’ interactive tutor and friend | TechCrunch

Heeyo built an AI chatbot to be a billion kids’ interactive tutor and friend

Ethics/Legal/Security

The EU’s AI Act is now in force | TechCrunch

The EU’s AI Act is now in force

Some SF landlords used AI to set rents higher – NBC Bay Area – https://www.nbcbayarea.com/news/local/making-it-in-the-bay/sf-landlords-using-ai/3609211/

OpenAI pledges to give U.S. AI Safety Institute early access to its next model | TechCrunch

OpenAI pledges to give U.S. AI Safety Institute early access to its next model

NIST releases a tool for testing AI model risk | TechCrunch

NIST releases a tool for testing AI model risk

Google

Character.AI CEO Noam Shazeer returns to Google | TechCrunch

Character.AI CEO Noam Shazeer returns to Google

Imagery

“@dr_cintas Leonardo Ai just announced its acquisition by Canva — aiming to accellerate innovation and expand research. It’s been impressive to watch this grow!

“Midjourney V6.1 is now live! V6.1 greatly improves image quality, coherence, text, and comes with brand-new upscaling and personalization models. It’s smarter, faster, clearer, and more beautiful. We hope you enjoy our best model yet <3

“Midjourney released a new update to it’s AI image generator with V6.1 Upgrades include improved image quality, coherence, and text rendering, along with new upscaling and personalization models, offering faster processing and enhanced overall aesthetics

“What’s new in V6.1? – More coherent images (arms, legs, hands, bodies, plants, animals, etc) – Much better image quality (reduced pixel artifacts, enhanced textures, skin, 8bit retro, etc) – More precise, detailed, and correct small image features (eyes, small faces, far away” / X

Locally Run AI Models

“Apple Intelligence Foundation Language Models technical report is out. By @ruomingpang and team. Link:

“Llama3.1 on a raspberry pi Thanks @JustineTunney for Llamafile 😍

“As Apple Intelligence is rolling out to our beta users today, we are proud to present a technical report on our Foundation Language Models that power these features on devices and cloud:  🧵
ttps://twitter.com/ruomingpang/status/1817983627340472642

Apple Intelligence Foundation Language Models – Apple Machine Learning Research

https://machinelearning.apple.com/research/apple-intelligence-foundation-language-models

Microsoft

“😂 this is so funny… Basically MSFT finance team cannot get their heads around Sam Altman’s doubling revenue every six months projections that drive the data center buildout… So they ended up justifying the buildout over a 15 year period instead of like a 5 year period.” / X

Multimodality

“Academics and data scientists are just starting to apply AI to traditionally difficult problems like labeling, analyzing, and categorizing qualitative data It is worth paying attention to the closely related idea in the computer science “LLM-as-a-judge” – lots of good tips there

“🚀Introducing MIA-Bench, a new benchmark for evaluating multimodal large language models on their ability to strictly adhere to complex instructions. Dive into our diverse set of image-prompt pairs and see how state-of-the-art models perform! 🍎GitHub:

MINT-1T: Scaling Open-Source Multimodal Data by 10x: A Multimodal Dataset with One Trillion Tokens – https://blog.salesforceairesearch.com/mint-1t/ 

OpenAI

“a few quick updates about safety at openai: as we said last july, we’re committed to allocating at least 20% of the computing resources to safety efforts across the entire company. our team has been working with the US AI Safety Institute on an agreement where we would provide” / X

Open Source

“Very sensible report from the National Telecommunication and Information Administration, which comes squarely in support of open-weight / open-source AI platforms. As @ClementDelangue says, it’s time to close the debate, abandon innovation-killing bills based on imaginary risks,” / X

White House says no need to restrict ‘open-source’ artificial intelligence — at least for now | AP News – https://apnews.com/article/ai-open-source-white-house-f62009172c46c5003ddd9481aa49f7c3  

Cohere

“Cohere is now part of GitHub Models, offering our state-of-the-art language models to over 100 million developers on @GitHub. Models such as Command R+ can be accessed through the @Azure AI Studio directly to the GitHub Marketplace.” / X

Meta/Llama

“The llama 3 Herd paper is the greatest source of alpha ever released by a big tech company. They literally asked and answered questions methodically about model construction with the compute and talent budget to conclusively determine which paths are dead ends.” / X

Podcasts/YouTube/Op-Eds

AI: Are we in another dot-com bubble? – by Kelvin Mu

https://kelvinmu.substack.com/p/ai-are-we-in-another-dot-com-bubble

“The importance of open source AI in The Economist by @martin_casado and UC Berkeley professor and Chairman of Databricks Ion Stoica.” / X

“if you want to understand how computers work at the hardware level. I’ve seen the first video banger so far. 

Publishing

Reddit CEO says Microsoft needs to pay to search the site – The Verge – https://www.theverge.com/2024/7/31/24210565/reddit-microsoft-anthropic-perplexity-pay-ai-search 

Perplexity brings ads to AI search, revenue-sharing with publishers – Fast Company

https://www.fastcompany.com/91163900/perplexity-selling-ads-ai-search

Why Big Tech Wants to Make AI Cost Nothing

https://dublog.net/blog/commoditize-complement

“😂 even TikTok became a ChatGPT wrapper in the end. And Zhang Yiming was a true pioneer in last generation ML work. Now it’s just a distribution channel. The difficulty collapse comes for us all in the end.” / X

“5 key takeaways: 🚀 LLMs could make 83% of reporter tasks and 76% of editor tasks way more efficient. “It’s important to emphasize that these figures are fundamentally about augmentation rather than automation”. 🧩 Four emerging job clusters to consider: AI-doers, AI-users,” / X

The $600 Billion Digital Ad Business Is Hanging on a Few Words From Google – WSJ

https://www.wsj.com/business/media/the-600-billion-digital-ad-business-is-hanging-on-a-few-words-from-google-6b9fc4f5?mod=tech_lead_pos3

“Alongside OpenAI’s SearchGPT, Microsoft’s Bing is testing a new generative search experience It combines traditional search results with AI-generated content — aiming to reshape how users interact with search engines.

Microsoft says OpenAI is now a competitor in AI and search

https://www.cnbc.com/2024/07/31/microsoft-says-openai-is-now-a-competitor-in-ai-and-search.html

“GenAI could boost the efficiency of a majority of editorial tasks in journalism. But that means recruiting, hiring, and retaining people with a host of AI skills. So, how is the labor market in the news industry evolving post ChatGPT? Read here:

“Someone is going to figure out a clever way to incentivize employees to share their AI productivity gains and lessons learned with their boss. But for now it feels like the right answer is to keep your mouth shut!” / X

Robotics and Embodiment

“LiDAR SLAM is so flipping cool Two different 3D maps captured by robots with an Ouster LiDAR puck + IMU + wheel encoders being merged together into one

Figure

“Figure 02 coming 8/6

“No doubt the biggest company in our lifetime will be within humanoid robotics A humanoid robot company will exceed $1T of value purely from operating in commercial factories 🏭 A robot in the home will bring about a >$20T company 🏡 @Figure_robot is working on both 🦾” / X

“Figure is one of the few very good robotic companies. And it’s no secret that OpenAI has chosen figure to implement their LLM in robotic. The latest videos of how figure is already being used in the factory at BMW is very promising. I’m really excited to see what they’re going to” / X

NVIDIA

“Jensen Huang says the next wave of AI is physical AI, which will require three computer systems to implement: AI, robot and Omniverse

“The coolest use case of the Apple Vision Pro I’ve seen: Using it to operate a humanoid robot in real-time. According to NVIDIA researchers, from the human’s point of view, you feel ‘immersed’ in another body, like in Avatar. Here’s how it works: → Human operators use Apple” / X

“@dr_cintas NVIDIA’s Project GR00T introduced a new approach to scale robot data. It uses the Apple Vision Pro for teleoperation, RoboCasa for environment simulation, and MimicGen for motion, potentially revolutionizing data collection in robotics.

“Exciting updates on Project GR00T! We discover a systematic way to scale up robot data, tackling the most painful pain point in robotics. The idea is simple: human collects demonstration on a real robot, and we multiply that data 1000x or more in simulation. Let’s break it down:

“✨Just announced: Representation agnostics physics simulation. ➡️

NVIDIA Accelerates Humanoid Robotics Development | NVIDIA Newsroom

https://nvidianews.nvidia.com/news/nvidia-accelerates-worldwide-humanoid-robotics-development

Science and Medicine

Neuralink rival Synchron offers thought control with Apple Vision Pro

https://www.cnbc.com/2024/07/30/neuralink-rival-synchron-offers-thought-control-with-apple-vision-pro-.html

Healx, an AI-enabled drug discovery platform for rare diseases, raises $47M | TechCrunch

Healx, an AI-enabled drug discovery platform for rare diseases, raises $47M

“Neuralink rival Synchron’s brain implant now lets people control Apple’s Vision Pro with their minds Synchron announced on Tuesday it has connected its brain implant to the Apple Vision Pro headset in an industry first. The company is building a brain-computer interface that will

AI reprograms glioblastoma cells into dendritic cells for cancer immunotherapy – https://www.news-medical.net/news/20240731/AI-reprograms-glioblastoma-cells-into-dendritic-cells-for-cancer-immunotherapy.aspx

“Academics and data scientists are just starting to apply AI to traditionally difficult problems like labeling, analyzing, and categorizing qualitative data It is worth paying attention to the closely related idea in the computer science “LLM-as-a-judge” – lots of good tips there

“AI is going to save so many lives. Researchers from MIT recently created an AI model that can detect breast cancer 5 years before it develops. How it works: → The model analyzed chromatin images from 560 tissue samples (122 patients), identifying 8 distinct cell states across” / X

AI could predict patient’s risk of bowel cancer returning, study finds | Shropshire Star

https://www.shropshirestar.com/news/uk-news/2024/07/31/ai-could-predict-patients-risk-of-bowel-cancer-returning-study-finds

Video

Black Forest Labs/FLUX

“New kids on the block alert for text-to-image and text-to-video models: @bfl_ml 🚀🎨 The models • FLUX.1 Text-to-Image suite released • Claims strongest ELO score • Model weights on @HuggingFace

“🔥 I am so damn excited to announce the launch of Black Forest Labs. We set ourselves on a mission to advance state-of-the-art, high-quality generative deep learning models for images and video, and make them available to the broadest audience possible. Today, we release FLUX.1” / X

“Blown away by the new @bfl_ml model. Try it out:

apolinario 🌐 on X: “Try out the official FLUX.1 [schnell] demo official demo from the goated new @bfl_ml team, state of the art in 4 steps! 🦶 ▶️ https://t.co/cEbKIxF3gx https://t.co/9KNTQZSvrH” / X – https://twitter.com/multimodalart/status/1819022021168083383

Announcing Black Forest Labs – Black Forest Labs

Announcing Black Forest Labs

The Rest of Video News

Vimeo Announces AI-Powered Video Translation with Authentic Voice Cloning

https://finance.yahoo.com/news/vimeo-announces-ai-powered-video-123000755.html

Introducing Stable Fast 3D: Rapid 3D Asset Generation From Single Images — Stability AI

https://stability.ai/news/introducing-stable-fast-3d

“Stability AI unveiled Stable Video 4D, its a new AI model that can turn single object videos into multiple videos from eight different angles Lots of potential applications for this, including game development, video editing, and virtual reality

Stability AI steps into a new gen AI dimension with Stable Video 4D | VentureBeat

Stability AI steps into a new gen AI dimension with Stable Video 4D

Twitter/X/Grok

“Elon Musk says he expects there to be 10 billion humanoid robots, which will be so profound that it will mark a fundamental milestone in civilization

For Optimus and robot news: see the Robotics/Embodiment category

Tech Papers, Training, and Development

“❤️ the dataviz: Model Release Heatmap by @calebfahlgren 

Model Release Heatmap – a Hugging Face Space by cfahlgren1

https://huggingface.co/spaces/cfahlgren1/model-release-heatmap

Introducing GitHub Models: A new generation of AI engineers building on GitHub – The GitHub Blog

https://github.blog/news-insights/product-news/introducing-github-models/

Credits/Sources


Medium shot.  A modern computer chip factory.  In the background, a large metal box is stampted with the text “Ethan B. Holland” in large letters.   In the foreground, an orange robotic arm is extended toward the camera, holding a bouquet of fresh flowers.  A delicate tag on the flowers reads “Credits” in a clear font. – FLUX.1 [dev]

Most of these weekly links come from just a few prolific oversharing sources. Please follow them, as they work hard to find the news each week and they make it a lot easier for me to compile.

For previous issues, please visit the archives!

MidJourney Prompt: a cozy home library at a beach house, a gorgeous summer afternoon. a humanoid robot reads a book. –chaos 20 –ar 4:3 –style raw –v 6.1 Upscaled with Magnific.ai. GPT suggested Lato for the August/Summer font because it is “a friendly and warm sans-serif font with soft curves, making it feel fresh and approachable.”

Thanks for reading!

Leave a Reply

Trending