📫 AI in the news today: Struggling AI startups, Figure 0 robot, chipmakers

- OpenAI Co-Founders Schulman and Brockman Step Back
https://finance.yahoo.com/news/openai-co-founders-schulman-brockman-010542796.html

- Struggling AI Startups Look for a Bailout from Big Tech
"More exits—either pseudo-acquisitions or real ones—are coming, investors say, as a bubble built by the excitement around generative AI is showing signs of peaking."
https://www.wsj.com/tech/ai/struggling-ai-startups-look-for-a-bailout-from-big-tech-3e635927?mod=rss_Technology

- Did Google Just Pay $2.5 Billion to Hire Character's CEO?
https://www.theinformation.com/articles/did-google-just-pay-2-5-billion-to-hire-characters-ceo

- Figure’s new humanoid robot leverages OpenAI for natural speech conversations
Figure has unveiled its latest humanoid robot, the Figure 02.
The most notable addition this time out arrives by way a longstanding partnership with OpenAI, which helped Figure raise a $675 million Series B back in February, valuing the South Bay firm at $2.6 billion.
https://techcrunch.com/2024/08/06/figures-new-humanoid-robot-leverages-openai-for-natural-speech-conversations/

- World’s Five Leading Chipmakers Have Now Promised U.S. Investment
The Biden administration award up to $450 million in grants to a South Korean chipmaker, SK Hynix, to help build its new chip facility in Indiana
The US now has commitments from all five of the world’s leading-edge semiconductor manufacturers to construct chip plants in theUS with financial assistance from the administration
https://www.nytimes.com/2024/08/06/business/economy/chipmakers-promise-investment.html

Yahoo Finance

OpenAI Co-Founders Schulman and Brockman Step Back

(Bloomberg) -- Greg Brockman, the president of OpenAI and a central figure at the company, said he is is going on leave until the end of the year. At the same time, researcher John Schulman is leaving for rival artificial intelligence startup Anthropic. Most…

Idefics3-Llama is out! 💥💥
Model:
HuggingFaceM4/Idefics3-8B-Llama3

Demo:
HuggingFaceM4/idefics3


It's a multimodal model based on Llama 3.1 that accepts an arbitrary number of interleaved images with text with a huge context window (10k tokens!) ✨

Supported by Hugging Face transformers 🤗

They have also released the dataset, SA-V
This dataset is truly huge, with 190.9K manual annotations and 451.7K automatic! 📊

📄 Paper: https://ai.meta.com/research/publications/sam-2-segment-anything-in-images-and-videos/

📝 Blog: https://ai.meta.com/sam2/

🔗 Demo: https://sam2.metademolab.com/demo

💾 Model Weights: https://github.com/facebookresearch/segment-anything-2/blob/main/checkpoints/download_ckpts.sh

📁 Dataset: https://ai.meta.com/datasets/segment-anything-video-downloads/

Zapier - Automate Like a Pro-Today’s Sponsor
Struggling to keep up with the rapid pace of technology?

ZapConnect 2024 is your gateway to the future of productivity, offering a free half-day virtual event packed with automation insights and tech-savvy strategies.

Learn from 42 automation experts, including Zapier's CEO

Choose from 30 innovative sessions on AI, marketing, and more

Network with 8,000+ fellow productivity enthusiasts in 60+ virtual lounges

Upgrade your tech skills and turn intimidating innovations into powerful allies.

Claim your spot at ZapConnect 2024 and future-proof your productivity!

Check out ZapConnect 2024

🤖How DatoCMS solved tag invalidation with Turso-(https://turso.tech/blog/how-datocms-solved-tag-invalidation-with-turso)

🪐Turso is SOC2 Type II compliant with zero issues (https://turso.tech/blog/turso-achieves-soc2-compliance)

📗Generate/store OpenAI Vector Embeddings in Turso tur.so/QfcmnJq

🔥Save Resend email events to your Turso database tur.so/eVLsNdi

➳ANN search with DiskANN in libSQL tur.so/2TBNQFv

✨Set your teams free with Turso Embedded Replicas tur.so/4GsGAPJ

🏅An environmentally friendly GitHub Actions trick tur.so/QKQQ0H8

📜 Here's why you should try .NET + libSQL tur.so/3OQueK0

AI Tools of the week
🔌 Thunderbit - Deploy AI apps and automations with this no-code solutions for web assistance, summarization, and task automation.

🎓 CourseGenie - Accelerate course creation with auto-generated descriptions, outlines, activities, and assessments.

💡 TextMine - Analyze, manage, and smart-search thousands of documents so that you can make better decisions quickly.

🚀 AI SEO by Leap - Supercharge your website with intelligent SEO tools for content creation, keyword research, and optimization.

🎓 inncivio - Change the way you and your team learn with AI-powered and personalized education.

🖋 Humanizar Texto - Convert AI-generated text into natural, engaging writing that evades detection and plagiarism checks.

📄 MyReport - Automate report and essay creation with AI-powered data collection, writing, and source citation.

🧑‍💻 Modelfuse - Build no-code AI workflows combining your own data with text, image, video, and audio LLMs from top providers.

💬 PSY - AI Therapists - Access confidential mental health support 24/7 through multilingual AI therapist chats.

✨ AI PDF Summarizer(https://theresanaiforthat.com/ai/ai-pdf-summarizer) - Convert lengthy PDFs into digestible insights, with multilingual summaries and document chat features.

📸 Executive Headshots (https://theresanaiforthat.com/ai/executive-headshots)- Transform everyday photos into polished, professional headshots for your digital presence.

💼 Aion(https://theresanaiforthat.com/ai/aion/) - Your AI co-pilot for navigating CEO-level company management, strategy, goal-setting, and market analysis.

📷 AI Cam Lens(https://theresanaiforthat.com/ai/ai-cam-lens) - Instantly analyze your surroundings and get answers to visual questions using your phone's camera.

🎥 AI Image To Video Generator(https://theresanaiforthat.com/ai/ai-image-to-video-generator) - Breathe life into static images by transforming them into captivating videos for social media.

There's An AI For That

AI PDF Summarizer And 2 Other AI Alternatives For Pdf summaries

The AI PDF Summarizer is a free AI-based tool that is designed to quickly and efficiently generate summaries of PDF documents. This tool utilizes AI technology to analyze the content of PDFs and provide a concise and accurate overview of the document. This…

OpenAI's new models gpt-4o-2024-08-06 and gpt-4o-mini begin to support structured output. Unlike the previous JSON mode, the current structured output can not only define the JSON Schema, but also ensure that the output JSON is 100% correct.

Some limitations:

1. Only a part of JSON Schema is allowed: String, Number, Boolean, Object, Array, Enum, anyOf, oneOf and allOf are not supported, which is enough for normal use

2. All fields are required and cannot be optional

3. The nesting cannot exceed 5 levels and cannot exceed 100 attributes

4. Some reserved words cannot be used as attribute names, such as string types cannot use minLength, maxLength, etc.

5. The first API response with a new Schema will generate additional delays, which will be cached later. The general delay will not exceed 10 seconds, but complex Schemas may require up to a minute of preprocessing time

6. Structured output does not prevent all types of model errors. For example, the model may still make mistakes in the values ​​of the JSON object (e.g., a wrong step in a math equation). If errors are made, it is recommended to provide examples in the prompt words or split the task into simpler subtasks.

Original article: https://openai.com/index/introducing-structured-outputs-in-the-api/

Openai

Introducing Structured Outputs in the API

We are introducing Structured Outputs in the API—model outputs now reliably adhere to developer-supplied JSON Schemas.

HubSpot - Turn AI Into Your Personal Assistant
Ready to master the art of AI delegation?

Learn how to turn AI into your personal productivity powerhouse with HubSpot’s highly anticipated AI Task Delegation Playbook. It’s time to optimize your workflow like never before!

Unlock time-saving efficiencies.

Elevate your overall productivity.

Seamlessly streamline your workflow.

Get ready to save time and boost efficiency with easy-to-use templates and calculators.

Don’t miss out—grab your copy of HubSpot’s Playbook and start transforming your workday!

Check out HubSpot’s Playbook—-https://offers.hubspot.com/ai-delegation

Hubspot

Free AI Task Delegation Playbook [Download Now]

This bundle of templates is designed for individuals like you to streamline task delegation and enhance productivity through AI technology.

Breaking News
The latest developments in AI

🚀 Google - A new tiny AI model, Gemma 2 2B, has been released challenging tech giants and even outperforming many of them. Alongside Gemma 2 2B, Google released ShieldGemma, a suite of safety content classifiers, and Gemma Scope, a model interpretability tool.

🎙 OpenAI - ChatGPT's new advanced voice mode impresses early users with diverse capabilities. Demos show it telling stories as an airline pilot with appropriate audio effects. While some accents aren't perfectly native, it handles interruptions well and can laugh or cry during conversations.

🎥 Runway - Gen-3 Alpha Turbo, a faster version of Runway's AI video model, has been unveiled. It's claimed to be 7x faster than the original while maintaining quality. The move is aimed at reducing costs, encouraging increased usage, and staying competitive in the AI video generation market.

Researchers at IISc have come up with a clever fix for AI's diversity problem in image generation. Their "attribute-aware" method is like a gentle nudge for biased algorithms, helping them create more inclusive pictures without a complete system overhaul. In tests, the AI adjusted gender ratios in professional images with impressive finesse. While it's not quite ready for the big leagues, this breakthrough could pave the way for fairer AI art down the road. It's a small step for code, but potentially a giant leap for AI representation. Looks like India's AI hub is making waves in the world of ethical AI! 🧔🏻‍♂️👩🏽https://timesofindia.indiatimes.com/

AI Tools of the week
1.📚 Opal - Transform study sessions with intelligent notes, flashcards, and quizzes that help absorb knowledge like a sponge. https://learnopal.com/

2.🎓 Thesify - Sharpen academic writing with AI tools that hunt references and distill articles, all while preserving scholarly integrity.

3.💡 BrainyBear - Build chatbots that absorb your business knowledge, turning your website into a 24/7 customer support hub.

4.🗨 Meta AI Studio - Create custom AI characters that chat like you, entertain audiences, or give form to your wildest imaginings.

5.💌 InstaBotGPT - Deploy an AI email assistant that mimics and replies in your voice, slashing response times from hours to minutes.

6.📊 ExcelDashboard AI - Turn Excel files into comprehensive dashboards and chat with your data for deeper insights.

7.🎨 AI Color Master - Elevate your designs with color palettes that capture emotions and optimize usability.

8.🔧 Yugo - Boost your web services by seamlessly connecting APIs with cutting-edge AI features.

9.🌐 AITranslator - Break language barriers with swift, expert AI translations for globally-minded businesses.

10.🧑‍💼 QSourcer - Uncover top-tier talent across LinkedIn, GitHub, and StackOverflow with AI-powered Boolean searches.

11.🎥 Video AI One - Streamline video creation by unifying multiple platforms to craft content from scripts, images, or extensions.

12.🏡 RapidREI - Accelerate real estate deals with AI that spots motivated sellers, crunches numbers, and generates smart offers.

13.🎞 Pix AI Video - Breathe life into stories with AI-crafted videos, blending visuals and rich audio to captivate your audience.

14.📖 AI Bedtime Stories for Kids - Craft custom bedtime stories for little ones, turning lights-out into an enchanting adventure.

15.🎨 CannyArt - Awaken your inner manga artist with AI doodles that blossom into anime art, zero artistic skills required.

Opal

Opal - The #1 AI powered study tool

Opal is the #1 AI powered study tool that helps you learn faster and remember more. It's like having a personal tutor in your pocket. Try it now for free!

Guidde - Create how-to video guides with AI
Tired of explaining the same thing over and over again to your colleagues?

It’s time to delegate that work to AI. Guidde is a GPT-powered tool that helps you explain the most complex tasks in seconds with AI-generated documentation.

Share or embed your guide anywhere

Turn boring documentation into stunning visual guides

Save valuable time by creating video documentation 11x faster

Simply click capture on the browser extension and the app will automatically generate step-by-step video guides, complete with visuals, voiceover, and call to actions.

The best part? The extension is 100% free. https://www.guidde.com/

Running Gemma 2 2B at 41.66 tokens/s on my MacBook 💻🚀

- MLX Community's swift conversion
- One-line download from the Hub
- Small yet powerful on-device model

Try it yourself:
mlx-community/google-gemma2-667dca89bc9abbfa34080066


#GemmaAI #OnDeviceAI #MachineLearning https://cdn-uploads.huggingface.co/production/uploads/647f36a8454af0237bd49574/0Hx-zcve_R0aaM01b11BG.mp4

I’ve always wondered why holography hasn’t had much progress since its inception. Imagine what being able to harness and manipulate light with your bare hands in meaningful ways would be like: 3D photorealistic calls, truly immersive workspace. Given that it’s depicted in every futuristic scifi movie, one could not help but vision a future as such. This paper gives a clear overview why:

Turns out it’s incredibly difficult to compute and render photorealistic 3D data in real-time. The author claims immense computational power is needed for high data transmission rates, and compute of large number of phase pixels required for realistic 3D holography. The latest significant breakthrough in holography was 9years ago published in this paper - wherein they were able to achieve mid-air touchable/interactive 3D holography using a Femtosecond laser system considered safer than nanosecond lasers. Quite astounding work: arxiv.org/pdf/1506.06668

Realizing this breakthrough at scale is an unavoidably tempting research endeavor, super exciting especially with recent developments in machine learning and neural network algorithms demonstrating that computer-generated holograms can approach real-time processing. https://cdn-uploads.huggingface.co/production/uploads/6438a9027de34e8ea7e4b257/gjxRB-y9qSITf1iWIxU5F.qt

I can solve the Traveling Salesman Problem using the same methods the scientists used to solve it with 1 qubit, except I do not need quantum computers to do it. I am kind of tired of screaming this from the rooftops at this point. I can create an imaginary probability space, then I can put a bunch of imaginary agents in the imaginary box, and solve real problems in seconds. Problems that would take minutes, hours, or years to solve via other algorithms. Here is a demo of me solving the Traveling Salesman problem using 50 agents to probabilistically sample at once: https://colab.research.google.com/drive/1XplG72nQDO_-2h4DUllERLp0Dr2pI2J2?usp=sharing