❌ Battle of the Titans: Musk Sues OpenAI

Elon Musk filed a lawsuit against OpenAI and its founders, Sam Altman and Greg Brockman, accusing the company of prioritizing profits over the interests of humanity and breaching the contract.

According to Musk, Altman violated the agreement in 2023 when he released the GPT-4 closed-source code, to which only OpenAI and Microsoft have access:

«OpenAI has been transformed into a closed-source de facto subsidiary of the largest technology company, Microsoft. The company is not just developing but is actually refining AGI to maximize profits for Microsoft, rather than for the benefit of humanity».

Musk demands OpenAI to return to the open-source and nonprofit company model, calling for Altman to be fired and the company to be required to give up all money received as a result of the practices alleged to be unlawful.

He is also seeking a ruling that GPT-4 and other new technologies would be considered AGI and, therefore, outside of Microsoft's license to OpenAI.

OpenAI reacted with a memo that it "categorically disagrees" with Musk’s suit, suggesting the billionaire’s claims have to do with the fact that he’s no longer involved with the company.

😳 The trial promises to be one of the most high-profile in the industry, although the feud between Elon Musk and Sam Altman began long ago.

➡️ Key Milestones:

2015: Musk and Altman start what becomes OpenAI. Musk and others invest $100 million.

2018: Unhappy with OpenAI, Musk wants to take over. Rejected, he left and stopped funding despite promising to invest about $1 billion.

2019: OpenAI becomes for-profit, getting $1 billion from Microsoft.

December 2022: Musk limits OpenAI's access to Twitter data after ChatGPT's launch.

July 2023: Musk starts his AI company, xAI, with $135 million in funding.

November 2023: Musk launches OpenAI competitor Grok, a chatbot with access to the X platform's real-time data. Altman criticizes Grok, calling its sense of humor "cringey."

🤔 Excited for what's next ...

#news @hiaimediaen

Hello! The most exciting updates from the world of AI for the 9th week of 2024 in our Sunday digest.

TOP NEWS

🔝 Will Nvidia become the world's most valuable company? Nvidia has reached a market capitalization of $2 trillion 💵 Сompany is now trailing just behind tech giants Microsoft and Apple. Nvidia is worth more than Russia's or Australia's GDP. Given the explosive growth of generative AI, it could become the most valuable company in the world.

SAVE THIS — IT'S HELPFUL

☄️ SDXL Lightning for real-time image generation. ByteDance, the developer of TikTok, has introduced a high-speed image generation model based on text descriptions. It works as fast as you type the prompt.

AI NEWS

🇫🇷 Mistral AI has launched its new flagship model, Mistral Large. The model reaches top-tier reasoning capabilities, excelling in understanding and generating multilingual text and code. Mistral has also announced a collaboration with Microsoft.

💲 10 facts about generative AI in numbers. For instance, 70% of Zoomers already use AI; by 2026, 90% of content created will be generative, and by 2032, the market size will grow by more than 1 trillion.

TO READ 📚

❌ Battle of the Titans: Musk sues OpenAI. Elon Musk filed a lawsuit against OpenAI and its founders, Sam Altman and Greg Brockman, accusing the company of prioritizing profits over the interests of humanity and breaching the contract.

✝️ 🤖 Father Paolo Benanti is a Franciscan monk and advisor to the Pope on AI. Benanti begins each day with prayers and meditations but also calls himself a "geek" and is inseparable from his Apple Watch.

Happy Spring, friends! 🌷

#AIweek | @hiaimediaen

🪄 How to edit images with AI?

🎨 At the end of last year, Adobe launched a web application to access their AI model Firefly. The service includes a helpful feature called Generative Fill, previously exclusive to Photoshop.

▶️ Generative Fill allows you to edit images using simple text prompts. You can replace backgrounds, clothing, and people; add or remove almost any object. An example is shown in our video 🔼

This tool quickly became popular on social media for creating memes, reimagining movie posters, and redesigning music album covers.

❓ How to use it?

1. Create and log in to your account here;
2. Upload an image;
3. Mark the area you want to edit;
4. Specify what you want to change;
5. Click the "Generate" button ✔️

In a few seconds, the service will offer three image variations. Choose one, and you can continue editing.

⚠️ Note: Adobe restricts the generation of copyrighted images, such as Godzilla, King Kong, or the Microsoft logo, to respect intellectual property rights.

💰 Firefly provides 25 free generation attempts. Paid plans start at €5.5 monthly.

What other tools do you use for image editing?

#manual @hiaimediaen

⚡️ Claude 3: a new AI model from the main competitor of OpenAI

The Anthropic company has announced the Claude 3 model family, which sets new industry benchmarks across various cognitive tasks. The family includes three models in ascending order of capability: Haiku, Sonnet, and Opus.

📊 The most intelligent model, Opus, outperforms its peers on most common evaluation benchmarks for AI systems, including language understanding (MMLU), common knowledge, basic math, and more.

Look at the results of comparative tests with GPT-4 and Gemini 🔼

⭐️ Language tasks capabilities

Claude 3 Opus exhibits near-human levels of comprehension and fluency on complex tasks. All models show increased capabilities in analysis and forecasting, nuanced content creation, code generation, and conversing in non-English languages like Spanish, Japanese, and French.

The context window is 200K tokens.

👀 Image recognition

The Claude 3 models have sophisticated vision capabilities on par with other leading models. They can process various visual formats, including photos, charts, graphs, and technical diagrams.

📎 How to try?

Claude 3 is now available in 159 countries on the claude.ai website.

ℹ️ Anthropic is a startup founded in 2020, focusing on AI development and research. Google and Amazon are among the company's investors.

Source: https://www.anthropic.com/news/claude-3-family

#Claude @hiaimediaen

🍷 AI now understands wine

A team of scientists from Portugal, France, Switzerland, and the US have trained an algorithm to trace wines to their origins. The AI processed information about the chemicals in wines, identified their formulas, and built a scheme with similar wines grouped into clusters. To the scientists' surprise, the clusters’ positions mirrored the locations of estates where the wines were from.

After analyzing 80 wines from 12 vintages harvested from 1990 to 2007 at seven estates in the Bordeaux region of France, the AI "guessed" the place of production with over 99% accuracy and correctly identified the vintage in 50% of cases.

Why does this discovery matter?

📌 Thanks to AI tools, winemakers will have a quick way to detect fraud (fake wines are responsible for €3 billion in lost sales annually in Europe alone).

📌 AI sommeliers can be used for quality control in wine production.

📌 Digital olfaction — the fast-growing area of AI — will be able to enrich the wine experience and help train sommeliers and wine enthusiasts.

We could use machine learning to figure out how to blend wines to optimize quality. Having tools like this would make it much cheaper to make great blends, which would benefit everybody, — Alexandre Pouget, a neuroscience professor at the University of Geneva in Switzerland.

How does the wine industry already use AI?

📌 In Napa Valley, AI-powered sensors installed in the vineyards collect and analyze massive amounts of data on vines, soil moisture, sunlight exposure, temperature, and nutrient levels and suggest irrigation and harvest times.

📌 Some wineries in Bordeaux employ AI-powered robots. They fertilize the soil, remove weeds, and harvest the crops.

📌 Researchers from Denmark and California analyzed thousands of consumer reviews and other data and developed an AI to predict individual wine preferences accurately.

What do you think about the idea of a robot sommelier?

🔥 — I'm all for it
🙈 — some things are better left to humans

#news @hiaimediaen

⏱ How to quickly create a presentation using AI?

Hello everyone! Last year, we wrote about Gamma.app for creating presentations. But there are still many questions on this topic, so we decided to share it with you again.

🎨 Crafting a presentation from scratch can be a time-consuming task. The AI-powered service, Gamma, will help create beautiful, engaging content with none of the formatting and design work. Gamma will generate the layout, choose relevant images, and format slides. You can refine and customize everything.

❓ How to use it?

1⃣ Register on the Gamma website. The site supports a large number of languages.

2⃣ To start, click the Generate button and specify the presentation topic. For example: "History of Memes".

3⃣ Click the Generate Outline button: Gamma will suggest slide titles, theses, and images. Everything can be edited.

4⃣ Choose the presentation design. You can change it later if you wish.

5⃣ Click Generate, and in 30 seconds, the slides are ready.

Here's our "History of Memes" slides 😐

And here's "The Secret Life of Emojis" 🙂

⬇ The Share button allows you to share the presentation link, export it in PDF or PowerPoint format, or provide access for collaboration.

⭐ Bonus: the presentation adapts to the device and will look equally beautiful on a computer and phone.

Summary: if you need to quickly come up with and design an idea into beautiful slides, then Gamma is here to help 🪄

#manual @hiaimediaen

🐶 Lost in Translation: Will AI soon help us talk to animals?

Many animals and birds communicate with each other. Digital bioacoustics and generative AI are capable of capturing, classifying, and analyzing animal sounds. Machine learning can also generate the voices of animals and birds.

🙊 What has already been "translated"?

The Earth Species Project team created AI to separate overlapping animal sounds — such as monkey calls — into individual tracks and filter background noise.

🐳 In the CETI project, zoologists are recording the sounds of sperm whales off the coast of the Dominican Republic. Sperm whales gather in small family groups near the water surface and communicate using Morse-like "codes," each clan having its own "dialect." Having processed 20% of the recordings using AI, the scientists found that the model could predict the remaining 80% of the sperm whale "code".

🐱 Researchers at the University of Lincoln use AI to understand cat expressions. They identified that cats have 276 facial expressions around other cats, which are not the same as the ones they show to humans.

🐱 Researchers from Tel Aviv University used an AI voice-recognition program to analyze 15,000 bat calls. They found that these animals have different communication styles, for example, mothers use a "baby language" with cubs.

🐷 Scientists from Copenhagen managed similarly to "translate" the language of pigs and separate their sounds into happy and displeased.

What are the prospects?

✅ An animal language translator can help us learn about the animals’ moods, health, behavior, and even what they want to "tell" people.

✅ Decoding animal vocalizations will help in studying the evolution of communication skills in humans and the conservation and welfare of endangered species.

➡️ Audio article in Financial Times

Do you need an interpreter from the animal language?

❤️ — yes, of course
🙊 — no, it's obvious

@hiaimediaen