Tech Giants Cut Corners to Harvest Data for A.I.

Plus Elon Tries to Raise $3 Billion for xAI

Apr 09, 2024

Today’s Highlights:

📰 News: Tech giants' quest for AI dominance drives aggressive data harvesting tactics, with OpenAI and Meta exploring controversial methods.

💰 Funding: Investors in Talks to Help Elon Musk’s xAI Raise $3 Billion at an $18 Billion Valuation

⚡️ Top News Stories:

1. OpenAI developed Whisper, a speech recognition tool, to transcribe YouTube videos for additional conversational text, aiding in the training of its AI systems like GPT-4.

Despite potential policy violations with YouTube, OpenAI transcribed over one million hours of video content, involving OpenAI’s president, Greg Brockman, in the collection process.
The tech industry's race for AI leadership has led to aggressive data collection strategies, with companies like OpenAI, Google, and Meta pushing ethical and legal boundaries.
Meta discussed acquiring Simon & Schuster to access extensive written content and considered collecting copyrighted data across the internet, risking legal action.
These actions reflect the tech giants' urgent need for vast amounts of digital data to advance AI technology, often sidelining corporate policies and legal considerations.

2. OpenAI's assisted fine-tuning and custom model programs aim to tailor AI models for specific organizational needs, enhancing domain-specific performance.

Assisted fine-tuning leverages advanced techniques and hyperparameters for performance maximization, aiding organizations in setting up efficient training and evaluation systems.
OpenAI also offers fully custom-trained models for organizations with extensive proprietary data, aiming to teach the model new knowledge or behaviors for highly specific use cases.
Harvey, an AI-native legal tool, partnered with OpenAI to create a custom-trained large language model for case law, resulting in an 83% increase in factual responses and a 97% preference rate over GPT-4 by attorneys.

3. FCC report of the Humane AI pin provides insight into the device's sophisticated use of Snapdragon technology for on-device AI projections.

The Humane AI pin, designed to project generative AI chatbot responses onto surfaces like a user's palm, utilizes a Snapdragon 720G processor for its operations.
A detailed FCC report, including a photographic teardown, reveals the internal components of the Humane AI pin, showcasing the device's reliance on Qualcomm's technology.
The Snapdragon 720G processor is chosen for its ability to handle on-device AI tasks with low power consumption, suitable for mobile and wearable devices.

4. Meta announces plans to extend its 'Made with AI' labels to a wider array of content in May in response to evolving AI-generated media challenges.

Meta plans to expand its "Made with AI" labeling to include a broader range of AI-generated video, audio, and image content starting in May, acknowledging the current policy's limitations.
The initiative responds to the need for a more robust system to identify AI-generated and manipulated content, highlighted by incidents like a misleading video of President Biden.
Labels will be applied based on self-disclosure by users, advice from fact-checkers, or Meta's detection of AI content markers.
The policy update follows recommendations from Meta's independent Oversight Board to broaden the scope beyond just videos altering speech.

5. Microsoft AI's new London hub to spearhead advancements in language models and AI tooling, enhancing global collaboration.

6. Spotify enhances music discovery with AI Playlist beta, offering a creative and user-friendly approach to playlist creation.

7. In his annual letter, Jamie Dimon, CEO of JPMorgan Chase places AI at the forefront of strategic considerations, underscoring its expected transformative impact.

💰 Top Funding News:

1. Elon Musk's AI startup, xAI, is in talks to raise $3 billion in a funding round, potentially valuing the company at $18 billion, with interest from Gigafund and Steve Jurvetson, among others.

2. Lumana, a next-generation video security platform leverages AI and a distributed hybrid cloud to unlock an organizations' visual data, has secured $24 million in seed funding from Norwest and S Capital.

3. Archetype AI is developing Newton, an AI model that allows users to communicate with houses, cars, and factories through AI, transforming sensor data into plain language. The startup just raised $13 million in seed funding led by Venrock, with participation from Amazon Industrial Innovation Fund, Hitachi Ventures, and more.

4. Jony Ive, former Apple design lead, and Sam Altman, CEO of OpenAI, are collaborating to start a new company focused on an AI-powered personal device. The project aims to raise at least $1 billion in funding, with potential investors including Thrive Capital and Emerson Collective, founded by Laurene Powell Jobs.