• The AI Pulse
  • Posts
  • šŸ¤– OpenAI Accidentally Deletes Evidence in Copyright Lawsuit

šŸ¤– OpenAI Accidentally Deletes Evidence in Copyright Lawsuit

PLUS: AI Detects Wildfires Before They Spread, DeepSeek Unveils Powerful ā€œReasoning AIā€

Welcome back AI enthusiasts!

In todayā€™s Daily Report:

  • šŸ—‘OpenAI Accidentally Deletes Evidence in Copyright Lawsuit

  • šŸ§‘ā€šŸš’AI Detects Wildfires Before They Spread

  • āš™ļøDeepSeek Unveils Powerful ā€œReasoning AIā€

  • šŸ› Trending Tools

  • šŸ’°Funding Frontlines

  • šŸ’¼Whoā€™s Hiring?

Read Time: 3 minutes

šŸ—žRECENT NEWS

OPENAI

šŸ—‘OpenAI Accidentally Deletes Evidence in Copyright Lawsuit

Image Source: Canvaā€™s AI Image Generators/Magic Media

Lawyers for The New York Times (NYT) say OpenAI engineers accidentally deleted crucial training datasets that couldā€™ve supported their case.

Key Details:
  • The NYT filed a lawsuit against OpenAI for illegally leveraging millions of copyrighted articles without permission to train AI models like ChatGPT.

  • Earlier this year, OpenAI agreed to provide two Virtual Machines for lawyers to conduct searches for copyrighted content in OpenAIā€™s training datasets.

  • A Virtual Machine is a digitized version of a physical computer that stores data, runs applications, and connects to networks.

  • The lawyers outlined how they spent over 150 hours searching the training datasets. Then, OpenAI engineers erased all the data stored on one of the Virtual Machines.

  • OpenAI tried to recover the training datasets and was somewhat successful. However, the folder formats and file names were ā€œirretrievablyā€ lost.

  • So, the recovered training datasets canā€™t be used to determine where the copyrighted articles ā€œwere used to build the AI models.ā€

Why Itā€™s Important:
  • OpenAI maintains that training AI models using publicly available internet materials falls under Transformative Use: Adding new expression, meaning, or messaging to the original work.

  • OpenAI believes the AI models ā€œlearnā€ how to generate human-like text from the articles rather than directly copying them, meaning OpenAI isnā€™t required to license the articles.

šŸ©ŗ PULSE CHECK

Do you think OpenAIā€™s situation falls under Transformative Use?

Vote Below to View Live Results

Login or Subscribe to participate in polls.

GOOGLE RESEARCH

šŸ§‘ā€šŸš’AI Detects Wildfires Before They Spread

Google Research developed ā€œFireSat,ā€ an AI framework that detects and tracks wildfires with high-resolution satellite imagery for the U.S. Forest Service (USFS).

Key Details:
  • ā€œFireSatā€ relies on a network of high-resolution satellites to capture images of the Earthā€™s surface every 20 minutes.

  • Then, it uses AI to compare the high-resolution satellite imagery with thousands of previous pictures of the same location to highlight any alarming changes.

  • ā€œFireSatā€ can detect wildfires roughly the size of a classroom. If it detects anything, it sends real-time alerts to emergency responders outlining the wildfireā€™s size, location, and likelihood of spreading.

Why Itā€™s Important:
  • The USFS currently uses low-resolution satellite images that update a few times a day, making it difficult to ā€œdetect fires smaller than a soccer field.ā€

  • ā€œFireSatā€ helps USFS make informed decisions about resource allocation and evacuation plans to protect local communities and wildlife ecosystems.

AI RESEARCH

āš™ļøDeepSeek Unveils Powerful ā€œReasoning AIā€

Image Source: DeepSeek API Docs/ā€œDeepSeek-R1-Lite-Preview is Now Live: Unleashing Supercharged Reasoning Power!ā€/Screenshot

Chinese research firm DeepSeek just launched DeepSeek-R1-Lite-Preview, a powerful AI model that leverages a Reasoning Engine to excel in reasoning tasks.

Everyone is trying to make significant strides in developing AI models with advanced reasoning capabilities that solve complex problems using ā€œa series of intermediate steps.ā€

This process requires a Reasoning Engine, which creates specific rules and training datasets that enable AI models to mimic a humanā€™s decision-making capabilities and ability to apply logic to unseen scenarios.

In other words, a Reasoning Engine allows AI models to analyze information, draw conclusions, and solve problems based on the knowledge itā€™s been given, similar to how we might reason through a situation in our heads.

DeepSeek-R1-Lite-Previewā€™s initial scores on the Math Word Problem Solving (MATH) benchmark, which covers 12,500 algebra, geometry, and pre-calculus problems, beat OpenAIā€™s ā€œOpenAI o1-preview.ā€

šŸ› TRENDING TOOLS

šŸ§ Dokko makes your chatbots smarter with less effort.

šŸ§©I/O Crossword creates awesome crossword puzzles.

šŸ“Šthisorthis.ai does side-by-side AI model comparisons.

šŸ“–ReadPartner summarizes websites, documents, and videos.

šŸ›‹CoolAIid is for architects, designers, and builders looking to speed up their workflow.

šŸ”®Browse our always Up-To-Date AI Tools Database.

šŸ’°FUNDING FRONTLINES

  • Cyera secures a $300M Series D to offer data security solutions for AI models.

  • Wordware raises a $30M Seed Round to build, iterate, and deploy AI models for you.

  • Blue Bear Capital lands $160M to back AI founders in climate, energy, and industry sectors.

šŸ’¼WHOā€™S HIRING?

  • X (Palo Alto, CA): Engineering Intern, Summer 2025

  • Wisk (Mountain View, CA): Software Engineering Intern, Summer 2025

  • Cleric (San Francisco, CA): Software Engineering Intern, Summer 2025

  • Parsons (Boulder, CO): Machine Learning {ML} Engineer Intern, Summer 2025

  • Nvidia (Austin, TX): Software Engineer, Ethernet Switching, New College Grad 2024

šŸ¤–PROMPT OF THE DAY

MARKET RESEARCH

šŸ”Research Your Competitors

Act like a market research consultant. Provide me with a detailed analysis of my top five competitors in [Specific Industry] with [Target Audience] and [Market Conditions].

Specific Industry = [Insert Here]

Target Audience = [Insert Here]

Market Conditions = [Insert Here]

šŸ“’FINAL NOTE

FEEDBACK

How would you rate todayā€™s email?

It helps us improve the content for you!

Login or Subscribe to participate in polls.

ā¤ļøTAIP Review of The Day

ā€œThis newsletter is so darn reliable.ā€

-Dan (1ļøāƒ£ šŸ‘Nailed it!)
REFER & EARN

šŸŽ‰Your Friends Learn, You Earn!

You currently have 0 referrals, only 1 away from receiving āš™ļøUltimate Prompt Engineering Guide.

Refer 3 friends to learn how to šŸ‘·ā€ā™€ļøBuild Custom Versions of OpenAIā€™s ChatGPT.

Reply

or to participate.