• The AI Pulse
  • Posts
  • šŸ¤– OpenAIā€™s ā€œGPT-o1ā€ Is Smarter Than Most Humans?!

šŸ¤– OpenAIā€™s ā€œGPT-o1ā€ Is Smarter Than Most Humans?!

PLUS: AI Creates Open-World Video Games, LLMs Collaborate For Smarter Solutions

Welcome back AI enthusiasts!

In todayā€™s AI Report:

  • šŸ“ŠOpenAIā€™s ā€œGPT-o1ā€ Is Smarter Than Most Humans?!

  • šŸŽ®AI Creates Open-World Video Games

  • šŸ’¬LLMs Collaborate For Smarter Solutions

  • šŸ› Trending Tools

  • šŸ’°Funding Frontlines

  • šŸ’¼Whoā€™s Hiring?

Read Time: 3 minutes

šŸ—žRECENT NEWS

OPENAI

šŸ“ŠOpenAIā€™s ā€œGPT-o1ā€ Is Smarter Than Most Humans?!

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

OpenAIā€™s ā€œGPT-o1ā€ reportedly scored an IQ of 120 on the Norway Mensa IQ Test, making it the first AI model to surpass the average human IQ level.

Key Details:
  • ā€œGPT-o1ā€ is a new series of AI models designed to spend more time thinking before they respond.

  • ā€œGPT-o1ā€ uses a Chain of Thoughts (CoT) framework to break down complex problems into manageable steps.

  • The Norway Mensa IQ Test is an online exam that requires you to solve 35 visual pattern puzzles within 25 minutes.

  • The visual pattern puzzles get progressively more complex, and you earn points for each correct answer.

  • ā€œGPT-o1ā€ correctly solved 25 out of 35 visual pattern puzzles on a version of the Norway Mensa IQ Test that contained new unpublished problems.

Why Itā€™s Important:
šŸ©ŗ PULSE CHECK

Is ā€œGPT-o1ā€™sā€ IQ of 120 a sign of human intelligence becoming obsolete?

Vote Below to View Live Results

Login or Subscribe to participate in polls.

TENCENT

šŸŽ®AI Creates Open-World Video Games

Chinese multinational technology conglomerate Tencent unveiled ā€œGameGen-O,ā€ an AI model that generates playable open-world video game content.

Key Details:
  • ā€œGameGen-Oā€ leverages a diffusion transformer framework that combines two techniques:

    1. Diffusion Models: Add ā€œnoiseā€ (i.e., random information) to training data and then learn to remove it, creating high-quality patterns.

    2. Transformers: Learn the context and relationships in sequential data (i.e., data arranged in a specific sequence).

  • ā€œGameGen-Oā€™sā€ diffusion transformer framework generates video game assets like characters, environments, and items.

  • It also allows gamers to interact with the generated video game assets.

  • It was trained on ā€œOGameData,ā€ an open-world video game dataset of over 4,000 hours of high-quality gaming footage.

Why Itā€™s Important:
  • ā€œGameGen-Oā€ is one of the worldā€™s first AI-powered game engines to generate playable open-world video game content without manually written software.

  • Tencent is competing with Google DeepMindā€™s ā€œGameNGen,ā€ which creates entirely new video game worlds in real-time tailored to each playerā€™s decisions. It can theoretically generate infinite video game worlds, ensuring thereā€™s always something new to explore.

AI RESEARCH

šŸ’¬LLMs Collaborate For Smarter Solutions

Image Source: Massachusetts Institute of Technology (MIT)/ā€œLearning to Decode Collaboratively with Multiple Language Modelsā€ Screenshot

Researchers from Massachusetts Institute of Technology (MIT) proposed ā€œCo-LLM,ā€ a method to make multiple Large Language Models (LLMs) collaborate to generate responses to user queries.

Have you ever been asked a question you only partially know the answer to? In response, you may ask a colleague with more knowledge on the subject matter to offer a better response to the question.

This collaborative process can also help LLMs improve their accuracy. However, teaching LLMs to recognize when they should collaborate is difficult.

ā€œCo-LLMā€ combines general-purpose base LLMs with a specialized AI model to generate responses to user queries.

To decide when the general-purpose base LLMs need help from the expert AI model, ā€œCo-LLMā€ uses a ā€œswitch variableā€ that measures the competence of each Token within the general-purpose base LLMs responses.

In other words, ā€œCo-LLMā€ reviews each Token within the general-purpose base LLMs responses to determine where the expert AI model can offer a more accurate answer.

Tokens are the smallest units of data used by an AI model to process and generate text. Similarly, we break down sentences into words or characters. You can experiment with OpenAIā€™s GPT-4 tokenizer here.

šŸ› TRENDING TOOLS

šŸ•øFramer publishes your site with AI in seconds.

šŸŽØdeblank instantly generates personalized color palettes.

šŸŖ¶Typefully writes better content to grow your audience faster.

šŸ“±AIphone is an AI-powered phone call app with live language translation.

šŸ«Marshmallow is a community crafted to unite like-minded people through hosting local events.

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

šŸ’°FUNDING FRONTLINES

  • 11x.ai closes a $24M Series A to build digital AI employees.

  • PathPresenter raises a $7.5M Series A to accelerate digital pathology workflows.

  • Operant AI secures a $10M Series A to bring active runtime protection to every layer of the ā€œcloud within the cloud.ā€

šŸ’¼WHOā€™S HIRING?

  • Galaxy (New York, NY): 2025 Marketing Intern, Summer 2025

  • Symbotic (Wilmington, MA): Front End Software Engineer Intern/Co-Op, Summer 2025

  • ServiceNow (Santa Clara, CA): Software Engineering Intern, Digital Technology {DT}, Summer 2025

  • NVIDIA (Santa Clara, CA): Deep Learning {DL} Performance Architect, New College Grad 2024

  • Applied Intuition (Mountain View, CA): Software Engineer, New College Grad 2024/2025

šŸ¤–PROMPT OF THE DAY

PRODUCT OR SERVICE COSTS

šŸ›’Impact of Product or Service Costs

Analyze the impact of product or service costs for [Business] with [Product/Service] in [Industry] with [Target Audience].

Provide actionable insights into determining optimal pricing options and outline potential pricing strategies to mitigate adverse market reactions.

Business = [Insert Here]

Product/Service = [Insert Here]

Industry = [Insert Here]

Target Audience = [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€œI feel like The AI Pulse is getting better and better! Itā€™s a sign youā€™re implementing constructive feedback.ā€

-Berend (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.