• FavTutor
  • Posts
  • OpenAI’s o1: AI That Can Think Before It Speaks

OpenAI’s o1: AI That Can Think Before It Speaks

Including the latest AI news of the week

Hello, AI Enthusiasts!

Welcome to FavTutor’s AI Recap! We’ve gathered all the latest and important AI developments from the past week in one place, just for you.

In Today’s Newsletter: 😀 

  • Google’s new AI Turns Your Notes into Podcast

  • OpenAI’s o1: AI That Can “Think” Before It Speaks

  • Google launched DataGemma AI Especially for Statistics

  • Adobe previews its new Gen AI Tools with Firefly Model

  • Codeforces bans GPT & Gemini in its Competitions

OpenAI
🧠 OpenAI’s o1: AI That Can “Think” Before It Speaks

OpenAI, the company behind ChatGPT, has unveiled its latest breakthrough: a new AI model called o1 (previously code-named “Strawberry“). The biggest differentiation of the o1 model when compared to other LLMs is that it ‘THINKS’. It processes before answering the prompt, instead of using predictive sequences of words.

Insights for you:

  • OpenAI has launched its mystery ‘Strawberry’ model named ‘o1’.

  • It is already rolled out to ChatGPT to all Plus and Team users.

  • Coding is one of the big areas where it is said to work better than other LLMs.

PARTNERSHIP
The best new app for newsletter reading.

Meco is a distraction-free space for reading and discovering newsletters, separate from the inbox. Add your newsletters in seconds and liberate your inbox today!

Meco connects to Gmail & Outlook, allowing you to instantly move your newsletters to the app (and clear your inbox)

You can try the experience for free and available on iOS and the web (desktop).

Google
🎙️ Google’s new AI Turns Your Notes into Podcast

Google’s NotebookLM app now offers an experimental feature called ‘Audio Overview’ that transforms research notes into AI-generated podcasts. Along with the discussion on the notes, the hosts use casual language, modern phrasing and banter with each other. They will also make connections between the topics to make it sound like real two people.

Insights for you:

  • The ‘Audio Overview’ AI feature in Google’s NotebookLM can create podcasts with two hosts discussing your research.

  • The hosts will summarize the content of the notebook, as well as engage in banter to make it sound real and fun.

  • However, it is currently limited to available in the English language only and might have some inaccuracies.

Google
📊 Google launched DataGemma AI Especially for Statistics

Google is not stopping with new AI models as they launched two new DataGemma models last week: DataGemma-RAG-27B-IT and DataGemma-RRIG-27B-IT. These are instruction-tuned models built upon the Gemma family to reduce hallucinations. DataGemma will harness the knowledge of Data Commons, a knowledge graph containing over 240 billion rich data points.

Insights for you:

  • Google’s new DataGemma models are open-source models to provide more accurate outputs in case of statistical data.

  • The models use Retrieval-Interleaved Generation and Retrieval-Augmented Generation to mitigate hallucinations.

  • The models are available on Hugging Face for research purposes.

Adobe
🛠️ Adobe previews its new Gen AI Tools with Firefly Model

Adobe has specially showcased three new Generative AI Tools: Text-to-video, Generative Extend, and Image-to-video. Some examples shared by Adobe were made in under 2 minutes. The more detailed your prompts are, the better the model can leverage that depth to generate inspirational imagery and b-roll.

New AI Tools:

  • Text to Video: Like SORA, Runway and other tools, you can video clips using prompts with control of camera angles, motions, and zoom.

  • Generative Extend: This tool allows you to extend existing shots in the timeline.

  • Image to Video: You can also upload your images to turn them into videos.

Codeforces
⏹️ Codeforces bans GPT & Gemini in its Competitions

Codeforces, a popular online programming platform has banned the use of AI like GPT, Gemini, and Claude in its competitions. The new rule allows for limited use like for basic code (boilerplate) or problem statements, but not for creating algorithms. This rule applies strictly to in-contest participation.

Insights for you:

  • Codeforces explicitly limits the use of AI systems like GPT, Gemini, Gemma, Llama, and Claude for solving programming problems.

  • Participants can’t receive ready-made code or natural language descriptions of the solution, nor they can resolve errors based on system feedback.

  • The ban came after when openAI’s new o1 model was performing excellently in the case of coding.