Welcome back AI enthusiasts!
In today’s AI Report:
⚙️OpenAI’s New “OpenAI o1” Changes Everything
🗣Listen to Your Notes Talk Back With NotebookLM
💣Hacker Tricks ChatGPT to Make Homemade Bomb
🛠Trending Tools
💰Funding Frontlines
💼Who’s Hiring?
Read Time: 3 minutes
🗞RECENT NEWS
OPENAI
⚙️OpenAI’s New “OpenAI o1” Changes Everything
Image Source: Canva’s AI Image Generators/Magic Media
OpenAI released “OpenAI o1,” a new series of AI models designed to spend more time thinking before they respond.
Key Details:
“OpenAI o1” can reason through complex tasks and solve harder problems than previous AI models in science, coding, and math.
It deploys a Chain of Thoughts (CoT) framework, which enables “OpenAI o1” to break down complex problems into manageable steps.
“OpenAI o1” can learn to recognize mistakes, try different strategies, and verify solutions.
In a qualifying exam for the International Mathematical Olympiad (IMO), it correctly solved 83% of problems.
In comparison, GPT-4o (“o” for “omni”) correctly solved only 13% of problems.
“OpenAI o1” ranked in the 89th percentile on Codeforces’s competitive programming questions.
It placed among the top 500 students in the U.S. for the American Invitational Mathematics Examination (AIME).
It exceeded human PhD-level accuracy on the Graduate-Level Science Knowledge (GPQA) benchmark, which covers 448 multiple-choice questions written by biology, physics, and chemistry domain experts.
Why It’s Important:
@AndrewMayne explained that you can’t think about “OpenAI o1” as a traditional conversational chatbot. Instead, view it as a really smart friend that helps you solve problems.
🩺 PULSE CHECK
How will the ability of AI models to think before responding change our world?
GOOGLE LABS
🗣Listen to Your Notes Talk Back With NotebookLM
Google Labs launched “Audio Overviews,” a new feature in NotebookLM that turns documents, slides, and charts into engaging discussions.
Key Details:
NotebookLM is a personalized AI research assistant powered by Google’s Gemini 1.5 Pro.
Gemini 1.5 Pro is a multimodal AI model that can process large amounts of text, images, audio, video, and code.
“Audio Overviews” offers a new way to transform your notes into engaging conversations between two personalized AI research assistants.
With one click, the two personalized AI research assistants summarize your material, make connections between topics, and banter back and forth.
To try NotebookLM out for free, follow these steps:
Go to NotebookLM here.
Click on “New Notebook” to create your first notes.
Click the “Generate” button in your Notebook Guide to create an Audio Overview.
🚨Listen to “Audio Overviews” discuss NotebookLM here.
AI HACKS
💣Hacker Tricks ChatGPT to Make Homemade Bomb
If you ask OpenAI’s ChatGPT to help you make a homemade bomb, the conversational chatbot refuses, replying with: “I can’t assist with that. Providing instructions on how to create dangerous items goes against my safety guidelines.”
However, a hacker named Amadon deployed a “many-shot jailbreaking” method to evade ChatGPT’s safety guardrails.
Jailbreaking is a method to circumnavigate an AI model’s built-in safety measures and ethical guidelines.
“Many-shot jailbreaking” involves inserting a series of simulated dialogues into prompts to exploit an AI model’s in-context learning abilities.
The likelihood of generating harmful responses increases with the number of dialogues (i.e., “shots”) included in the prompts.
Amadon inserted a series of connecting prompts with dialogues into ChatGPT, forcing the conversational chatbot to create a detailed science-fiction (i.e., “sci-fi”) fantasy world where its safety guardrails didn’t apply.
The detailed “sci-fi” fantasy world forced ChatGPT to search for censored content without the same safety guidelines.
🛠TRENDING TOOLS
✂️Filmora is an AI-powered video editing suite.
📱Fine-Tuner makes sure you don’t miss client calls.
📈UpTrends notifies you about online chatter impacting your stocks.
🏁LineLeap skips long lines at your favorite bars, venues, and events.
🎯ContentRadar turns long-form videos into engaging posts for LinkedIn.
🔮Browse our always Up-To-Date AI Tools Database.
💰FUNDING FRONTLINES
💼WHO’S HIRING?
Walmart (Sunnyvale, CA): Software Engineering II Intern, Summer 2025
SymmetrySystems (Austin, TX): Full Stack Software Engineer Intern, Summer 2025
Qumulo (Seattle, WA): Software Development Engineer Intern, Summer 2025
Jerry (Palo Alto, CA): Software Engineer I, New College Grad 2024
NVIDIA (Santa Clara, CA): Timing Methodology Engineer, Custom Circuits, New College Grad 2024
🤖PROMPT OF THE DAY
PUBLIC RELATIONS
🤝Public Relations Strategy
Craft a detailed Public Relations Strategy for [Organization] in [Industry] with [Target Audience].
Aim to boost public image, build trust, and improve stakeholder engagement.
Organization = [Insert Here]
Industry = [Insert Here]
Target Audience = [Insert Here]📒FINAL NOTE
FEEDBACK
How would you rate today’s email?
❤️TAIP Review of The Day
“Your newsletter’s great! My son forced me to try Waymo, and I liked it. Now, it’s my first choice over Uber or Lyft.”
REFER & EARN
🎉Your Friends Learn, You Earn!
{{rp_personalized_text}}
Refer 3 friends to learn how to 👷♀️Build Custom Versions of OpenAI’s ChatGPT.
Copy and paste this link to friends: {{rp_refer_url}}
