🤖 OpenAI Postpones Search Engine Reveal

PLUS: The Transformer Killer?, Can Teachers Detect AI?!

Welcome back AI enthusiasts!

In today’s AI Report:

  • 📆OpenAI Postpones Search Engine Reveal

  • 🪦The Transformer Killer?

  • 🎓Can Teachers Detect AI?!

  • 🛠5 Trending Tools

  • 💰Venture Capital Updates

  • 💼Who’s Hiring?

Read Time: 3 minutes

🗞RECENT NEWS

OPENAI

📆OpenAI Postpones Search Engine Reveal

Image Source: Steve Jennings/Getty Images for TechCrunch

OpenAI postponed a developer conference at the company’s headquarters, at which the startup’s leaders were rumored to reveal a ChatGPT-powered search engine that rivals Google Search.

Key Details:
  • Pete Huang, the host of The Neuron podcast, posted on X: “Search (dot) ChatGPT (dot) com.”

  • Currently, there’s little information about the soon-to-launch ChatGPT-powered search engine. However, a Y Combinator Hacker News forum discovered that OpenAI registered a new domain under search.chatgpt.com with an SSL certificate.

  • When OpenAI CEO Sam Altman appeared on the Lex Fridman podcast in March, he said: “The intersection of LLMs plus search, I don’t think anyone has cracked the code yet. I would love to do that. I think that would be cool.”

Why It’s Important:
  • OpenAI executives have been active on social media platforms, mocking the startup’s current product offerings and laughing at how limited current Big Tech GenAI products are.

  • At the Milken Institute Q&A event, OpenAI COO Brad Lightcap stated, “We’ll look back in a year and realize how laughably bad they were.”

  • OpenAI might challenge search giants like Google Search by offering a new way to find information through LLMs like GPT-4.

🩺 PULSE CHECK

Would you use ChatGPT-powered search as your default search engine?

Vote Below to View Live Results

Login or Subscribe to participate in polls.

SEPP HOCHREITER

🪦The Transformer Killer?

Image Source: Canva AI Image Generator

Sepp Hochreiter, a Deep Learning pioneer and the original creator of Long Short-Term Memory (LSTM), unveiled a revolutionary new LLM architecture called “Extended Long Short-Term Memory (xLSTM).”

Key Details:
  • LSTM networks are a type of recurrent neural network capable of learning order dependence in sequence prediction problems like machine translation or speech recognition.

  • Simply put, LSTM networks can learn the order of things in a sequence, like words in a sentence, and use that knowledge to make predictions.

  • However, LSTM networks have three major weaknesses:

    1. The Inability to Revise Storage Decisions

    2. Limited Storage Capacities

    3. A Lack of Parallelizability Due to Memory Mixing

  • Simply put, LSTM networks can forget things they shouldn’t, can’t remember everything forever, and take time to process information because they can’t do it all at once.

  • To solve these issues, Hochreiter’s research team developed “xLSTM,” enabling LSTM networks to revise storage decisions and enhance storage capacities.

Why It’s Important:
  • The “xLSTM” architecture efficiently handles various aspects of long context prompts and has better validation perplexities than Transformers.

  • Most popular LLMs rely on Transformers, a type of neural network that transforms an input sequence into an output sequence. “xLSTMs” might replace the Transformer.

🚨Explore the paper here.

AI RESEARCH

🎓Can Teachers Detect AI?!

A series of studies by several German universities revealed that novice and experienced teachers struggled to distinguish between student-written and AI-generated texts.

ScienceDirect outlined the two studies, showcasing that 89 novice and 200 experienced teachers had difficulty identifying texts generated by OpenAI’s ChatGPT when asked to evaluate them on the dimensions of “language,” “structure,” and “content.”

The first study tested pre-service teachers enrolled in a Master of Education (M.Ed.) program at German universities. The participants correctly identified only 45.1% of AI-generated texts.

The second study tested experienced teachers with English as their first language. The participants correctly identified just 37.8% of AI-generated texts.

These two studies revealed that novice and experienced teachers were overconfident in their ability to spot AI-generated text.

🛠TRENDING TOOLS

🚀AgentCrew offers 100% free AI workflows to automate any tasks.

📱Meaning is the world’s first AI screen time coach to monitor addictive apps.

✏️AFFiNE is an AI partner that helps you write better, draw faster, and present better.

⚙️Plumb is a collaborative, no-code, node-based builder.

👨‍🚀Postman builds, debugs, documents, monitors, and publishes APIs.

🔮Browse our always Up-To-Date AI Tools Database.

💰VENTURE CAPITAL UPDATES

  • Checkfirst raises a $1.5M pre-seed to apply AI to remote inspections and audits.

  • Biden will announce a $3.3B investment by Microsoft to build an AI data center.

  • Atlan raises a $105M Series C to help companies get data AI-ready.

💼WHO’S HIRING?

  • Turing Video (San Jose, CA): React Developer Intern, Summer 2024

  • Amazon (Seattle, WA): Hardware Development Engineer Intern, Fall 2024

  • Disney (Burbank, CA): ABC Signature Casting Intern, Fall 2024

  • BMW (Silicon Valley, CA): Business Administration Intern, Fall 2025/Spring 2025

  • Persona (San Francisco, CA): Junior Software Engineer, New Grad

🤖PROMPT OF THE DAY

COMPETITIVE LANDSCAPE

🕵️‍♂️Research Competitors

Act like a market research consultant. Provide me with a detailed analysis of my top five competitors in the [Specific Industry].

Specific Industry = [Insert Here]

📒FINAL NOTE

If you found this useful, follow us on Twitter or provide honest feedback below. It helps us improve our content.

How was today’s newsletter?

❤️TAIP Review of the Day

“You don’t miss! The quality of this newsletter is unmatched.”

-Michael (⭐️⭐️⭐️⭐️⭐️Nailed it!)
REFER & EARN

🎉Your Friends Learn, You Earn!

You currently have 0 referrals, only 1 away from receiving ⚙️Ultimate Prompt Engineering Guide.

Refer 9 friends to enter 🎰May’s $200 Gift Card Giveaway.

Reply

or to participate.