- The Rundown AI
- Posts
- 🤖 OpenAI drops GPT-5 hints
Sign Up | Advertise | Tools | View Online
Good morning, AI enthusiasts.
There’s been a ton of GPT-5 gossip floating around on Twitter.
And now, there’s finally some evidence. Let’s get to it!
In today’s AI rundown:
🤖 OpenAI files trademark for GPT-5
🎵 Meta’s new tool for creating music from text
🔎 Google announces BIG AI search updates
🛠️ 6 New tools and 3 quick hits
Read time: 3 minutes
P.S. Keep an eye out and click the star emoji hidden somewhere in today’s email for a free entry into our AI starter pack giveaway (win a free M2 MacBook Air)!
Additionally, you’ll get another free entry for every subscriber who joins our newsletter using your unique link: https://www.therundown.ai/subscribe?ref=PLACEHOLDER
You can track how many referrals you’ve made here. Happy hunting!
LATEST DEVELOPMENTS
OPENAI
Source: Blockchain News
The Rundown: OpenAI has filed an application for the trademark of “GPT-5” with the United States Patent and Trademark Office.
Key points:
The application was filed on July 18, 2023, and is currently in process.
It includes artificial speech, conversion of audio data files into text, voice and speech recognition, and machine-learning-based language and speech processing.
The application also extends to software for developing and implementing artificial neural networks.
Of note: Back in March, Siqi Chen, founder of Runway quoted: “GPT-5 is scheduled to complete training this December and that OpenAI expects it to achieve AGI.”
We’ll see if this timeline remains true. One thing’s for sure: the world is going to look a whole lot different if so.
TOGETHER WITH MONSTER API
The Rundown: Monster API's no-code fine-tuner is a groundbreaking tool that simplifies the complex and time-consuming fine-tuning process of LLaMA.
LLaMA is a remarkable Large Language Model – but it faces challenges in handling specific problems.
Fine-tuning is the solution, turning LLaMA into a specialized powerhouse. Monster API supports fine-tuning in various models, including LLaMA 7B, 13B, Falcon 7B, and more!
Start experimenting with 5,000 free API credits – Available using the code “ROWAN”.
META
Image source: Meta
The Rundown: Meta has just released a new open-source AI codenamed AudioCraft for generating high-quality, realistic audio and music from text.
Details:
It consists of three models: MusicGen, AudioGen, and EnCodec.
The models are now open-sourced, enabling researchers and practitioners to train their own models with their datasets.
The models were trained on extensive data, including 20,000 hours of music owned or licensed by Meta, and public sound effects.
Noteworthy: Many music artists around the world are lobbying against using AI in music, while others such as Grimes encourage its use.
Where do you stand? Let us know in the poll below:
Are you pro-AI music, or anti-AI music? |
Image source: Google
The Rundown: Google is adding images AND video to the summary box of AI-powered SGE search results, aiming to create a more colorful and informative experience for those who have enabled this feature in Search Labs.
Key points:
CEO Sundar Pichai emphasizes SGE as the future of Google Search.
Specific moments from YouTube videos are now linked in search results.
Google is adding publish dates to articles in the summary box, and it appears to be testing in-line links within the AI summary.
According to Google, recent improvements have cut the generation time of AI overviews in half.
Why it matters: Google’s AI endeavors haven’t always gone so smoothly, but it’s clear to see their efforts to continuously improve their AI features.
I, for one, do think they will get it right. Eventually.
TRENDING TOOLS
🪄 Magical- Send your AI to meetings to take notes and follow up on your behalf (link)*
📚 Supermanage- AI that magically distills your team’s public Slack channels into a real-time brief on any employee (link)
🎙️ Interview AI- Prepare for your next job interview with AI (link)
🧠 Decktopus- An AI-powered presentation generator (link)
📆 Vimcal- Calendar app on steroids. Assists users in scheduling and organizing events efficiently (link)
QUICK HITS
Kickstarter is implementing a new policy requiring projects that use AI tools to disclose details about how they'll use AI content, credit content creators, and indicate sources of training data. The policy will take effect on August 29th.
Researchers at Carnegie Mellon University discovered a vulnerability in popular chatbots like ChatGPT, Google’s Bard, and Claude from Anthropic. They found that adding specific strings to the prompts can make these chatbots produce harmful and prohibited responses.
Warren Buffett is cautiously fascinated by AI and ChatGPT, seeing the potential for everyday use but admitting he doesn't understand it enough to invest. He acknowledges the risks associated with AI and is uncertain about its long-term benefits for humanity.
SPONSOR US
🦾 Get your product in front of over 250k+ AI enthusiasts
Our newsletter is read by thousands of tech professionals, investors, engineers, managers, and business owners around the world. Learn more or get in touch today.
* This is sponsored content
THAT’S A WRAP
If you have anything interesting to share, please reach out to us by sending us a DM on Twitter: @rowancheung & @therundownai
How was today's newsletter?Feedback helps us improve! |
❤️ Rundown review of the day
Reply