Top 10 AI News #weekly
This week, 23 startups from seven Asia-Pacific countries embarked on Google’s inaugural AI Academy, starting with a three-day boot camp in Singapore.
Imagine a world where AI helps you make smarter choices, gives you more personalized experiences and helps everyone feel included. This is what 23 AI startups from across the Asia-Pacific, all selected for Google for Startups' first AI Academy in the region, are trying to create. The program — where startups will gain mentorship, product training and networking opportunities to refine their AI products and grow their business — started this week with an interactive bootcamp in Singapore.
Hailing from Japan, Korea, India, Indonesia, Malaysia, Pakistan and Singapore, these innovators are building AI collaborators for lawyers, democratizing influencer marketing with AI-generated avatars and expanding access to underserved areas like financial services and healthcare via AI-powered platforms. By mentoring these startups, we’re investing in a future where AI is used to solve pressing societal issues and create more opportunities.
Adobe demonstrated the amazing Project Perfect Blend, where AI can seamlessly blend people and animals with completely mismatched lighting, exposure, and colors into photos with just one click.
This kind of inline AI feature improves the efficiency of our work without changing the usual workflow.
With Perplexity’s groundbreaking new Spaces feature, you’ll have our own personal research assistant who not only answers your questions but creates organized, shareable collections of knowledge.
These collaborative environments allow users to:
• Upload and store files
• Search files in addition to the web
• Pick an AI model of your choice
• Write custom instructions for the answers you want
• Invite others and search collaboratively
Notice: only Enterprise Pro users can use Space.
The new model appears to be a fine-tuned version of Meta’s Llama-70b.
Nvidia unceremoniously launched a new artificial intelligence model on Oct. 15 that’s purported to outperform state-of-the-art AI systems including GPT-4o and Claude-3.
OpenAI has introduced the model GPT-4 Audio Preview.
Key features include:
- Generating voice responses based on text.
- Analyzing the emotion, tone, and pitch of audio input.
- Supporting voice-to-voice interaction, where audio can serve as both input and output.
Difference from Realtime API:
ChatGPT audio preview: Mostly for generating voice, analyzing voice emotion, and voice-to-voice interaction, paying much attention to the elaborate handling of voice input and output in tone and emotion.
Realtime API: It focuses on real-time data processing, for immediate feedback scenarios like real-time speech to text, instant translation, and other continuous interactive applications.
The new model appears to be a fine-tuned version of Meta’s Llama-70b.
Nvidia unceremoniously launched a new artificial intelligence model on Oct. 15,2024, which is purported to outperform state-of-the-art AI systems including GPT-4o and Claude-3.
ChatGPT Plus, Enterprise, Team, and Edu users can start testing an early version of the Windows desktop app.
Get faster access to ChatGPT on your PC with the Alt + Space shortcut.
The ability to use AI to evaluate AI reliably offers a glimpse at a possible pathway toward building autonomous AI agents that can learn from their own mistakes.
The idea of being self-taught and able to self-evaluate is basically crucial to the idea of getting to this sort of super-human level of AI.
Google is continuing to build on the viral success of NotebookLM with a handful of updates today. With millions of users to date, Google’s NotebookLM is losing the “experimental” status/badge in the top-left corner.
"We’re not just advancing AI infrastructure—we’re reinventing it," said Lightmatter co-founder and CEO Nick Harris. "With Passage, the world’s fastest photonic engine, we’re setting a new standard for performance and breaking through the barriers that limit AI computing. This funding accelerates our ability to scale, delivering the supercomputers of tomorrow today."
Best Open Source Alternatives to Proprietary Software#weekly
Crawl4AI Web crawler
Crawl4AI automates a web crawl or data extraction process for large language models (LLMs) and AI applications. Learn more about its features, pros and cons.
Open Canvas
Open Canvas, inspired by ChatGPT Canvas, is an open-source web application for collaborative document writing with the help of intelligent agents.
This Week's Summary
In this week's edition of AI Weekly News, we’ve learned about some of the most important breakthroughs in AI. Packed with launches, including the launch of Google's AI Academy announcing 23 startups in the Asia Pacific bound to innovate across industries, and Adobe introduces its Project Perfect Blend to make photo editing all the more engaging.
We then discovered an Nvidia AI model outperforming GPT-4o benchmarks, OpenAI bringing improvements to voice interactions with a preview of GPT-4 Audio, and Meta presenting a model that can assess its own performance. That's all in the name of quick changes within AI technology and its applications.
