insideAI News AI News Briefs – 8/31/2023

Welcome insideAI News AI News Briefs, our timely new feature bringing you the latest industry insights and perspectives surrounding the field of AI including deep learning, large language models, generative AI, and transformers. We’re working tirelessly to dig up the most timely and curious tidbits underlying the day’s most popular technologies. We know this field is advancing rapidly and we want to bring you a regular resource to keep you informed and state-of-the-art. Enjoy!

Habana and Hugging Face Demonstrate Top AI Training Performance on BridgeTower Vision-Language Model – recall the joint development work by Intel AI researchers and Microsoft Research to create BridgeTower, a pre-trained multimodal transformer that delivers state-of-the-art vision-language tasks. This model was recently integrated into Hugging Face, the leading open source library for machine learning. Hugging Face published an update to a blog post that recaps AI training performance benchmarks for Habana Gaudi2 AI accelerators and NVIDIA’s H100 GPU. Overall, the results seem to show that Gaudi2 outperforms H100 for training this multimodal transformer BridgeTower model and achieves 2.5 times the speed of the A100.

OpenAI Launches ChatGPT Enterprise, Business-Focused Service – The GPT-4-powered Enterprise service is the biggest OpenAI announcement since ChatGPT’s debut according to CNBC. It includes access to GPT-4 without usage caps, performance up to 2x faster than previous versions, API access, enhanced privacy and data analysis, and performance and customization options. One key differentiator is that Enterprise will allow clients to input company data to train and customize ChatGPT for their own industries and use cases. This feature unlocks data analysis customization features such as creating charts and performing math tasks.

SeamlessM4T, a Multimodal AI Model for Speech and Text Translations – potentially laying the way for a space-aged “Universal Language Translator.”

NVIDIA Beats on Earnings Up 843% YOY – The AI chip leader surpassed expectations with revenue of $6B for Q2 2024 driven by demand for its AI chips. Q3 projections point to an even higher level at $16B.

Release of Meta’s Code Llama – The code-generating tool built on top of the Llama 2 LLM is now live in an effort to compete head-to-head with services like Microsoft’s GitHub Copilot and Amazon’s CodeWhisperer.

Microsoft and Epic expand AI collaboration – Meant to accelerate generative AI’s impact in healthcare, addressing the industry’s most pressing needs.

Business Tap into ChatGPT’s Interface with Fine-Tuning Release – OpenAI announced GPT-3.5 Turbo + fine tuning and has partnered with Scale to empower more companies to create custom state-of-the-art models for their specific business needs. In context with this month’s acquisition of product design firm Global Illumination to work on their core products, it appears that Open AI has big plans to offer richer digital experiences for users. It comes with a digestible cost structure: $0.008 for training, $0.012 for input, and $0.016 per 1K tokens for output.

Ikigai Lands $25M to Automate Data Science – GenAI for tabular data which generates synthetic data for training ML models to improve accuracy and efficiency.

AI Writing Tool Raises $2.75M – Lex.page assists writers get past roadblocks optimizing sentence structure, word choice, and more.

Google Supercharges Vertex AI with New Capabilities and Extensions – Vertex AI is getting a significant overhaul, enhancing its large language models and expanding developer tooling.

An LLM for Every Enterprise – HuggingFace is targeting mass adoption for AI builders after closing a huge series D led by Salesforce. Boosts over 10,000 enterprise customers, 500,000 AI models, and 250,000 data sets.

Modular’s AI Software to Challenge NVIDIA – Google has just backed this startup in a $100M round to build a cheaper alternative to AI development with its software, bypassing the need for expensive and in-demand NVIDIA chips.

Former Starbucks Exec Joins Atypical AI – The new head of engineering joins the startup making AI for personalized learning experiences. Atypical AI also just raised $4M in a pre-seed round from Bloomberg, Ascend, and SNR.

Shownotes.io – Using AI to summarize anything including show notes for podcasts.

Databricks Eyes Strategic Funding – The data analytics platform is reportedly in talks with T. Rowe Price on a new round to value the company at $43B.

U.S. Copyright Office Seeks Public Input on AI and Copyright Dilemmas – The U.S. Copyright Office is launching a public comment period on August 30th to explore the cloudy waters of AI and copyright issues.

Sign up for the free insideAI News newsletter.

Join us on Twitter: https://twitter.com/InsideBigData1

Join us on LinkedIn: https://www.linkedin.com/company/insidebigdata/

Join us on Facebook: https://www.facebook.com/insideAI NewsNOW

Comments

  1. Your AI News Briefs provide a snapshot of AI’s latest strides. The Habana-Hugging Face collaboration showcasing Gaudi2’s training prowess against H100 marks a milestone in AI accelerator capabilities. OpenAI’s ChatGPT Enterprise launch is a game-changer, offering speed, customization, and privacy enhancements for tailored industry applications. SeamlessM4T’s potential for a Universal Language Translator hints at groundbreaking developments in speech-text translations. These updates illuminate AI’s evolution, promising transformative applications across industries. Thanks for keeping us at the forefront of AI’s dynamic landscape!

  2. Great & Helpfull Post!