Issue 64 6th August 2024

Highlights

Accelerating AI app development with Azure AI and GitHub | Microsoft Azure Blog

Microsoft empowers AI developers by bringing a curated list of Azure AI models to a community of more than 100 million on GitHub. Learn more.

Making AI models ‘forget’ undesirable data hurts their performance

So-called “unlearning” techniques are used to make a generative AI model forget specific and undesirable info it picked up from training data, like sensitive private data or copyrighted material. But current unlearning techniques are a double-edged sword: They could make a model like OpenAI’s GPT-4o or Meta’s Llama 3.1 405B much less capable of answering […]

Introducing GitHub Models: A new generation of AI engineers building on GitHub

We are enabling the rise of the AI engineer with GitHub Models – bringing the power of industry leading large and small language models to our more than 100 million users directly on GitHub.

OpenAI’s Sam Altman is becoming one of the most powerful people on Earth. We should be very afraid

Sam Altman’s ChatGPT promises to transform the global economy. But it also poses an enormous threat. Here, a scientist who appeared with Altman before the US Senate on AI safety flags up the danger in AI – and in Altman himself

Paper of the week

Estimate the difficulty of a music piece using AI

We think you’ll like this week’s paper for its quirkiness. Because it’s summertime, we feel it’s a great time for something different. In this paper, they explore how to use machine learning to estimate the difficulty of learning and playing a music piece. Enjoy!

Articles

AI chip startup Groq lands $640M to challenge Nvidia

Groq, a startup developing chips to run generative AI models faster than conventional processors, said on Monday that it’s raised $640 million in a new funding round led by Blackrock. Neuberger Berman, Type One Ventures, Cisco, KDDI and Samsung Catalyst Fund also participated.

Productionizing a RAG App with Prefect, Weave, and RAGAS

A few months ago, I released the Film Search app, a Retrieval-Augmented Generation (RAG) application designed to recommend films based on user queries. def evaluate_with_ragas(query, model_output):

There’s a Tool to Catch Students Cheating With ChatGPT. OpenAI Hasn’t Released It.

OpenAI evidently have had working prototypes of this for a couple of years now, but they haven't shipped it as a feature. I think this is the key section for understanding why.

Introducing torchchat: Accelerating Local LLM Inference on Laptop, Desktop and Mobile

Today, we’re releasing torchchat, a library showcasing how to seamlessly and performantly run Llama 3, 3.1, and other large language models across laptop, desktop, and mobile.

Google rehires founders of consumer chatbot startup Character.AI

Character.AI Inc., a well-funded startup with a chatbot service for consumers, today announced that its founders will join Google LLC. Noam Shazeer and Daniel De Freitas are returning to the search giant about two years after leaving their engineering roles to launch Character.AI. The chatbot maker, in turn, will continue operating under new leadership.

AI scientist: 'We need to think outside the large language model box'

AI21 Labs co-founder Yoav Shoham says something extra is needed to build better and smarter generative artificial intelligence.

Precision home robots learn with real-to-sim-to-real

CSAIL researchers introduce a novel approach allowing robots to be trained in simulations of scanned home environments, paving the way for customized household automation accessible to anyone.

Data Value Lineage, meaning at last?

“Data value lineage is the process of reconciling at all times the enterprise data assets (including their maintenance costs) and the main delivery tasks performed within a data organisation with the measured contribution to enterprise value drivers”

OpenAI pledges to give U.S. AI Safety Institute early access to its next model

OpenAI CEO Sam Altman says that OpenAI is working with the U.S. AI Safety Institute, a federal government body that aims to assess and address risks in AI platforms, on an agreement to provide early access to its next major generative AI model for safety testing.

Code

Handwritten Text Recognition for Xournal++ Using Deep Learning

Taking handwritten notes digitally comes with many benefits but lacks searchability of your notes. Hence, there is a need to make your handwritten notes searchable. This can be achieved with "handwritten text recognition" (HTR), which is the process of assigning searchable text to written strokes.