AI Powered Tools

A run down of the latest SaaS products and services which leverage AI to give you back time and focus on what’s important.

• ChatSpot - free AI-powered assistant that combines the power of ChatGPT with unique data sources, including the HubSpot CRM, to help supercharge your work

• Manatal - for recruiters. Designed to source and hire candidates faster. Tailored for HR teams, recruitment agencies, and headhunters.

• Motion - for busy people who need more time. Automated scheduling and planning with A.I.

News

Data Integrity: The Key to AI Success

The success of Artificial Intelligence (AI) projects is tied to the integrity and management of data. From capturing relevant data to efficient ingestion, curation, and the timely deprecation and deletion of outdated information, each step is crucial for ensuring quality outcomes. Implementing solid data management practices not only enhances trust in AI-driven analytics but also unlocks the potential for significant operational improvements. This article emphasizes the foundational role of data integrity in achieving impactful and sustainable AI solutions. Read the full article on Bright Journey AI

China Unveils World's First AI "Child" - Tong Tong

Chinese scientists from the Beijing Institute for General Artificial Intelligence (BIGAI) have introduced the world's first AI "child" named Tong Tong. This virtual AI avatar represents a significant leap toward developing a general artificial intelligence (AGI) capable of thinking and reasoning like a human. Unlike conventional AI that focuses on single tasks, Tong Tong demonstrates autonomy, emotional intelligence, and the ability to learn, reflecting human-like cognition and decision-making abilities. The introduction of the Tong test by BIGAI further aims to advance AGI by assessing AI on a broad spectrum of capabilities including vision, language, and social values.

For more details, visit the original article on Interesting Engineering.

OpenAI and Common Sense Media: AI Education and Safety Initiative

OpenAI collaborates with Common Sense Media to develop AI guidelines and educational materials for teens and families. OpenAI CEO Sam Altman expressed at a conference the aim to make AI beneficial, especially for kids lacking AI access. Common Sense is creating an AI ratings system for understanding AI's risks and benefits. This partnership seeks to establish AI usage guidelines and family-friendly GPT models, aligning with Common Sense's standards. Their goal is to educate about safe, responsible AI use and prepare for AI's societal impacts.

Read the full article here

Google Chrome Introduces Generative AI Features

Google Chrome has launched three new generative AI features to enhance user experience: a tab organizer for efficient browsing, AI-generated custom themes for personalization, and writing assistance for web-based text input. These innovations are part of Chrome's ongoing integration of machine learning and AI technologies, aimed at making web navigation easier, safer, and more customizable. Initially available in the U.S. on Macs and Windows PCs, these features represent Chrome's commitment to leveraging AI for improved functionality and user engagement.

Read the release here

Large Language Models

No doubt LLMs are the flavour of the month. We like other model architectures too but this is where there is pace, so we need to help you get to the important stuff.

Reducing GPT-4 Costs with LLMLingua Prompt Compression

Microsoft has developed LLMLingua, a new technique for compressing prompts in large language models (LLMs) like GPT-4 and Claude, without compromising response quality. This innovation can compress prompts by up to 20 times, significantly reducing the costs and making advanced LLMs more accessible. LLMLingua works by eliminating irrelevant parts of prompts, optimizing their length while preserving essential information. This development holds promise for broader application of LLMs by reducing computational costs and accommodating longer contexts more efficiently. Microsoft has also made LLMLingua available as an open-source library for developers.

Read the full article on bdtechtalks.com

Open Source LLM Revolution by AI2

The Allen Institute for AI (AI2) has launched OLMo, a groundbreaking open-source Large Language Model (LLM), marking a pivotal advancement in AI. Unlike existing models, OLMo offers complete access to its training code, data, and toolkits, coupled with an open-source license. This initiative aims to foster transparency and accelerate responsible AI development, challenging the norm of proprietary models. OLMo's comprehensive framework enables deeper scientific exploration of LLMs, promising to enhance AI safety and trustworthiness.

Read the full article