AI News & Launches

Y Combinator's online founder-matching tool helps Hona, a GenAI medical records startup, secure $3M seed funding. Founders Yoesep and Steinle, reconnecting after respective careers, were joined by AI specialist Zhang after matching on YC's platform. Despite initial rejection from YC, Hona persisted and was accepted in the winter 2024 batch, with Duke and Harvard MDs testing the product.

Joseph Jones in Scientific American opines that the true threat to humanity isn't AI itself, but rather its owners who exploit AI to undermine privacy and democracy. He outlines the unethical use of technology for surveillance and manipulation, discussing topics from Incognito mode's false privacy to corporate data harvesting, and stressing the need for human principles to guide tech advances.

Scrapegraph-ai is an innovative open source library designed to simplify web scraping tasks using the power of AI. It provides easy installation, fast implementation, and focuses on automating the scraping process to save hours of time.

InstructLab is an open-source command-line interface tool designed for creating, training, and interacting with large language models. It's geared towards allowing users to add new knowledge, generate synthetic training data, and re-train models to enhance performance. The project incorporates alignment tuning methods, and supports customization and community contributions.

Apple introduces Final Cut Pro for iPad 2 with Live Multicam and AI editing features on Mac. Advanced capabilities with M4 chip-powered iPad Pro, wireless multicam recording using Final Cut Camera app, and storage support for external drives bring new efficiencies to video production.

Microsoft is reportedly progressing on an extensive in-house artificial intelligence model named MAI-1 which is said to feature around 500 billion parameters, surpassing the scale of other open source models it has trained. This initiative is poised to rival AI models developed by tech giants like Google, and companies like OpenAI, in which Microsoft is also an investor.

AFFiNE AI is a multifaceted AI Partner that enhances writing, drawing, and presenting capabilities. Offering tools such as content summarization, tone perfection, insight provision, and image generation, AFFiNE AI aims to streamline the creative process. With an emphasis on privacy and a user-friendly experience, it's designed for both professionals and casual users.

Storyville offers a unique service to create personalized bedtime stories for kids using AI. It offers the ability to generate tailored stories with beautiful illustrations and educational content, designed to foster empathy and social skills. The app also provides a simple pricing model with the promise of future premium features, including audio stories.

Experience Adobe Express, featuring add-ons that enhance creativity through templates, design assets, and plugins. Ideal for making social media content, logos, and more.

AFFiNE AI offers an AI-powered platform to create content, transform ideas into images and presentations, and facilitate better writing. It features a downloadable app, a range of resources, and a pricing plan of $8.9 per month (billed annually) with a free trial available. Users can look forward to AI integration with databases and workspaces, as well as the upcoming features like auto-tagging.

This hypothetical YouTube video discusses the latest developments in technology including new smartphone releases, advances in AI, and cybersecurity trends.

Stack Overflow and OpenAI announce a partnership to integrate Stack Overflow's OverflowAPI with OpenAI's AI models. This collaboration aims to provide developers with trusted and vetted technical knowledge and to improve the performance of OpenAI models using the Stack Overflow community's content.

iLus AI specializes in leveraging artificial intelligence to drive innovation, offering solutions that range from machine learning models to data analytics.

Logome is an AI-powered logo generator that allows users to effortlessly create stunning logos and brand kits without the need for design skills. It provides a seamless process for generating logos and creating personal websites featuring AI-designed logos, making brand building simple and efficient.

Touch is a platform designed to automate the sales prospecting process. It seamlessly integrates with existing sales tools to create ideal customer profiles, curate weekly updated lists of leads, and generates personalized emails sent to targeted customers in seconds.

Recruiters are adapting to the influx of AI-generated applications by either incorporating more generative AI tools or relying on traditional methods. Tech layoffs since 2022 have increased job applications, leading to AI being used more in recruitment processes. However, concerns about AI bias and quality of candidates remain. Major job platforms like LinkedIn and Indeed are enhancing their AI tools for both recruiters and job seekers to improve the efficiency of recruitment.

Ken Liu addresses machine unlearning, focusing on the removal of training data influence from ML models to mitigate issues like private data retention and misinformation without retraining from scratch. He explores various unlearning approaches, their challenges, and the concept's rising importance in 2024.

MindPal offers robust AI workflows to efficiently automate any tasks, significantly boosting productivity for businesses. Free to use, this platform is driven by a strong foundation that quickly adapts to user needs and continually enhances features.

Actionize allows users to connect their GPT to thousands of apps using familiar no-code platforms like Zapier and Make.com. It offers a simple interface to manage Actions for the GPT and a one-click code snippet export for use in ChatGPT.

ContentPie is an AI-powered content generation platform aimed at increasing organic traffic and improving SEO rankings by automating content creation. It offers features like on-brand AI images, article ideas, web publishing, and bulk articles management with support in over 150 languages.

Sequoia introduces a scalable, robust, and hardware-aware speculative decoding framework for serving large language models like Llama2-70B with low latency on consumer GPUs like RTX4090. Sequoia achieves notable speed improvements over previous systems, facilitates faster model inference, and adapts to different hardware with no loss of output quality. Its robustness is attributed to handling speculation budgets efficiently and integrating a sampling without replacement algorithm to ensure consistency in generated temperatures.

An exploration of text compression using a large language model trained to predict text sequences. The author tests the ability to extract and reproduce copyrighted text, achieving significant compression on known texts like 'Alice's Adventures in Wonderland'.

This paper presents a Bayesian learning model to understand the behavior of Large Language Models (LLMs). Authors Siddhartha Dalal and Vishal Misra explore the optimization of LLMs, presenting a generative text model and examining the LLMs' approximation of it. They introduce the Dirichlet approximation theorem for prior approximations and discuss the implications of in-context learning.

CBMC is a Bounded Model Checker for C and C++ programs, supporting various C/C++ versions and compiler extensions. It verifies memory safety, checks exceptions, undefined behavior, and has user-specified assertions. CBMC is available for Linux, Windows, and MacOS X, and can be integrated with multiple SMT solvers like Boolector, CVC5, and Z3.

The field of robotics is witnessing significant advancements akin to the AI breakthroughs that made ChatGPT possible. By leveraging techniques like generative AI and reinforcement learning, researchers are empowering robots with the ability to learn new skills and adapt to various environments, even in complex, unpredictable home settings. Notable examples include the PR2 robot and its smaller successor, Stretch, which offer a vision of practical and autonomous home robots that could transform daily living for individuals like Henry Evans, who, after suffering a stroke, found a new level of autonomy with these robotic innovations.

Pressmaster.ai is an AI-powered platform designed to streamline the PR publishing process, allowing users to plan, write, and share articles instantly with over 3000+ magazines using pre-made templates. It boasts features like a Magic Editor for creating compelling content, a drag & drop newsroom, and detailed metrics and KPIs tracking.

Creatoor AI offers a platform to quickly create premium videos using AI, catering to various use cases with ready-to-use features, and a blog for updates.

AI Copilots, increasingly present in software development, are shifting coding education from syntax memorization to emphasizing higher-level problem-solving skills. Universities are adapting curricula to include critical testing and debugging skills, and educators stress the importance of AI as a tool rather than a replacement for learning.

The website discusses DrEureka, a novel approach for sim-to-real transfer utilizing Language Models (LLMs) to automate the design of reward functions and physics parameter randomization for robust robot skill acquisition. It achieves notable success without human intervention in tasks like quadruped Globe walking and dexterous cube rotation, as well as handling unexpected disturbances like kicking or deflating a yoga ball.

Ontario family doctor Dr. Rosemary Lall states new AI notetaking technology saved her job by significantly reducing the time spent on patient-related paperwork. The AI Scribe program helps compile SOAP notes quickly, improving work balance and patient care.

UNA leverages AI to offer personalized Canadian immigration solutions across various programs such as skilled worker, provincial nomination, and family reunification. Utilize tools like CRS Calculator and Express Entry Invitations to navigate the process efficiently.

Fotographer.ai offers marketing content creation tools powered by AI, including personalized backgrounds, photorealistic images generated from text, and AI model stock photos. It includes templates for easy creation and aims to enhance e-commerce conversion rates through professional-looking images. Free and paid plans are available with various features.

HoshAI offers advanced AI-based content generation services, harnessing the power of OpenAI's ChatGPT 4 Turbo and tools like DALL-E 3 for producing text and images. Users can benefit from custom templates, a multilingual system, and secure payment options for a comprehensive digital experience.

Georgia Tech collaborates with Meta to create an expansive open-source dataset, designated OpenDAC, to enhance the development of AI-driven carbon capture solutions. The massive dataset and AI model aim to expedite direct air capture technology design, reducing costs and potentially accelerating climate mitigation efforts. Their joint research has been published in ACS Central Science.

Wizardshot is a Chrome extension that offers a simple and quick way to create AI-powered, step-by-step tutorials directly from your browser. With its easy three-click process, users can capture their screens to make tutorials without manually writing instructions, enhancing productivity and knowledge sharing.

Synthesia is an AI video creation platform that allows users to produce professional videos leveraging AI avatars and voiceovers in over 130 languages. The platform's key offerings include creating videos as easily as making a slide deck, providing a wide range of AI avatars, natural voice options, 1-click translations, and numerous video templates. It caters to various use cases such as training, marketing, sales, customer service, and more, offering scalable and affordable content creation solutions.

Braintrust AIR is a groundbreaking AI technology aimed at reinventing the hiring process by significantly reducing the time, cost, and effort required to hire top talent. It offers AI-powered job description generation, candidate matching, and asynchronous AI interviews, aiming to streamline the entire recruitment process.

GitHub repository 'npi' by 'npi-ai' is a platform providing Tool-use APIs for AI agents in virtual worlds. It is designed for developers needing AI-driven automation, offering quickstart guides, installation help, command-line tools, and access to Python SDKs.

Microsoft has revised their usage terms for Azure OpenAI Service to ban U.S. police departments from utilizing their generative AI technologies, which includes language models like GPT-4. The decision comes amidst concerns over potential biases and inaccuracies, and follows the announcement of Axon's new product leveraging AI to summarize police bodycam audio. While global law enforcement agencies are now prohibited from real-time facial recognition via Azure OpenAI Service, questions remain about the specifics of these restrictions and existing law enforcement contracts.

Researchers demonstrate a methodology using non-expert crowdsourcing to obtain high-quality surgical tissue annotations for training AI models in real-time colorectal surgery assistance. This approach saved hundreds of expert hours and produced a model, Bowel.CSS, accurate in segmenting surgical tissues.