AI News & Launches

Prep offers real-time, AI-powered mock interviews with instant feedback to help job seekers land their dream jobs. It boasts a user base of over 36,000 and provides personalized evaluation to boost confidence for interviews.

Cerebella is an iOS flashcard app designed to make studying efficient and engaging, featuring AI-powered tailored learning, an intuitive interface, visual tools, and a spaced repetition algorithm.

FileTranscribe provides an AI-powered transcription service that allows for accurate instant transcriptions, featuring diarization, multilingual support, and additional features such as summaries, and emotion, purpose, and topic detection.

Alan Turing and Christopher Strachey pioneered AI writing with the love letter generator, a program that penned peculiar love letters and became one of the world's first computer-generated writings. The duo's efforts in the early 1950s at the University of Manchester signified a groundbreaking moment in artificial intelligence, and had deeper personal significance as both men were gay during a time when homosexuality was criminalized.

Tenstorrent, an AI hardware startup, has launched PCIe development cards and systems, expanding access to their chips for all developers. The Grayskull e75 and e150 cards, from 2023, are succeeded by the Wormhole n150 and n300 cards, and the high-performance TT-LoudBox and TT-QuietBox workstations, all aiming to boost AI development by improving accessibility and connectivity.

Britannica Education provides an educational guide with practical tips and a quiz to distinguish real images from AI-generated ones, emphasizing the importance of media literacy in the digital age.

orch is a Rust framework for LLM (language learning model) orchestration. It provides a library for building language model powered applications and agents, intended for usage in magic-cli or other Rust applications.

Awan LLM is an unrestricted and cost-effective LLM Inference API platform offering unlimited tokens and various features such as AI Assistant, AI Agents, Roleplay. Their competitive edge is providing unlimited token generation by owning their data centers and GPUs.

Screenflick is a web application that allows users to search for famous movie dialogues. This tool can be found at https://screenflick.ashutoshrath.me/, providing a simple interface to generate dialogues from various films.

ChartFast is an AI-powered data analysis and visualization tool that converts data work into a streamlined, efficient process, eliminating the need for repetitive tasks such as manual data cleaning and chart generation. It offers affordable pricing plans for both small and large projects.

StudyMap.ai is an AI-driven learning platform designed to enhance the study experience by providing personalized learning maps for users. It offers tailored study paths to help users focus on their goals and boost learning efficiency.

MIT researchers have developed a machine-learning framework that predicts phonon dispersion relations in materials up to 1,000 times faster than other AI techniques, significantly speeding up the design of efficient energy-conversion systems and faster microelectronic devices. The method overcomes the complexities of modeling how heat is carried through materials by phonons, which can save time and advance material sciences.

An article titled 'When ChatGPT summarises, it actually does nothing of the kind' by gctwnl explores the author's findings that LLMs, like ChatGPT, do not truly understand or summarise content. The author's experiments show that LLMs tend to shorten text rather than summarise it, failing to capture the essence of a document.

AMD's Ryzen AI 9 HX 370 APU outperforms competitors in the PassMark benchmark, with a 12-core configuration and impressive single-core and multi-core scores, making it faster than the 20-core Intel 14700HX and 14-core Apple M3 Max.

CodiumAI addresses the complexities of large-scale enterprise codebases with their Retrieval Augmented Generation (RAG) pipeline. The blog by Tal Sheffer highlights the challenges and solutions in applying RAG to 10k+ code repositories, focusing on intelligent chunking, maintaining context, enhancing code embeddings, and advanced retrieval techniques for improved AI-powered code generation.

Supermemory is a hub designed to organize, search, and utilize information effectively. Features include a search engine, writing assistant, and canvas, enhancing productivity for students, professionals, and internet users.

Reactor Chat is an AI-driven platform that offers a smart messaging system capable of conversational interactions. It provides an intuitive chat interface that is designed to streamline communication and enhance user engagement.

Concurrence AI specializes in managing online communities by providing AI-powered moderation tools. It offers 24/7 filtering for spam, ads, and customizable content across various languages, with an uptime guarantee and usage-based payment plans.

Tenstorrent announced its latest Wormhole AI processors utilizing RISC-V architecture, aiming to provide high scalability and performance at a competitive price. The company unveiled PCIe cards in two configurations, n150 and n300, designed for AI applications and boasts cost-effectiveness with prices starting at $999. Two workstation offerings, TT-QuietBox and TT-LoudBox, were also introduced, featuring multiple Wormhole AI cards.

Mathstral is a new LLM from Mistral AI, fine-tuned for math reasoning, boasting a 32k context window and available under Apache 2.0. This article provides a guide to run mathstral locally using WasmEdge and LlamaEdge stack, detailing the setup steps for installations and serving via API. It demonstrates Mathstral's precision in solving a simple math comparison, positioning it as an excellent solution for specialized areas, even surpassing larger models.

Mantle revolutionizes software development by using Large Language Models (LLMs) to dramatically reduce the time and effort required for code conversion. Capitalizing on Gemini LLMs' latest advancements and their million-token context windows, the Mantle team was able to convert a prototype project written in R into production-quality code using Golang and ReactJS while maintaining its original logic and structure. This approach saved months of developer time by handling 80% of repetitive coding, allowing engineers to focus on high-value final touches.

Revent AI develops AI-driven chatbot solutions to automate customer service and support, leveraging machine learning to provide real-time, conversational experiences across multiple platforms.

Discovery Outcomes is an AI-powered product management platform designed to streamline the process of innovation and growth. It focuses on optimized workflows, customer-driven growth, and seamless collaboration to increase productivity.

MyMathSolver.ai is an AI-powered math solver that effortlessly solves complex math problems online for free using GPT-4o. It offers features like step-by-step solutions, interactive problem-solving with a math AI bot, and supports versatile input methods such as text or images.

New Portal is an extension for your browser that provides a more personalized and enhanced new tab page, allowing users to access frequently visited websites, bookmarks, and apps easily. It offers customization options to create a unique browsing experience.

Cohesive AI offers tools for web scraping and AI enrichment directly within Google Sheets, allowing users to enable outbound at scale without the hassle of learning a new platform.

Flot AI provides a cutting-edge AI memory solution designed to enhance artificial intelligence by enabling it to remember and learn from previous interactions and experiences. This innovative approach aims to boost AI's problem-solving capabilities and adaptiveness over time.

OpenAI's new gpt-4o mini model has reached an inflection point with its low cost, making it feasible to generate AI content supported by ad revenue. With costs at $0.15 per 1M input tokens and $0.60 per 1M output tokens, a monetized blog could profit from ad impressions exceeding AI content generation costs.

Yi Tay discusses the shift from encoder models like BERT to a new era of large language models (LLMs), explaining technical nuances like Encoder-Decoders, PrefixLM architectures, and Denoising objectives. He presents insights on why BERT-style models are deprecated, how models like T5 offer more flexibility, and the role of denoising objectives in complementing causal language modeling (CLM).

ChatGPT Chrome Extension by EliseiNicolae is an open-source project to keep the Temporary Chat feature enabled in ChatGPT. It has 5 stars, 1 fork, and is licensed under the MIT license.

Membots.ai is a platform where users can mint AI-powered memecoins for free, without fees or gas. It's a place to transform memes into interactive cryptocurrencies that you can communicate with, and it also offers an option to create AI waifus as memecoins.

Fastn.ai is a no-code, AI-powered orchestration platform that enables developers to connect data flows and create app integrations effortlessly. It supports various integrations, decreases time-to-market, and promotes infrastructure automation.

Wrenly offers a variety of HR tools to enhance workplace culture, including Anytime Feedback Suggestion Boxes, Surveys, Recognition & Rewards platforms, and AI Insights with a Culture Score. It provides free alternatives to paid services, like Kudoboard, for creating recognition boards to celebrate team milestones.

EasyMedia is an AI-powered online application that quickly transforms YouTube videos into optimized content for various social media platforms, offering a seamless integration and tailored output to fit format-specific requirements.

ai-renamer is a Node.js Command Line Interface (CLI) tool that utilizes AI models from Ollama and LM Studio, such as Llava, Gemma, Llama, etc., to intelligently rename files according to their contents. It supports various operations such as specifying the provider, setting API keys for OpenAI, choosing the model, and setting the base URL for different ports, along with other options.

LLM Optimize is designed to revolutionize SEO by offering professional audits to improve website rankings in AI Engines like ChatGPT and Google's AI Overview. With affordable one-time payments and thorough analysis, LLM Optimize positions websites to outrank competitors and capitalize on generative engine traffic.

Wasps is an AI-powered code review tool integrated in VSCode, developed by Gitsecure. It helps developers understand their codebase quickly, detects and fixes issues, and provides instant feedback and fix suggestions, all within minutes.

Xspiral is an innovative 3D design and collaboration platform aimed at everyone from beginners to professionals. With a no-download, real-time rendering environment, it integrates AI to assist in productivity and creativity. Features include 2D/3D design tools, animations, interactivity, and exporting options.

Leonardo.ai is an advanced AI-based platform offering a suite of tools for creators and professionals to generate art, videos, and transparent PNGs using AI technology. With features like the foundational Phoenix model and an array viewing options like Anime XL and Photoreal, it caters to various creative needs including design, marketing, and architecture.

TSMC CEO C.C. Wei predicts the AI chip shortage may continue until 2025 or 2026, as the company strives to balance supply and demand. Amidst high customer demand, TSMC is doubling efforts to expand CoWoS packaging capacity and pushing for overseas expansion with new plants in Arizona and Kumamoto. Despite comments by Donald Trump affecting TSMC's stock prices, Wei confirms no change in strategy.