Zero One Thing completed hundreds of millions of dollars in financing, and an international strategic investor and Southeast Asian consortium joined
Zero One Thing, an AI big model unicorn company founded by Kai-Fu Lee, has completed a new round of financing, with the amount reaching hundreds of millions of dollars. People familiar with the matter said that the participants in this round of financing include a certain international strategic investment, Southeast Asian consortium and other institutions.
Founded in May 2023, Zero One Everything was founded by Kai-Fu Lee, chairman of Sinovation Ventures and former global vice president of Microsoft. The core team members come from companies such as Google, Microsoft, IBM, and Baidu. The company's product strategy for the C-end is to first verify in overseas markets and then land in the domestic market. Zero One Everything has launched the AI office tool PopAi overseas and the productivity tool "Wanzhi" in China. On the B-end, Zero One Everything's large model has entered the closed-source commercialization stage and launched its first closed-source model Yi-Large. The current customer portrait shows a distinct global feature. (Intelligence Emerges)
AI-native human behavioral security firm Abnormal Security raises $250M at $5.1B valuation to boost AI-driven cyber protection
Founded in 2018, Abnormal Security offers a human security platform that uses machine learning to block sophisticated inbound attacks and detect compromised accounts in email and connected applications.
Abnormal Security Corp. has raised $250 million in late-stage funding at a valuation of $5.1 billion. The funding is intended to enable the company to continue its mission of protecting humanity from cybercrime through artificial intelligence and understanding human behavior to detect and prevent social engineering attacks. Wellington Management Co. led the Series D round, with participation from existing investors Greylock Partners, Menlo Ventures Management, Insight Partners, and CrowdStrike Falcon Fund.
SK Hynix receives $950 million in grants and loans from US government
SK Hynix is about to receive a $450 million CHIPS Act grant and a $500 million loan from the U.S. government to build an AI chip factory in Indiana that focuses on advanced packaging and research, which will strengthen the U.S.'s capabilities in the artificial intelligence supply chain. The total investment in this factory project is $3.87 billion, and it is expected to create about 1,000 jobs. The factory will focus on providing high-bandwidth memory chip packaging services for AI chip manufacturers such as Nvidia.
So far, the Biden administration has announced subsidies for the world's top five chipmakers, including SK Hynix, Samsung Electronics, Micron Technology, Intel and TSMC. SK Hynix's major customers include Apple, Microsoft and Google.
Intelligent Robotics Company Yunshenchu Receives Series C Investment
Yunshenchu Technology is a developer of intelligent robot products, focusing on the development and marketing of legged robot products. Its "Chitu" quadruped robot is the first quadruped bionic robot in China that can run fast. Recently, the industrial and commercial change information showed that Hangzhou Yunshenchu Technology Co., Ltd. (DeepRobotics) completed its C round of financing, with new investors including Hansong Asset Management, Huajian Function Investment, Shenzhen Zhicheng Group and Moganshan High-tech Investment.
Core Interconnect, a digital-analog mixed signal chain chip design and development company, completes RMB 300 million Series C financing
Core Interconnect is a mixed-signal chain chip design and development company dedicated to empowering industry through technological innovation in multiple dimensions and levels, including chips, algorithms, artificial intelligence, communications, and software. It has the world's original chip agile design technology, AI industrial defect monitoring engine, and HPC high-performance computing and high-performance, low-power edge computing. It can deeply integrate application scenarios to provide one-stop signal chain solutions.
Core Interconnect has completed a C round of financing of more than RMB 300 million. This round of financing was led by CICC Capital, with participation from Guangzhou Zengcheng Industrial Investment Group and Zhongshan Torch Development Zone Huaying Investment.
SynSense Technology, a provider of full-scale brain-like perception and computing technology and application solutions, has completed a strategic financing of hundreds of millions of yuan, with heavy investment from major companies such as Ningbo Tongshang Fund and Samsung
SynSense Technology has always focused on edge computing application scenarios. It is a full-stack brain-like intelligent technology company that has high-efficiency perception + computing underlying technology and provides customized chip + software integrated hardware and software solutions.
This round of investment was led by Ningbo Tongshang Fund, and Samsung and other major companies became new shareholders. The funds will be used for mass production of products such as DVS high-speed cameras and the new generation of fusion cameras Aeveon™. At the same time, the company will further strengthen exchanges and cooperation with universities and research institutions such as the Chinese Academy of Sciences, develop brain-like high-computing edge inference chips, and accelerate the application of existing technologies and products in more fields such as industrial security, consumer electronics, aerospace, and autonomous driving.
Industrial analytics, AI and monitoring company Seeq raises $50 million in Series D funding
Seeq specializes in advanced analytics, artificial intelligence, and enterprise monitoring services for industrial companies. The company provides a self-service enterprise SaaS platform and solutions to accelerate critical insights and actions from historically untapped data.
Sixth Street Growth led the round, which brings the total funding to $165 million, with participation from existing investors including Insight Partners, Altira Group, Second Avenue Partners and Saudi Aramco Energy Ventures. Nari Ansari, managing director of Sixth Street Growth, will join Seeq’s board of directors.
Open source data science platform Knime raises $30 million
Knime has created what it claims is the world's most complete "end-to-end data science" open source platform. It is used for AI development and includes tools for creating and deploying analytical models, sharing insights, and using them across the organization. The platform is centered around an intuitive, low-code user interface, combined with generative AI, allowing anyone to quickly build simple and complex analytics for AI models. It provides tools for tasks such as data preparation for training AI models, including large language models. It has now received $30 million in new investment from its long-term supporter Invus.
Apromore, a provider of AI-driven process mining and digital solutions, raises $15 million in Series B-1 funding
Apromore's technology enables business teams to quickly use advanced data science techniques for digital transparency and operational services in a completely code-free environment. Building on its digital twin capabilities and code-free user experience, the company will continue to expand the platform's capabilities through an AI optimization engine to recommend opportunities to simplify, automate and standardize activities and workflows, and for efficient discovery, analysis and simulation of end-to-end processes.
The round was led by existing investors, with participation from Tin Alley Ventures and other investors. This financing brings the total Series B funding to $25.3 million/$38.7 million.
AI cybersecurity company Aurascape AI receives $12.8 million in seed round funding
Aurascape AI is leveraging patent-pending AI innovations to improve cybersecurity for the widespread adoption of AI by enterprises. The company is currently conducting proof of concepts (POCs) with select customers and plans to launch its first commercial product in late 2024.
The round was led by Mayfield Fund. Other investors and shareholders include Celesta Capital, StepStone Group and AISpace, as well as several industry leaders including Mark McLaughlin, former chairman and CEO of Palo Alto Networks.
Digital employee robot company "Yida Technology" completed B+ round of financing: will increase investment in large models and build a digital employee robot platform
Yida Technology focuses on the large supply chain (manufacturing, logistics distribution, wholesale and retail), relying on its original six core hyper-automation technologies to build a digital employee robot hyper-automation platform, helping thousands of leading large supply chain companies to achieve business hyper-automation. This round of financing was led by CDH VGC (Innovation and Growth Fund), followed by New Shang Capital, and Index Capital continued to serve as the exclusive financial advisor.
High-precision map comprehensive solution service provider "Kuandeng Technology" completed nearly 100 million yuan in B+ round financing
Kuandeng Technology is committed to promoting the widespread application of autonomous driving through the intelligent crowdsourcing advanced map business model. Its core technologies include deep learning, image recognition, 3D vision, intelligent robots, map construction, and big data cloud services based on them. It is the first technology startup in China that focuses entirely on crowdsourcing high-precision maps, providing complete data services for the entire autonomous driving ecosystem. This round of financing was jointly invested by Guangdong Rongtai Capital and Zhejiang Deqing Government Industry Fund.
AI career platform Impactpool receives €3.6 million in Series A funding
Impactpool, an AI-powered career platform that helps organizations in impact-driven industries connect with talent across the entire work spectrum and makes it easier for everyone, regardless of socioeconomic background, to find a job, announces €3.6 million in Series A funding led by Mediahuis and Fort Knox.
AI-driven robotics software company Phinxt Robotics raises £2 million in funding
Phinxt Robotics provides a
platform
that enables any type of robot to collaborate
, creating an autonomous ecosystem for warehouses and the entire supply chain. It adopts a per-robot rental model to enable low-cost adoption and provide customers with the most suitable robot. The platform can coordinate edge robots and easily add new robots at a low cost. Starting with warehouse automation, e-commerce providers can use it to double profit margins, increase productivity fourfold, and get more work done with less expenditure.
The round was led by Sure Valley Ventures, with participation from Ada Ventures, Atlas Ventures and Heartfelt Ventures.
Web3AI network DeAgentAI completes $6 million seed round of financing
DeAgentAI is committed to building the world's first AI agent intelligent neural network in the Web3 field, relying on its self-developed LLM model and AGI system to solve the problem that the existing agent system cannot meet the endogenous needs of users. Through its neural network, DeAgentAI can efficiently analyze user intentions and meet user needs with the highest intelligence.
This round of financing was led by Web3.com Ventures and Vertex Capital, with participation from Higgs Capital, Kernel Labs, Waterdrip Capital, Tido Capital, PANONY, Chaincatcher, Goplus, UXLINK and others.
AI startup DemoSquare raises €1.28 million for its artificial intelligence technology
DemoSquare provides an AI-driven SaaS platform designed to help businesses effectively navigate complex regulations and stay ahead of legislation. DemoSquare's founders Dr. Victor Kristof (CEO) and Dr. Jérémie Rappaz (CTO) conducted their PhD research in machine learning and artificial intelligence at the Ecole Polytechnique Fédérale de Lausanne (EPFL). QBIT Capital led the round and was supported by co-investor High-Tech Gründerfonds (HTGF).
AI improves the accuracy of space weather forecasts, Perceptive Space receives $2.8 million in funding
Perceptive Space's innovation lies in its "AI-first" strategy, which combines traditional space weather forecasting methods with sensor fusion technology and analyzes decades of space weather and operational data.
Suresh, the company's CEO, said that based on this method, Perceptive Space's prediction accuracy can reach 10 times that of existing models. Suresh founded the company in 2022 and has extensive experience in space weather modeling and artificial intelligence product development. The team currently consists of five experts, whose members come from well-known institutions such as NASA, Los Alamos National Laboratory, and Massachusetts Institute of Technology. This financing has been supported by several investment institutions focusing on artificial intelligence and space technology.
AI-driven software-as-a-service company Board Intelligence receives funding from K1
Founded in 2010 by founders Jennifer Sundberg and Pippa Begg, Board Intelligence provides software and services that drive better decision making in boards of directors. The company’s AI-powered report writing platform helps management teams prepare shorter, clearer reports for the board of directors.
The company has raised most of the funding from software-focused investor K1, which will be used to expand internationally, make more acquisitions, and expand its AI product range.
GenAI process automation solution Fisent raises funding
Under the leadership of founder and CEO Adrian Murray, Fisent's BizAI solution achieves end-to-end automation of repetitive business tasks by applying basic GenAI models, reducing the workload of manual interpretation of different types of data and content, thereby improving back-end efficiency.
The amount of its financing has not been disclosed, and its supporters include Pegasystems, Sand Dollar Capital, etc.
(Welcome to add WeChat
AIyanxishe2
to learn more about AIGC and financing, and chat about new AI products with like-minded friends)
OpenAI API introduces structured output function to help developers intelligently connect external tools
OpenAI introduced structured output functionality in its API, which enables developers to provide JSON schemas and ensure that the outputs generated by the model precisely conform to those schemas.
This feature supports two forms: function calls and response format parameters. Function calls are implemented by setting strict: true in the function definition, while the response format parameter allows developers to provide a JSON schema through the response_format parameter. This new feature is already supported in the gpt-4o-2024-08-06 model, which scored 100% in OpenAI's evaluation, significantly exceeding the gpt-4-0613 model's score of about 40%. The prices of the new model have also been adjusted, with the input price reduced by half and the output price reduced by one-third.
Reddit to test AI-powered search results page
Reddit plans to introduce AI-generated search result summaries. Reddit co-founder and CEO Steve Huffman announced in a recent earnings call that the company will test AI-driven search results pages to help users gain deeper insights into content and discover new communities. This decision comes in the context of Reddit's collaborations with OpenAI and Google, which have enabled Reddit to leverage large language models to build new features. In addition, Reddit is also pushing its AI language translation capabilities and plans to expand it to more languages.
Dark Side of the Moon Kimi Open Platform: Context Cache storage fee reduced by 50%
Kimi Open Platform under Dark Side of the Moon announced that it will reduce the storage fee of context cache by 50%, from 10 yuan/1M tokens/min to 5 yuan/1M tokens/min. The platform announced the start of public beta of context cache on July 1. This technology is particularly suitable for scenarios with frequent requests and repeated references to a large number of initial contexts, which can significantly reduce the cost of long text models and improve efficiency.
Tencent Yuanbao launches in-depth reading mode: supports up to 500,000 words of input, can extract papers, generate DuPont analysis charts
In the latest version of Tencent Yuanbao, when users upload URL links or files of professional content such as papers, financial reports, and research reports, in addition to obtaining a text summary, they can also enter the deep reading mode to read the long text carefully. The deep reading mode can natively support input of up to nearly 500,000 words, and can generate graphic and text content based on the understanding of professional content. It can not only quote the original image of the input content, but also draw high-level analysis diagrams through code.
Baidu Netdisk launches new photography industry solution
Baidu Netdisk released a new photography industry solution, including a new generation of commercial AI portrait photography product - Baidu Netdisk AI Photo Editing. It is reported that this product can help photo retouchers reduce more than 70% of repetitive work. At the same time, the "Super Canvas" has also been upgraded to the Pro version, providing a one-stop industry solution for photo studios.
Peking University/Institute of General Studies proposes a super-difficult benchmark to evaluate the understanding and generation of long texts, and large models are completely defeated
Peking University and the Beijing General Artificial Intelligence Research Institute have released
the LooGLE benchmark, which is used to evaluate the understanding ability of large language models for long texts. LooGLE contains 776 very long documents after 2022, each containing an average of 19.3k words, covering multiple fields such as academic, history, sports, etc. The benchmark is characterized by the inclusion of very long real documents, manually designed long-short dependency tasks, relatively novel documents, and cross-domain general data.
By evaluating 8 state-of-the-art LLMs, the study found that commercial models outperformed open source models. LLMs performed well on short-dependency tasks, but faced challenges on more complex long-dependency tasks. Context-based learning and thought chaining methods provided only limited improvements on long-context understanding, while retrieval-based techniques showed significant advantages in short question answering. All evaluation codes have been released on GitHub.
Code ????https://github.com/bigai-nlco/LooGLE
Scamnetic launches AI fraud detection platform that detects fraudsters in real time
Scamnetic Inc. announced that its AI-powered fraud detection platform is now open to software providers, and can be integrated into customers' software via APIs to provide comprehensive fraud protection. The platform can detect fraudulent activities in real time, including email, SMS, chat platforms, phone calls, and other channels, and assess potential fraud risks by scanning the content of communications. Scamnetic provides two tools, Scan&Score and IDeveryone, for automatically assessing communication risks and verifying the identity of counterparties, respectively. In addition, Scamnetic also provides fraud intervention tools to provide victims with 24/7 support to help them recover stolen assets.
Product Hunt
Hot List, Me.bot
Me.bot is an intelligent companion focused on personalized experience, designed to provide personalized guidance and support by learning and adapting to the user's unique experience and interaction. It helps users build memory archives through multimodal recognition and connects memory points to provide inspiration and support.
Founder Felix Tao, as the founder and CEO of Mindverse AI, led the team to develop the MindOS platform, which uses natural language to enable users to easily create autonomous AI agents. He served as the head of the Neuro-Symbolic Laboratory at Alibaba Cloud, focusing on giving AI systems cognitive capabilities. During his time at Facebook, he established a content understanding platform in news push. Felix Tao holds a degree in computer science from Tsinghua University and a Ph.D. in computer science from the University of Illinois at Urbana-Champaign. His doctoral research was supervised by Professor Jiawei Han, an expert in data mining. He has also participated in multiple projects, including research insight systems and the design of trusted mobile social networks based on ad hoc networks.
???? https://www.me.bot/?ref=producthunt
1. ByteDance’s “Jimeng AI”, which competes with Kuaishou, Keling and Sora, has been launched on Apple App Store and Android App Store
This app was developed by the Jianying and Xingtu teams (Shenzhen Lianmeng Technology Co., Ltd.) and uses an in-app purchase model. The membership service is 69 yuan per month (79 yuan per month) or 659 yuan per year, corresponding to the purchase of 505 points/month (which can generate about 2,050 pictures or 168 AI videos).
????https://jimeng.jianying.com/
2. Canva releases one-stop AI creation suite "Magic Studio": supports generating flower characters and transition animations
Canva announced that its one-stop AI creation suite "Magic Studio" has been officially launched in China. The suite is known as "a one-stop AI creation suite for everyone", and its functions include AIGC-based copywriting generation, image generation, flower font special effects generation, image editing, transition animation design generation, etc.
????https://www.canva.cn/magic/#features
3. DomoAI upgrades the “/move” function to support green screen and lip sync
The “/move” feature has been super upgraded, allowing users to not only make their favorite characters dance, but also supports green screen backgrounds and lip sync functions, allowing characters to dance in various scenes and achieve lip movements synchronized with the characters.