Skip to content
SummHub: Your AI-Powered News Companion
Menu
  • Home
  • AI
  • Tech
  • Icons
  • News
Menu

[2025-02-20] Global AI News

Posted on 20/02/202520/02/2025 by iflux

AI Guide

AI Laptops

Current AI Laptop Availability

Several AI-powered laptops showcased at CES 2025 are now available for purchase or pre-order. Noteworthy models include the Lenovo Yoga 9i 2-in-1 Aura Edition, featuring enhanced AI processing and a stylish design, available through Lenovo's store. The Acer Swift Go 14 AI, a more affordable Copilot+ PC equipped with a Qualcomm Snapdragon X Plus chip, is also an option. The Asus ZenBook A14, lauded for its design, battery life, and OLED display, along with the gaming-centric Asus ROG Flow Z13 tablet, powered by AMD's Ryzen AI Max+ processors, are also currently available.

image

News List

Mashable

2 hours ago

Heres how you can get some of our favorite AI laptops from CES 2025

CES 2025 showcased a strong presence of AI in laptops, with notable models from Asus, Acer, and Lenovo. While many are still awaiting release, some AI laptops are already available for purchase or pre-order. The Lenovo Yoga 9i 2-in-1 Aura Edition, featuring improved AI processing and a sleek design, is available on Lenovo’s store. The Acer Swift Go 14 AI, equipped with a Qualcomm Snapdragon X Plus chip, is a more budget-friendly Copilot+ PC. The Asus ZenBook A14, praised for its design, battery life, and OLED display, is also available, alongside the gaming-focused Asus ROG Flow Z13 tablet, utilizing AMD’s Ryzen AI Max+ processors.

AI Guide

AI Technology

AI-Powered Agents and the Lessons from Boo.com

image In late 1999, the online fashion startup Boo.com launched its website featuring a virtual shopping assistant named Miss Boo. Miss Boo was designed to offer styling tips and allow customers to view clothes on 3D models before purchasing. Although the concept is similar to current AI-powered agents, Boo.com collapsed shortly after launch due to a slow and unreliable website. The technology industry is now betting that AI-powered agents will be more successful.

News List

Reuters.Tech

2 hours ago

AI agents have clear mission, hazy business model

In late 1999 the online fashion startup Boo.com launched its website, which featured a virtual shopping assistant. The avatar, known as Miss Boo, was supposed to offer styling tips to customers and allow them to view jeans and shirts on three-dimensional models before buying. The concept may sound familiar to anyone who has recently listened to technology companies extoll the potential of agents using artificial intelligence. Boo.com collapsed a few months after it launched, in part because the website proved slow and unreliable. The technology industry’s hopes are riding on a bet that AI-powered agents will be more successful this time.

AI Guide

Food Industry

Spore.Bio Secures $23 Million for Rapid Microbiology Testing

Spore.Bio, a deep tech startup based in Paris, has raised $23 million in a Series A round led by Singular. The company is focused on revolutionizing microbiology testing for the food industry by using spectral signatures and deep learning algorithms to detect bacteria and pathogens rapidly. This technology significantly reduces testing time, from several days to a shorter duration, addressing the critical need for rapid testing to prevent contamination incidents. Spore.Bio has already signed commercial contracts covering up to 200 factories and has partnered with the Pasteur Institute.

News List

TechCrunch

4 hours ago

Spore.Bio raises $23M to apply machine learning to microbiology testing

Spore.Bio, a Paris-based deep tech startup focused on reinventing microbiology testing for the food industry, has raised a $23 million Series A round led by Singular. The company’s technology aims to provide faster microbiological tests, reducing the time from several days to a much shorter duration by using spectral signatures and deep learning algorithms to detect bacteria and pathogens. This innovation addresses the critical need for rapid testing in the food and beverage industry to avoid contamination incidents and long-lasting damage to brand reputation. Spore.Bio has already signed commercial contracts covering up to 200 factories and has partnered with the Pasteur Institute.

AI Guide

Artificial Intelligence

AI Model Development

Grok-3 hasn't been fully released yet, but it's expected to influence how other AI labs release future models.

AI-Powered Coding Startup Funding

Codeium, an AI-powered coding startup, is reportedly raising a new round of funding led by Kleiner Perkins at a $2.85 billion valuation, a significant increase from its $1.25 billion valuation six months prior. The company has reached approximately $30 million in annualized recurring revenue (ARR), resulting in a valuation multiple of roughly 95 times ARR. Codeium targets enterprise customers, with over 1,000 companies using its free tier, including Anduril, Zillow, and Dell. They also introduced Windsurf Editor in November, capable of autonomous code writing.

News List

TechCrunch

8 hours ago

AI-coding startup Codeium in talks to raise at an almost $3B valuation, sources say

AI-powered coding startup Codeium is raising a new round of funding at a $2.85 billion valuation, led by Kleiner Perkins. This comes just six months after a $150 million Series C at a $1.25 billion valuation. Codeium has reached about $30 million in annualized recurring revenue (ARR), implying a valuation of roughly 95 times ARR. Codeium targets companies rather than individual developers, with its free tier being used by over 1,000 enterprise customers, including Anduril, Zillow, and Dell. In November, it introduced Windsurf Editor, which can write some of the code without human involvement, similar to Cursor.

VentureBeat

9 hours ago

Breaking down Grok 3: The AI model that could redefine the industry

Grok-3 still hasn’t fully shipped yet. But it will surely set the tone for how other AI labs release future models.

AI Guide

Artificial Intelligence and Power Flow Models

Computational Efficiency of Power Flow Methods

PPFL methods demonstrate superior computational efficiency compared to DPFL approaches. This advantage stems from PPFL methods circumventing the training process, unlike DPFL methods, due to their reliance on predefined physical models. However, the accuracy of PPFL methods hinges on the assumption that all necessary physical parameters are known and accurate, which may not always hold true, especially in distribution grids. The computational costs of most DPFL methods grow at a superlinear rate.

image

News List

HackerNoon

13 hours ago

Why Some AI Power Flow Models Are Faster Than Others

The evaluation of computational efficiency compares DPFL and PPFL methods, provides a general review of DPFL methods, and individually examines specific groups of DPFL approaches. PPFL methods outperform DPFL approaches in computational efficiency, as PPFL methods avoid the need for a training process, unlike DPFL methods. The absence of training in PPFL methods is attributed to their irreliance on predefined physical models of the system, assuming all necessary physical parameters are known and accurate. However, this assumption might not always be valid, particularly in distribution grids. The computational costs of the majority of DPFL methods exhibit superlinear growth.

AI Guide

AI Development

Rabbit's Android Agent Progress

Rabbit is actively developing its "generalist Android agent", designed to control apps on a tablet, mirroring the intended functionality of the R1 device. Currently, engineers input requests via a laptop prompt box, which are then translated into actions on an Android tablet. Tasks include finding YouTube videos, locating cocktail recipes, and adding ingredients to a Google Keep grocery list. This AI agent is continually being updated, with recent improvements including the ability to train agents for specific tasks and prompt interface redesigns. Rabbit plans to reveal more about its "cross-platform multi-agent system" soon.

News List

The Verge

10 hours ago

Rabbit shows off the AI agent it should have launched with

Rabbit is still developing its “generalist Android agent,” which slowly controls apps on a tablet, similar to what Rabbit claimed its R1 device would do. The engineers type requests into a prompt box on a laptop, which translates them to actions on an Android tablet, such as finding a YouTube video, locating a cocktail recipe, and adding ingredients to a Google Keep grocery list. The AI agent is still a work in progress, and Rabbit has been steadily rolling out updates, like the ability to train its AI agents to complete specific tasks or prompt it to remake its own interface. The company promises to share more about its “upcoming cross-platform multi-agent system” in the coming weeks.

AI Guide

Gaming Industry

Microsoft's Muse AI Model for Xbox Game Development

Microsoft's Xbox division unveiled Muse, a generative AI model designed to assist game developers in creating visuals and gameplay. Trained on Ninja Theory’s Bleeding Edge, Muse understands game physics and 3D environments, generating visuals and reacting to player movements. One potential application is game preservation, where Muse AI could optimize classic titles for modern hardware. Microsoft aims to further explore generative AI to aid game teams in prototyping projects, conducting interviews with 27 game creators globally to guide the research. However, the announcement has faced criticism from developers and the online community, who argue that generative AI devalues the aesthetic efforts of game developers and artists, potentially leading to a loss of craft and empowering tool owners. Concerns are also rising about AI potentially displacing human workers in game creation amid industry layoffs.

image

News List

Wired

11 hours ago

Xbox Pushes Ahead With Muse, a New Generative AI Model. Devs Say ‘Nobody Will Want This’

Microsoft has announced Muse, a new AI model designed to assist Xbox game developers in building games. Muse, trained on Ninja Theory’s Bleeding Edge, understands game physics and 3D environments, generating visuals and reactions to player movements. One intriguing use case involves game preservation, where Muse AI could optimize classic titles for modern hardware. Microsoft aims to explore generative AI further to help game teams prototype projects, interviewing 27 game creators globally to shape the research. However, the announcement has been poorly received by developers and the online community. Critics argue that generative AI devalues the aesthetic efforts of game developers and artists, potentially leading to a loss of craft and empowering those who own the tools. Concerns are rising about AI pushing human workers out of game creation, amidst industry layoffs.

VentureBeat

11 hours ago

Xbox’s AI initiative with Muse is an attempt to read the tea leaves, not the room

Earlier today, the heads of Microsoft’s Xbox division revealed Muse, a generative AI model that intends to create both visuals and gameplay for games. The model, which was trained on the largely forgotten Ninja Theory multiplayer game Bleeding Edge, is not a shocking leap for Microsoft’s Xbox division. The company as a whole, from CEO […]

AI Guide

AI Model Updates

Mistral's Regional Language Model: Saba

Paris-based AI startup Mistral is focusing on providing large language models (LLMs) that understand regional-specific languages and are tailored to grasp cultural nuances. Mistral has released its first "specialized" regional language-focused model, Saba, a 24-billion-parameter model trained on datasets from across the Middle East and South Asia to meet a growing customer base in Arabic-speaking countries. Saba is similar in size to Mistral Small 3 but performs better at handling Arabic content and excels with South Indian languages like Tamil and Malayalam. Mistral's benchmark tests show that Saba outperforms Arabic-centric models and multilingual LLMs while providing more accurate and relevant responses than models over 5 times its size, and can also be a base to train highly specific regional adaptations.

News List

ZDNet.AI

12 hours ago

Mistral’s new AI model specializes in Arabic and related languages

Paris-based AI startup Mistral is focusing on providing large language models (LLMs) that understand regional-specific languages and are tailored to grasp cultural nuances. Mistral has released its first “specialized” regional language-focused model, Saba, a 24-billion-parameter model trained on datasets from across the Middle East and South Asia to meet a growing customer base in Arabic-speaking countries. Saba is similar in size to Mistral Small 3 but performs better at handling Arabic content and excels with South Indian languages like Tamil and Malayalam. Mistral’s benchmark tests show that Saba outperforms Arabic-centric models and multilingual LLMs while providing more accurate and relevant responses than models over 5 times its size, and can also be a base to train highly specific regional adaptations.

AI Guide

AI Applications and Research

Stytch Connected Apps

Stytch Connected Apps enables AI agents and third-party apps to authenticate, access data, and take action on behalf of users, streamlining authorization.

AI in Research

Google's new AI is aiding real-world research by generating hypotheses for researchers.

image

Mobile Technology

Apple's iPhone 16e Launch

Apple has launched the iPhone 16e, priced at $599, as its most affordable AI-equipped iPhone. It features a 6.1-inch full-screen design with Face ID, a USB-C port, a 48MP camera, and the Action button. Powered by the A18 chip, it supports Apple Intelligence, including generative AI writing tools, Genmoji, Image Playground, Visual Intelligence, and ChatGPT integration.

image

AI Regulation and Policy

Changes in US AI Regulation

President Trump overturned former President Joe Biden's October 2023 executive order on AI regulation, replacing it with his own. Government agencies were instructed to halt AI regulation activities, and the director of the USAI Safety Institute (AISI) resigned. Mass firings are reportedly planned at the National Institute of Standards and Technology (NIST), affecting AISI.

News List

ZDNet.AI

13 hours ago

The head of US AI safety has stepped down. What now?

A recent executive order signed by former President Joe Biden in October 2023, which included measures for regulating AI, was overturned by President Trump, who replaced it with his own order on AI in the US. Consequently, government agencies enforcing AI regulation were instructed to halt their activities, and the director of the USAI Safety Institute (AISI) resigned. The National Institute of Standards and Technology (NIST) is reportedly planning mass firings that would further affect AISI. Biden’s order established the USAI Safety Institute (AISI), which collaborated with developers to improve safety measures. Earlier this month, AISI director Elizabeth Kelly announced her departure from the institute. The fate of these initiatives, and the institute itself, is now uncertain.

Ars Technica

13 hours ago

Google’s new AI generates hypotheses for researchers

And it’s already helping with some real-world research.

VentureBeat

15 hours ago

A Stytch in time: Connected Apps untangles authorization tie-ups for AI agents

Stytch Connected Apps enables AI agents and third-party apps to authenticate, access data and take action on behalf of users.

Engadget

16 hours ago

Apple announces the iPhone 16e with Apple Intelligence for $599

Apple has launched its new entry-level iPhone 16e, succeeding the iPhone SE. Priced at $599, it’s the most affordable AI-equipped iPhone. The iPhone 16e resembles the iPhone 14, featuring a 6.1-inch full-screen design with Face ID, a USB-C port, and a single 48MP camera with 2x zoom. It also includes the Action button. Powered by the A18 chip, it supports Apple Intelligence, offering generative AI writing tools, Genmoji, Image Playground, Visual Intelligence, and ChatGPT integration. Available in 128GB, 256GB, and 512GB storage options, the iPhone 16e is competitively priced against the Google Pixel 8a and Samsung Galaxy A35.

AI Guide

AI Models and Applications

Microsoft's Interactive AI World Model

Microsoft's new model shows improvements but is primarily suited for low-resolution prototypes. image

xAI's Grok3

xAI released Grok3, its latest AI model, which it claims outperforms leading models on math and programming benchmarks. However, experts are questioning the value of these benchmarks and calling for independent testing authorities.

Klok – A Multi-LLM Chat App

Mira Network launched Klok, a multi-LLM chat app built on its decentralized infrastructure. Klok allows users to access multiple AI models like GPT-4omini and Llama3.370BInstruct for free and earn Mira Points. It aims to address AI hallucinations through consensus mechanisms. image

Google's Career Dreamer

Google has launched Career Dreamer, an AI-powered tool to help users explore career possibilities by analyzing their experiences, skills, and interests. It is currently available in the United States.

AI Hardware and Gadgets

Humane's AI Pin Debacle

The Humane AI Pin failed due to its high price ($700 plus subscription) and unclear value, struggling with basic tasks. HP acquired Humane for $116 million. image

Humane Bricking AI Pins

Humane is facing criticism for effectively bricking 700 AI Pins with limited refunds, demonstrating how not to treat early adopters.

Rabbit R1

Rabbit gained attention with its $199 R1 gadget, selling 40,000 units shortly after its launch. Recent updates include user-programmable tasks, but it was revealed that only 5,000 people actively use the device.

AI and Cybersecurity

AI-Powered Cyberattacks

Attackers are leveraging AI to craft hyper-personalized scams, poison training data, and fool real-world systems. Organizations should adopt better verification protocols, train employees, and use AI-powered phishing drills. It is important to consider that AI attacks are also expanding to the physical world. image

Financial and Policy Implications

China ETFs and AI

U.S.-listed exchange-traded funds focused on China have seen bullish options flows, driven by the emergence of AI startup DeepSeek and easing tariff concerns. image

US AI Safety Institute Cuts

The Trump administration is planning to cut the USAI Safety Institute, potentially eliminating 497 roles at NIST, impacting AI safety and regulation efforts and semiconductor production initiatives.

News List

Reuters.Tech

14 hours ago

China ETFs draw bullish options bets on AI play and eased tariff fears

U.S.-listed exchange-traded funds focused on China have drawn bullish options flows as the emergence of artificial intelligence startup DeepSeek brightens the outlook for Chinese tech shares, and tariff-related news seemed less threatening than before.

ZDNet.AI

14 hours ago

US AI Safety Institute will be ‘gutted,’ Axios reports

The Trump administration is planning to cut the USAI Safety Institute (AISI), potentially gutting it by eliminating 497 roles at the National Institute of Standards and Technology (NIST). These cuts will impact not only AI safety and regulation efforts, including partnerships with companies like Anthropic and OpenAI, but also semiconductor production initiatives under the “CHIPS for America” program. This decision comes shortly after AISI staff were excluded from the AI Action Summit in Paris and the resignation of AISI director Elizabeth Kelly, signaling a de-emphasis on AI safety in favor of “AI dominance.”

Engadget

14 hours ago

The Humane AI Pin debacle is a reminder that AI alone doesn’t make a compelling product

The AI Pin, launched by Humane, failed to gain traction due to its high price ($700 plus a $24/month subscription) and unclear value. Released in April 2024 to poor reviews, the device struggled with basic tasks like answering questions and taking photos, suffering from issues like overheating and a fire hazard-prone battery case. HP acquired Humane for $116 million after initial rumors suggested a $750 million to $1 billion valuation. The Pin’s failure highlights the immaturity of consumer-grade AI and the challenges of creating compelling AI hardware.

Ars Technica

14 hours ago

“Truly a middle finger”: Humane bricking $700 AI Pins with limited refunds

Humane’s showing how not to treat early adopters.

Ars Technica

14 hours ago

Microsoft shows progress toward real-time AI-generated game worlds

Despite improvements, Microsoft’s new model is still mainly useful for low-res prototypes.

TechCrunch

14 hours ago

This Week in AI: Maybe we should ignore AI benchmarks for now

xAI, Elon Musk’s AI startup, has released Grok3, its latest flagship AI model, which outperforms several leading models on benchmarks for mathematics and programming. However, the value of these benchmarks is questionable, as they often test for esoteric knowledge and are self-reported by AI companies. Experts are calling for better, independent testing authorities and benchmarks that align with economic impact. It’s suggested that less attention be paid to new models and benchmarks unless there are major AI technical breakthroughs.

TechCrunch

14 hours ago

Google’s ‘Career Dreamer’ uses AI to help you explore job possibilities

Google has launched Career Dreamer, an AI-powered experiment designed to help people explore career possibilities. The tool uses AI to analyze a user’s experiences, skills, and interests to suggest potential career matches and create a career identity statement. Career Dreamer also allows users to collaborate with Gemini, Google’s AI assistant, to workshop cover letters and resumes. Currently available only in the United States, Career Dreamer aims to simplify the career exploration process.

The Verge

14 hours ago

The Rabbit R1: all the latest news about this hare-raising AI gadget

Rabbit, an AI startup, has garnered attention with its $199 R1 gadget, selling 40,000 units within eight days of its launch at CES. The Rabbit R1 features Rabbit OS, a dedicated virtual assistant built on a Large Action Model that interacts with apps like Spotify and Uber. Designed in collaboration with Teenage Engineering, the device includes a touchscreen, scrolling navigation wheel, and rotating camera. Recent updates include allowing users to teach the R1 to perform tasks, though it was revealed that only 5,000 people are actively using the device at any given time.

HackerNoon

16 hours ago

AI Can Outsmart You, and Cybercriminals Know It

AI is being leveraged by attackers to craft hyper-personalized scams, poison training data, and fool real-world systems, necessitating a holistic and adversarial view of cybersecurity systems. Examples include deepfake impersonation in financial fraud and the alteration of medical reports. Organizations should adopt better verification protocols, train employees to recognize phishing attempts, and use AI-powered phishing drills. It is also important to consider that AI attacks are also expanding to the physical world and security measures should be considered to prevent them.

HackerNoon

17 hours ago

Mira Network Launches Klok: A ChatGPT Alternative with Multiple AI Models and Rewards

Mira Network has launched Klok, a multi-LLM chat app built on its decentralized infrastructure to ensure AI outputs are verified, unbiased, and uncompromised. Klok allows users to access multiple AI models like GPT-4omini and Llama3.370BInstruct for free and earn Mira Points. The platform addresses AI hallucinations by using consensus mechanisms among multiple LLMs. Mira has raised over $9 million from venture capital firms and aims to build a trust layer for AI and crypto.

AI Guide

AI

AI Security Concerns

Security firm Adversa AI jailbroke xAI's Grok-3 just a day after its release. The jailbreak revealed significant safety and security vulnerabilities, including revealing its system prompt and providing instructions for making a bomb. Adversa AI's report emphasizes that Grok-3's safety measures are weak.

AI in Legal Settings

A law firm condemns the careless use of AI in court, calling it "nauseatingly frightening".

image

News List

Ars Technica

15 hours ago

AI making up cases can get lawyers fired, scandalized law firm warns

“Nauseatingly frightening”: Law firm condemns careless AI use in court.

ZDNet.AI

15 hours ago

Yikes: Jailbroken Grok 3 can be made to say and reveal just about anything

Just a day after its release, xAI’s latest model, Grok-3, was jailbroken by security firm Adversa AI, revealing significant safety and security vulnerabilities. The firm successfully used linguistic, adversarial, and programming methods to get the model to reveal its system prompt, provide instructions for making a bomb, and offer gruesome methods for disposing of a body. Adversa AI’s report emphasizes that Grok-3’s safety measures are weak compared to its competitors, with every jailbreak approach succeeding. Despite Elon Musk’s claims of Grok-3’s advanced capabilities, its lack of guardrails, a deliberate design choice, raises concerns, especially given past instances of misinformation and dicey image generations.

AI Guide

Artificial Intelligence

AI-Powered Game Development

Microsoft is making strides in AI game development with its new Muse AI system. Muse can generate interactive 3D game worlds by analyzing gameplay footage, demonstrating advanced AI capabilities in understanding spatial environments and human interactions. Furthermore, Microsoft researchers have created Muse as a generative AI model specifically for game ideation.

AI-Driven Startup Predictions

Crunchbase is now using AI for predictive analysis, achieving 95% accuracy in forecasting startup funding rounds. This pivot involves abandoning historical data in favor of predictive intelligence that forecasts acquisitions and company growth, potentially changing investment strategies.

AI Language Tutoring

The author tested ChatGPT's Advanced Voice Mode as an AI language tutor while traveling in Japan, practicing basic travel phrases for transportation, restaurants, and retail shops. image

AI for Freelancers

Fiverr is launching the Personal AI Creation Model, allowing users to train AI on their work. This is part of the larger FiverrGo initiative, aimed at empowering freelancers with AI's automation capabilities. The goal is to make freelancers irreplaceable by using AI as a tool. image

AI Chatbot Competition

Mistral's AI assistant, Le Chat, has reached one million downloads shortly after its release, endorsed by French President Emmanuel Macron. This contrasts with OpenAI's ChatGPT, which has 350 million downloads, as well as competitors such as Google's Gemini and Microsoft's Copilot.

AI in Meetings

Google is enhancing Meet with new features including expanded caption history, allowing users to review the last 30 minutes of live captions. Additionally, AI-generated Next Steps using Gemini will analyze meeting notes to automatically identify and organize actionable tasks.

AI Accent Modification

The voice and speech recognition technology sector is projected to exceed $28.1 billion by 2027. Sanas, founded in 2020, develops AI-powered software that adjusts a speaker’s accent in real-time, aiming to enhance human connection in customer interactions. Sanas acquired InTone to strengthen its IP portfolio and serve a wider customer base, currently serving around 50 customers in healthcare, logistics, and hardware manufacturing.

Symbolic AI Resurgence

Symbolic AI, which emulates human thinking by manipulating symbols rather than numerical data, is regaining momentum. This approach, rooted in early AI history, uses knowledge representation with rules applied to symbols, allowing machines to solve problems in a human-like fashion by modeling and reasoning over symbols. image

Technology

Apple's New Affordable iPhone

Apple has unveiled the iPhone 16e, its newest affordable smartphone, priced at $600. This model offers access to Apple Intelligence features like the Clean Up tool, natural language search in the Photos app, and typing to Siri. The iPhone 16e features a 6.1-inch OLED display with a notch that includes FaceID integration.

News List

KDnuggets

16 hours ago

A Gentle Introduction to Symbolic AI

This article introduces symbolic AI, an area of AI that emulates human thinking by manipulating symbols instead of numerical data. Symbolic AI, rooted in early AI history and involving knowledge representation using rules applied to symbols, is regaining momentum, with some experts suggesting it as a future direction in the field. It allows machines to solve problems in a human-like fashion by modeling and reasoning over symbols.

Gizmodo

16 hours ago

The Apple iPhone 16E Is The Cheapest Way to Use Apple Intelligence

Apple has unveiled the iPhone 16e, its newest affordable smartphone, priced at $600. This model offers access to Apple Intelligence features like the Clean Up tool, natural language search in the Photos app, and typing to Siri. The iPhone 16e features a 6.1-inch OLED display with a notch that includes FaceID integration, a departure from the previous iPhone SE’s TouchID.

TechCrunch

16 hours ago

Sanas taps AI to change call center workers’ accents in real time

The voice and speech recognition technology sector is projected to exceed $28.1 billion by 2027. Sanas, founded in 2020, develops AI-powered software that adjusts a speaker’s accent in real-time, aiming to enhance human connection in customer interactions. Sanas acquired InTone to strengthen its IP portfolio and serve a wider customer base, currently serving around 50 customers in healthcare, logistics, and hardware manufacturing.

VentureBeat

16 hours ago

Crunchbase’s AI can predict startup success with 95% accuracy—will it change investing?

Crunchbase pivots to AI-powered predictions engine with 95% accuracy for startup funding rounds, abandoning historical data in favor of predictive intelligence that forecasts acquisitions and company growth.

ZDNet.AI

16 hours ago

Google Meet’s AI action items and caption history just changed the meeting game

Google is introducing new features for Meet, including expanded caption history and AI-generated “Next Steps.” Users can now scroll back and review the last 30 minutes of live captions, and Gemini will analyze meeting notes to automatically identify and organize actionable next steps. Expanded captions are available for all Google Workspace customers, while Next Steps is rolling out to specific Google Workspace users.

TechCrunch

16 hours ago

Mistral’s Le Chat tops 1M downloads in just 14 days

Mistral’s AI assistant, Le Chat, has reached one million downloads shortly after its release. French president Emmanuel Macron has endorsed Le Chat. OpenAI’s ChatGPT has been downloaded 350 million times. Other competitors include Google’s Gemini and Microsoft’s Copilot.

Mashable

16 hours ago

Fiverr now lets gig workers train AI on their own work and sell it

Fiverr is launching a program called Personal AI Creation Model that allows users to train AI on their own body of work. It’s part of a larger AI initiative called FiverrGo, designed to help freelancers use AI’s automation capabilities as a tool rather than being replaced by it. Fiverr’s CEO says the aim is to make freelancers irreplaceable, not obsolete.

VentureBeat

17 hours ago

Muse is Xbox’s generative AI model for gameplay ideation

Microsoft researchers have created the first generative AI model, Muse, for game ideation.

VentureBeat

17 hours ago

Microsoft’s Muse AI can design video game worlds after watching you play

Microsoft new Muse AI system learns to generate interactive 3D game worlds by watching gameplay footage, marking a major advance in artificial intelligence’s ability to understand spatial environments and human interactions.

Wired

21 hours ago

Before Going to Tokyo, I Tried Learning Japanese With ChatGPT

The author tested ChatGPT’s Advanced Voice Mode as an AI language tutor while traveling in Japan. The author used ChatGPT to practice Japanese and compared the experience to using Duolingo on a previous trip. The author approached ChatGPT as a synthetic language tutor, practicing basic travel phrases for transportation, restaurants, and retail shops.

AI Guide

Artificial Intelligence & Machine Learning

Quantum Computing Advancements

Microsoft has announced a new chip, suggesting that quantum computing is "years, not decades" away. This puts them in line with predictions from Google and IBM, indicating that this technology is approaching faster than previously anticipated.

image

Generative AI and Gaming

Microsoft Research has developed Muse, a generative AI model that can generate game environments from visuals or player actions. Trained on data from the Xbox game Bleeding Edge, Muse understands 3D game worlds and physics, allowing it to generate gameplay visuals and react to player interactions, and adapt to new gameplay elements in real-time. The company is exploring the potential of using Muse to preserve classic games and optimize them for modern hardware. Currently, Muse is limited to generating gameplay visuals at a resolution of 300×180 pixels.

image

AI Model Training and Accessibility

Fiverr is now enabling creators to train AI models on their work. Clients can then purchase access to these custom AI models and generate content. Freelancers can edit and set the price for their custom AI models, which are available across various services.

AI-Powered Research Tool

Google Gemini's DeepResearch is a virtual agent that performs online research. Accessible on the web, Android, and iOS, DeepResearch generates a research plan, searches the web for relevant information, and generates a detailed report. To use DeepResearch, you'll need the $20-per-month Gemini Advanced subscription, which is part of the Google One AI Premium plan.

Open-Source AI Development

DeepSeek, a Chinese AI startup, has developed powerful large language models (LLMs) at a lower cost. DeepSeek is open-source and its models perform as well as or better than other models.

HP Acquires Humane

HP has acquired Humane's CosmOS operating system and IP, including the tech for the AI Pin, for $116 million. HP intends to use the acquired technology to develop its own suite of AI-centric devices and integrate CosmOS into its PCs, smart printers, and connected conference rooms. The devices will be bricked by Feb. 28.

Consumer Electronics

Apple's New iPhone

Apple unveiled the iPhone 16e, a refresh of the budget-minded smartphone, starting at $599 and shipping on February 28. The iPhone 16e includes Apple Intelligence, featuring small models for text summaries, letter writing, and image generation. Powered by the A18 processor and Apple C1 modem, it's part of an exclusive group capable of running Apple Intelligence.

News List

Reuters.Tech

17 hours ago

Microsoft creates chip it says shows quantum computers are ‘years, not decades’ away

Microsoft on Wednesday unveiled a new chip that it said showed quantum computing is “years, not decades” away, joining Google and IBM in predicting that a fundamental change in computing technology is much closer than recently believed.

TechCrunch

17 hours ago

Apple’s  $599 iPhone 16e adds AI, ditches the fingerprint scanner

Apple unveiled the iPhone 16e, a refresh of the budget-minded smartphone, starting at $599 and shipping on February 28. Dropping the SE branding, the iPhone 16e includes Apple Intelligence, featuring small models for text summaries, letter writing, and image generation. Powered by the A18 processor and Apple C1 modem, it’s part of an exclusive group capable of running Apple Intelligence. The design updates include Face ID, USB-C, and the camera notch. This launch arrives during a challenging time for Apple, with a market share decline in China. The original iPhone SE had strong sales in China and India.

Engadget

17 hours ago

Microsoft trained an AI model on a game no one played

Microsoft Research has developed Muse, a generative AI model that can create game environments from visuals or player actions. Trained on data from the Xbox game Bleeding Edge, Muse understands 3D game worlds and physics, allowing it to generate gameplay visuals and react to player interactions. It can adapt to new gameplay elements in real-time, offering a way to control development costs. Muse is currently limited to generating gameplay visuals at a resolution of 300×180 pixels. Microsoft is exploring the potential of using Muse to preserve classic games and optimize them for modern hardware.

The Verge

17 hours ago

Microsoft’s Xbox AI era starts with a model that can generate gameplay

Microsoft is exploring the potential of using its new Muse AI model to help Xbox developers create parts of games and even preserve classic games by optimizing them for modern hardware. The Muse model, created by Microsoft Research, can generate a game environment based on visuals or players’ controller actions. It understands a 3D game world and game physics and can react to how players interact with a game. The Muse model was trained on a large amount of human gameplay data from the Xbox game Bleeding Edge.

ZDNet.AI

17 hours ago

How to run DeepSeek AI locally to protect your privacy – 2 easy ways

DeepSeek is a Chinese AI startup that has developed powerful large language models (LLMs) at a lower cost than US AI companies. DeepSeek is open-source, and its models perform as well as or better than other models. The company has released different models for different purposes (such as programming, general-purpose, and vision). DeepSeek seems to be having a conversation with itself, in the process of relaying information to the user. You can install DeepSeek locally and use it at your whim. You can use it on Linux, MacOS, or Windows, and it won’t cost you a penny.

ZDNet.AI

17 hours ago

How to use Gemini’s Deep Research to browse the web faster and better

Google Gemini’s DeepResearch is a virtual agent that performs online research for you. Accessible on the web, Android, and iOS, DeepResearch generates a research plan, searches the web for relevant information, and generates a detailed report for your review. With DeepResearch, Gemini is also capable of multitasking. To use DeepResearch, you’ll need the $20-per-month Gemini Advanced subscription, which is part of the Google One AI Premium plan. At the prompt, type a description of the topic you want to research. The AI creates a plan with a specific series of steps that it will follow.

The Verge

17 hours ago

Fiverr wants freelancers to create AI models

Fiverr will now let creators train AI models on their own work. Clients can then purchase access to custom AI models and generate content “with the speed of AI with the expertise of a pro,” the company says. Freelancers can edit and set the price for their custom AI models, which are available across services like voiceovers, songwriting, graphic design, illustration, copywriting, and digital marketing. Fiverr warns that exclusive outputs are not guaranteed, and similar results may be generated for different users. Clients will own the rights to the final AI-generated content they purchase.

Gizmodo

17 hours ago

Humane Bricks Its AI Pin as It Gets Acquired by HP for $116 Million

HP has acquired Humane’s CosmOS operating system and IP, including the tech for the AI Pin, for $116 million, signaling the end for Humane. The devices will be bricked by Feb. 28. HP intends to use the acquired technology to develop its own suite of AI-centric devices and integrate CosmOS into its PCs, smart printers, and connected conference rooms. Tuan Tran, HP’s former president of the company’s printing arm, noted CosmOS’ ability to use both on-device and cloud-based AI models as a key factor in the acquisition. Humane had been looking to sell shortly after releasing the AI Pin, which received largely negative reviews.

ZDNet.AI

17 hours ago

Will AI destroy human creativity? No – and here’s why

The question of whether AI could impact creativity and innovation is complex. Artificial intelligence has been publicized before. The New York Times recently ran a piece titled “Electronic Brain Teaches Itself,” spotlighting the Perceptron, an ambitious AI project funded by the US Navy. This early attempt at artificial intelligence was to be the first non-living machine capable of “perceiving, recognizing, and identifying its surroundings without human training or control.”

KDnuggets

18 hours ago

Math, Machine Learning & Coding Needed For LLMs

The requested URL was not found. If you are a human, go to the KDnuggets homepage and search for this blog from there. Or check KDnuggets latest opinions and tutorials.


With SummHub, you will enjoy the efficient and convenient news read experience brought by technology. Try SummHub immediately and start your creative journey! 🚀👇👇👇

<< IFlux: Revolutionizing Your News Experience with AI-Powered Intelligence
>> [2025-02-20] DeepSeek

Recent Posts

  • AI News Alerts: Best Newsletters and Social Media Sources
  • AI News: Transformative Impacts on Business
  • AI News Insights: Decoding the Jargon in Reports
  • OpenAI Acquires Jony Ive’s Startup to Build AI Devices
  • Google I/O 2025: Gemini AI, XR, and the Future of Search

🔥 Product

AI Intelligence

Powered by advanced AI technology for smart content curation, instant translations, and intelligent summaries that help you grasp complex topics effortlessly.

Global Perspective

Access diverse viewpoints from trusted global sources, with real-time updates and breaking news from around the world, all in one place.

Personalized Experience

Create unlimited custom topics, choose your preferred reading style, and get smart notifications tailored to your interests and schedule.

©2025 SummHub | Design: Newspaperly WordPress Theme