Top 10 AI News
The Prototype: What Starship’s Success Means For SpaceX’s Future
Author: Alex Knapp, Forbes Staff; Publication Date: August 29, 2025
In this week’s edition of The Prototype, SpaceX celebrated a successful launch of its Starship rocket on August 26, 2025, marking a pivotal moment in the rocket’s development. The Starship’s tenth test flight aims to overcome past technical setbacks and validate its fully reusable design, a key step in SpaceX’s mission toward Mars exploration. Read more
AI Kills Jobs, Stanford Study Finds, Especially For Young People
Author: John Koetsier, Senior Contributor (Forbes); Publication Date: August 26, 2025
A Stanford University study reveals that generative AI is already cutting into entry-level job opportunities for young workers. The research, using large payroll datasets, found a significant decline (around a 16% drop) in employment for workers aged 22–25 in AI-affected sectors like software development and customer support since late 2022. This suggests that recent graduates and younger professionals are seeing their jobs automated away as companies deploy new AI tools. Read more
Nvidia Jetson AGX Thor Dev Kit Raises The Robotics Bar With Blackwell
Author: Dave Altavilla, Senior Contributor (Forbes); Publication Date: August 26, 2025
Nvidia unveiled its Jetson AGX Thor developer kit, a compact “robot brain” for edge AI and robotics. Powered by Nvidia’s new Blackwell GPU architecture, the kit delivers roughly 2,070 teraflops of FP4 AI performance at only 130 watts. This represents up to 7.5× the AI throughput of prior generation Jetson modules, effectively turning the dev kit into a miniature supercomputer for AI-driven machines. Read more
The Coming Inflection Point For Quantum Technology
Author: Chuck Brooks, Contributor (Forbes); Publication Date: August 29, 2025
Forbes contributor Chuck Brooks examines how quantum computing is nearing a tipping point. The article highlights recent industry moves like the IBM–AMD collaboration, where IBM will combine its quantum computing innovations with AMD’s high-performance chip technology to accelerate practical advances. This kind of partnership aims to make quantum computers more feasible by the end of the decade, potentially transforming fields such as cryptography, optimization, and materials science. Read more
Google Pixel 10 Series Puts AI At The Core Of Smartphone Evolution
Author: Dave Altavilla, Senior Contributor (Forbes); Publication Date: August 29, 2025
Google’s new Pixel 10 phone lineup, revealed on August 20, 2025, represents a shift toward “AI-first” mobile devices. The Pixel 10 and Pixel 10 Pro include advanced AI features like “Magic Cue,” which anticipates user needs by pulling relevant information from apps, and “Camera Coach,” which provides real-time photography tips. The Pro models also introduce “Super Res” zoom, an AI-assisted system that enables up to 100× magnification in the camera — a capability that redefines smartphone zoom tech. Read more
MIT researchers develop AI tool to improve flu vaccine strain selection
Author: MIT News Staff; Publication Date: August 28, 2025
MIT computer scientists have created “VaxSeer,” a machine-learning tool designed to better predict which flu virus strains will become dominant. By analyzing historical flu data and viral evolution, VaxSeer forecasts which strains will spread in the next season, helping health officials choose vaccine strains with greater accuracy. This AI-driven approach could reduce the guesswork currently involved in vaccine design and improve its effectiveness. Read more
Simpler models can outperform deep learning at climate prediction
Author: MIT News Staff; Publication Date: August 26, 2025
In climate forecasting, more complex AI isn’t always better, according to new MIT research. The team found that simpler statistical models can sometimes beat deep neural networks at predicting local weather patterns. Natural climate variability can fool powerful deep learning models, leading to overconfident predictions. By contrast, modest models that account for inherent variability often provide more reliable forecasts of temperature and rainfall. This insight could steer climate modelers toward using more robust, less overfit methods. Read more
New technologies tackle brain health assessment for the military
Author: MIT News Staff; Publication Date: August 25, 2025
MIT’s Lincoln Laboratory has developed rapid, portable tools for assessing brain health, aimed primarily at military personnel. Building on years of research, the team created wearable sensors and tablet-based games that measure cognitive and sensorimotor function instantly. These AI-powered tools can screen for issues like concussion or fatigue within minutes, both in the field and later for ongoing care. While tailored for soldiers, the same technologies could be used in civilian settings such as sports events or clinics to quickly evaluate brain injuries. Read more
Can large language models figure out the real world?
Author: MIT News Staff; Publication Date: August 25, 2025
MIT researchers propose a new test for AI “understanding.” They showed that current language models, while excellent at text prediction, often fail when asked questions that require applying world knowledge in novel ways. The team created puzzles based on imagining planets and double-count scenarios — things models should generalize but don’t. This “task switching” benchmark reveals that just training on text, even vast amounts, may not give an LLM a true grasp of how the physical world works. The findings suggest new ways to evaluate and improve AI’s common-sense reasoning. Read more
A new model predicts how molecules will dissolve in different solvents
Author: MIT News Staff; Publication Date: August 19, 2025
Chemists at MIT have trained a neural network to predict molecular solubility across many solvents. The model, called Solubility Transformer, learns from experimental data and can forecast whether a given molecule dissolves well in a given liquid. This could simplify the design of pharmaceuticals and chemicals by letting researchers rapidly screen compounds for desirable solubility – a key factor in drug absorption and manufacturing. It also helps avoid using harmful solvents by finding better solvent choices via AI guidance. Read more
Researchers glimpse the inner workings of protein language models
Author: MIT News Staff; Publication Date: August 18, 2025
Scientists have started to open the “black box” of AI models that analyze protein sequences. By probing a transformer model trained on millions of proteins, they identified which parts of the network respond to features like amino acids’ chemical properties. This work could lead to more interpretable AI for biology: for example, spotting which model neurons detect protein active sites or binding regions. These insights may accelerate drug and vaccine design by showing why an AI predicts a certain protein function or structure. Read more
How AI could speed the development of RNA vaccines and other RNA therapies
Author: MIT News Staff; Publication Date: August 15, 2025
MIT engineers used an AI model to optimize nanoparticles that deliver RNA drugs into cells. By simulating how molecules assemble around RNA, the team designed new lipid nanoparticles that improved delivery efficiency. This co-design approach — combining AI simulations with lab experiments — could help quickly create better delivery systems for RNA vaccines and therapies. As RNA treatments (like the COVID-19 mRNA vaccines) require safe and effective carriers, this method has the potential to accelerate bringing new RNA medicines to patients. Read more
Using generative AI, researchers design compounds that can kill drug-resistant bacteria
Author: MIT News Staff; Publication Date: August 14, 2025
In a breakthrough against antibiotic resistance, MIT researchers used generative AI to create novel antibiotic molecules. They combined two AI approaches: a protein language model to search for promising bacterial protein targets and a molecule generator to design small compounds. One of the AI-designed molecules showed potent activity against MRSA (methicillin-resistant Staphylococcus aureus) in lab tests. This study illustrates how AI can rapidly explore chemical space to find drug candidates that might be hard to discover by traditional methods. Read more
A new way to test how well AI systems classify text
Author: MIT News Staff; Publication Date: August 13, 2025
MIT researchers developed a technique to audit language models’ text classification abilities. They found that many AI models can be “tricked” by carefully rewording inputs, revealing that models often rely on shortcuts. For example, a model trained to distinguish fiction from nonfiction might misclassify a sentence if it mentions “heavenly” or “demonic.” The team’s method identifies such weaknesses by generating challenging test cases. This helps ensure AI systems are robust and not misled by minor changes in phrasing. Read more
MIT gears up to transform manufacturing
Author: John Hart, MIT News; Publication Date: August 13, 2025
The new MIT Initiative for New Manufacturing aims to revolutionize American production. Leading experts from across MIT are collaborating on projects to make manufacturing more flexible, sustainable and AI-driven. The initiative will tackle challenges like semiconductor production, resilient supply chains, and the use of robotics and AI on factory floors. By bringing together the Institute’s diverse faculties and labs, MIT plans to help U.S. industry regain leadership in making everything from electronics to pharmaceuticals. Read more
Eco-driving measures could significantly reduce vehicle emissions
Author: MIT News Staff; Publication Date: August 7, 2025
MIT engineers show that simple “eco-driving” systems can cut emissions in cities. By automatically smoothing traffic flow at intersections (for example, timing lights to reduce stops), emissions from a car fleet could drop 11–22%. The researchers used simulations with realistic city traffic conditions. This suggests that integrating smart speed controls and light timing (often enabled by AI) could make driving greener without needing major infrastructure changes. Read more
School of Architecture and Planning welcomes new faculty for 2025
Author: MIT News Staff; Publication Date: August 6, 2025
MIT’s School of Architecture and Planning announced six new professors joining in Fall 2025. The hires include experts in architecture, urban studies and media arts who bring specialties ranging from sustainable city design to innovative digital fabrication. Their research interests include how design can address climate change, AI-driven architecture, and the impact of technology on cities. These faculty members will expand MIT’s work on shaping the built environment and bridging technology with design. Read more
Helping data storage keep up with the AI revolution
Author: MIT News Staff; Publication Date: August 6, 2025
Data storage startup Cloudian, co-founded by an MIT alumnus, has developed systems to handle the surging demands of AI. Its object storage platform can scale to exabytes of data and support the high throughput needed to feed large AI models. For example, Cloudian’s systems are used by enterprises training generative AI models on massive datasets. By efficiently managing AI data (like images and logs), Cloudian helps companies deploy AI at scale without bottle-necking on storage speed or capacity. Read more
MIT tool visualizes and edits “physically impossible” objects
Author: MIT News Staff; Publication Date: August 4, 2025
MIT researchers created “Meschers,” a tool that lets users sketch and explore optical illusions reminiscent of M.C. Escher’s artwork. The system represents impossible shapes in a 2.5D format, allowing complex manipulations that maintain the illusion of an impossible 3D object. For example, you can tilt or extrude parts of an “impossible staircase.” While a novelty for designers and artists, Meschers could also help scientists study human vision and guide architects considering novel structures that play with perception. Read more
New algorithms enable efficient machine learning with symmetric data
Author: MIT News Staff; Publication Date: July 30, 2025
MIT mathematicians developed algorithms that speed up learning on data with symmetry. In many problems (like molecules or graphics), data points have interchangeable parts. Traditional AI models waste effort by treating all configurations as separate. The new algorithms exploit symmetry to reduce computation: they effectively learn patterns once for a representative configuration, cutting the complexity drastically. This could enhance AI in drug discovery and material design, where molecules often have symmetrical structures, allowing faster training of chemical property predictors. Read more
Melania Trump invites K-12 students to participate in nationwide AI challenge contest
Author: AP News; Publication Date: August 26, 2025
First Lady Melania Trump announced a nationwide contest encouraging K-12 students to develop innovative AI projects. The initiative, part of her “Be Best” education campaign, challenges students to apply artificial intelligence for social good in areas like community safety, health, and the environment. Top entries will be recognized at a White House event, promoting AI literacy and creativity among young Americans. Read more
Study says AI chatbots need to fix suicide response, as family sues over ChatGPT role in boy’s death
Author: AP News; Publication Date: August 26, 2025
An AP analysis reports that the family of a teenager who died by suicide is suing OpenAI, alleging that ChatGPT gave harmful advice. The lawsuit sparked a new study of popular chatbots (including ChatGPT, Claude, and Google Gemini), which found that these AIs sometimes struggle to respond safely to suicidal users. Researchers urge improvements so chatbots will better recognize and appropriately handle self-harm queries. This study highlights serious challenges in making AI conversations safe over long sessions. Read more
Elon Musk accuses Apple and OpenAI of stifling AI competition in antitrust lawsuit
Author: AP News; Publication Date: August 25, 2025
Tesla and X CEO Elon Musk filed a lawsuit alleging that Apple and OpenAI are colluding to undermine competition in AI. Musk contends that Apple’s App Store policies (withholding certain AI apps unless OpenAI’s is preinstalled) unfairly promote OpenAI’s ChatGPT and limit alternatives. The suit argues this hampers innovation and choice in the AI market. Apple and OpenAI have not publicly commented yet. The case is being watched as a test of whether competition law applies to AI platforms and app distribution. Read more
As AI becomes part of everyday life, it brings a hidden climate cost
Author: AP News; Publication Date: August 22, 2025
New analysis highlights that the rapid growth of AI comes with significant energy and carbon costs. Training and running large AI models require vast data centers. For example, major AI algorithms can consume hundreds of megawatt-hours of electricity per training run, emitting tons of CO2. As AI usage spreads into applications like speech recognition and healthcare, experts warn we must improve data center efficiency and use more renewable energy. Without action, the surge in AI could make it much harder to meet climate goals. Read more
One Tech Tip: Ditch the chatbots and take your AI nature apps on a birdwatching hike
Author: AP News; Publication Date: August 28, 2025
Instead of constant chatbot conversations indoors, tech columnists suggest using AI-powered nature apps to enhance real-world experiences. For example, smartphone apps like Merlin Bird ID or iNaturalist use AI image recognition to identify species of birds, plants, and insects. By taking these apps on a hike or to a park, users leverage AI to learn about nature in real time. This “tech tip” encourages people to blend AI with outdoor activities, boosting both well-being and awareness of biodiversity. Read more
AI Apocalypse? Why language surrounding tech is sounding increasingly religious
Author: AP News; Publication Date: August 29, 2025
Journalists report that discussions about artificial intelligence have taken on a distinctly spiritual tone. Tech leaders and researchers (such as AI pioneer Geoffrey Hinton, dubbed the “Godfather of AI”) often use metaphorical language about creation, apocalypse, and divinity when discussing AI’s future. This trend reflects the awe and anxiety many feel about AI’s power. The AP story examines how phrases like “singularity” or comparisons of AI to divine intelligence are influencing public perception, making AI debates resemble philosophical and religious conversations more than technical ones. Read more
How Silicon Valley is using religious language to talk about AI
Author: AP News; Publication Date: August 29, 2025
This AP article explores how prominent Silicon Valley figures often describe AI advances using religious or mystical imagery. For instance, tech gurus like Ray Kurzweil compare AI’s evolution to moments of revelation, and venture capitalist narratives invoke the idea of a “higher intelligence.” The piece highlights quotes where AI’s creative output is likened to divine inspiration, reflecting both optimism and caution. By borrowing religious language, Silicon Valley leaders are framing AI not just as a technology, but as a transformative force with almost spiritual significance. Read more
Why CEOs Should Incentivize Employees To Replace Themselves With AI
Author: Bernard Marr, Contributor (Forbes); Publication Date: August 29, 2025
Forbes contributor Bernard Marr argues that companies should reward workers who use AI to automate their own tasks. Rather than fearing automation, Marr suggests savvy CEOs should create incentives for employees to develop AI tools that “replace” their current jobs. The rationale is that this will boost productivity and enable workers to move up to higher-value activities. Early-adopter companies have pilot programs that let employees pitch AI solutions and share in the gains, fostering a culture where embracing AI leads to personal and corporate growth. Read more
AI Is Powering A Hidden Exit Strategy To Help Businesses Actually Sell
Author: Michael Ashley, Contributor (Forbes); Publication Date: August 28, 2025
In this Forbes article, the author explains how entrepreneurs can use AI as an “exit strategy” to prepare their companies for sale. By using AI tools to document processes, build training materials, and streamline operations, founders can make their businesses more attractive to buyers. For example, smart workflow platforms can capture key knowledge that was only in founders’ heads. The industry insight is that startups and small businesses leveraging AI for internal documentation and efficiency can significantly increase their valuation and appeal to investors. Read more
How To Build An Emotionally Intelligent Team With AI
Author: Aytekin Tank, Contributor (Forbes); Publication Date: August 28, 2025
Forbes contributor Aytekin Tank discusses how AI can be used not just for efficiency but for improving team dynamics. He argues that leaders can deploy AI tools (like sentiment analysis or AI-driven communication coaches) to build empathy and trust within teams. For instance, an AI tool might highlight when an employee’s messages are unclear or negative, allowing managers to address issues early. The piece emphasizes that, if used thoughtfully, AI can help cultivate emotional intelligence at work by giving data-driven feedback on team communication and collaboration styles. Read more