 |
|
|
|||||
The global race for better batteries has never been more intense. Electric vehicles, drones, and next-generation aircraft all depend on high-performance energy storageyet the traditional approach to battery R&D is struggling to keep pace with demand. Innovation and investment alone wont solve the problem, unless we compress the timeline. Speed is now the defining barrier between potential and impact. Even as AI speeds up materials discovery, battery lifetime still dictates success: each charge-discharge cycle lasts about six hours, so proving out 500 cycles can take up to eight months, turning lifetime testing into the key bottleneck for promising chemistries. Thats changing. Physics-informed AI is redefining battery development. National labs like NREL have shown how neural networks can diagnose battery health 1,000 times faster than conventional models, bringing real-time insight into degradation and performance. The Real Cost of Traditional Testing Battery development has always been a waiting game. Consider the mathematics: testing at a standard C/3 rate allows for just two complete cycles per day. Multiply that across different chemistries, protocols, and form factors, and you’re looking at years of validation before a single product reaches market. This isn’t just inefficientit’s becoming unsustainable. While battery researchers methodically work through their testing cycles, the market landscape shifts beneath them. New competitors emerge, customer requirements evolve, and breakthrough technologies risk becoming obsolete before they’re even validated. The industry needs a fundamental shift in how it approaches innovation. Why Conventional AI Isn’t the Answer Many companies have turned to traditional machine learning, hoping to accelerate their development cycles. But conventional AI tools face critical limitations in battery applications: Data scarcity: Unlike consumer tech, battery research generates relatively small, messy datasets that resist standard ML approaches. Black box problem: Correlation-based models might identify patterns, but they can’t explain why those patterns exist, which is a nonstarter in a field governed by strict electrochemical and thermodynamic principles. Regulatory challenges: Engineers and regulators need to understand not just what an AI predicts, but why it makes those predictions. Enter Physics-Informed AI Physics-informed AI represents a fundamental departure from conventional approaches. Instead of learning patterns from data alone, these models embed physical laws directly into their architecture. The result is AI that doesn’t just recognize correlationsit correlates with the underlying physics. This approach transforms how we think about battery development. Rather than waiting months for empirical validation, physics-informed models can simulate real battery behavior with remarkable accuracy. They account for aging, degradation, thermal stress, and mechanical factorsall grounded in established scientific principles. At Factorial, we’ve achieved something that seemed impossible just years ago: predicting cycle life outcomes after just 12 weeks of early testing, compared to the 36 months typically required. Software-Driven Breakthroughs The impact extends beyond faster testing. Using our newly launched Gammatron platforma proprietary physics-informed AI systemwe recently optimized a fast-charging protocol without altering any physical components. The result: a twofold improvement in cycle life, achieved entirely through software. Gammatron, developed to simulate and predict battery behavior with high accuracy, has transformed our approach to development with Stellantis. By forecasting long-term performance from just two weeks of early data, the platform helped accelerate validation timelines and informed protocol adjustments that significantly extended battery lifespan, without changing chemistry or hardware. Were not the only ones seeing this level of transformation. At The Battery Show Europe, Monolith CEO Richard Ahlfeld shared that his company, working with Cellforce Group, is using AI to reduce battery materials testing requirements by up to 70%, while maintaining or even improving discovery rates. These aren’t theoretical savings. Monolith reports 2040% reductions in testing across active partner projects today, accelerating products to market by months. This represents a new paradigm in battery developmentone where software innovations can drive hardware-level gains. As our models continuously learn from new lab data, they evolve in real time, accelerating innovation throughout the entire product lifecycle. his combination of AI and lab data enables a feedback loop that isnt seen in traditional AI models. Transforming Industry Standards Physics-informed AI enables capabilities that were previously impossible: Precision matching: Align specific chemistries with target applications based on predictive performance modeling rather than trial and error. Virtual prototyping: Simulate performance outcomes before investing in physical prototypes, dramatically reducing development costs and timelines. Intelligent optimization: Fine-tune charging protocols for optimal speed and safety without extensive physical testing. Predictive monitoring: Identify potential failure modes early in the development cycle, reducing both risk and cost. Perhaps most importantly, these tools support continuous learning throughout the product lifecycle. As new materials, processes, and data become available, the models evolve, enabling rapid adaptation across diverse battery platforms and applications. The Simulation-First Future We’re witnessing the emergence of a new development paradigmdigital cell design. Tomorrow’s battery breakthroughs will begin not in physical labs, but in sophisticated simulations that combine domain expertise, experimental validation, and intelligent AI modeling. This shift from hardware-first to data-first innovation will separate industry leaders from followers. Companies that can seamlessly integrate these capabilities will unlock longer range, faster charging, and greater resilience, solving what are fundamentally systems challenges rather than just materials challenges. The tools exist today. The question isn’t whether this transformation will happen, but how quickly companies will adapt to leverage these capabilities.
Category:
E-Commerce
Matt Losak got interested in evictions when he was nearly subject to one. He was working for a union, which required frequent travel. One month he accidentally left his rent check on his refrigerator and left. He called the property manager, who told him it wouldnt be a problem as long as he gave her the check when he returned and paid a $25 late fee, all of which he did. Six weeks later, he received an eviction notice for failure to pay rent. Losak fought the eviction off by proving that the property owners had signed the eviction papers after having already cashed his check. I caught them by the toe, he said. But the experience focused him on the fact that, in most places, landlords can move to evict tenants for any reason. Its one part of why evictions have been climbing steadily since the pandemic, even outpacing pre-pandemic levels in some places. Losak helped found Renters United Maryland in 2017, and enshrining good cause protection from eviction was his number-one issue. Good cause laws require landlords to have a specific reason for evicting someone or terminating their lease, although typically tenants can still be removed for things like violating lease agreements or failing to pay rent. The lack of good cause protections, is a major loophole, he says, allowing a landlord to, based on whim and caprice, rip somebody out of their home. When Losak first started advocating for good cause, he says state lawmakers literally laughed in my face. They didnt buy that it was a big problem, not to mention that landlords and their lobbyists, who oppose such laws, hold huge sway. But his organization has worked to build support over the years. It pushed the idea that a lack of stable housing leads to concrete costs in mental and physical health declines, lower academic achievement for children, and higher crime rates. Theres all types of costs, financial and social, that people are beginning to recognize are directly related to unstable or poor quality homes, he says. The Pandemic Effect Then came the pandemic, which clearly was an accelerant and focused state lawmakers attention on housing policy, he says. In 2020, experts and analysts put the number of potential evictions at anywhere between 17 million and 28 million. That fear of widespread evictions led courts and others to look at the eviction process differently and try things out, says Sarah Gallagher, vice president of state and local innovation at the National Low Income Housing Coalition. Eventually eviction moratoria and rental assistance would help to keep the number of evictions low. Now, however, those protections are gone and rents are rising rapidly. The housing crisis, Gallagher says, is even worse than it was before the pandemic. Rents rose 29% between 2019 and 2023 and homelessness is at a record high. So lawmakers have been looking for other tools to continue helping people stay housed. Maryland was among a number of other states that considered good cause eviction protections during this years legislative session. After a bottleneck against advancing legislation was ousted, the bill came close to passage. Losak was able to marshal support from a wide coalition that included the ACLU, NAACP, unions, and even the governor. Legislation passed the state house, only to die in the senate under pressure from landlord lobbies. Losak noted that next year is an election year for the Maryland legislature. There is a bubble of anger and frustration that is growing among a huge number of Marylanders that might very well have a political impact, he says. Were optimistic. Fights are now brewing in Chicago and Rhode Island. These cities and states are looking to join a growing movement: 11 states and 27 localities have now passed good cause laws. The majority of those laws were passed in the years since 2020. It is gaining momentum, Gallagher says. Reducing eviction rates Research has found that these laws reduce eviction filing rates, keeping people both housed and firmly rooted in their communities. Academics have also found that good cause protections in California, Oregon, and New Hampshire didnt decrease housing production after they passed. New York State passed good cause eviction protection last year that went into effect immediately in New York City and allowed other places to opt in. It used to be that, as long as a New York landlord filled out all the forms correctly, it could evict someone. But now landlords cant just say, Well, I want this person out, says Judith Goldiner, attorney in charge of the civil law reform unit at The Legal Aid Society. Under the new law, landlords have to offer their tenants a lease renewal unless they can prove they have a good reason not to, such as illegal behavior or the need to demolish a building. It also capped rent increases. But the law came with a long list of exceptions, including any building built after 2009, luxury buildings, and buildings owned by someone with a portfolio of 10 or fewer units. The carveouts mean that, while its a useful tool, Goldiner says, it is often unduly complicated and hard to explain. It can be hard even to figure out what entity owns buildings, let alone how many others they own. The law is still new, so there havent been decisions hammering out exactly what counts as a good cause for evicting someone. Still, Goldiner knows of a number of tenants who were able to use it to negotiate leases instead of ending up in eviction proceedings. Definitely its been really helpful for a lot of people to actually avoid litigation, she says. Expanding tenant protections Good cause laws wont stop evictions all on their own. There isnt one protection thats going to do it all, Gallagher says. You have to look at putting in tenant protections at all stages of the housing and eviction process. That includs starting with regulating how tenants are screened by landlords so they arent discriminated against all the way to sealing eviction records after tenants are forced out so they arent penalized when trying to get housing in the future. Still, good cause laws are particularly important because they help insulate tenants from retaliation for asserting any rights they do have. They enable tenants to speak out against unsafe or unhealthy housing conditions or other abuses. Under good cause laws, landlords cant kick people out just because they, say, asked for repairs. And policies like right to counsel, which guarantees legal representation in eviction proceedings, are only useful if there are actual rights for lawyers to defend. If tenants dont have the ability to advocate for their rights, then what good is the protection? Gallagher says.
Category:
E-Commerce
Pick up an August 2025 issue of Vogue, and youll come across an advertisement for the brand Guess featuring a stunning model. Yet tucked away in small print is a startling admission: She isnt real. She was generated entirely by AI. For decades, fashion images have been retouched. But this isnt airbrushing a real person; its a person created from scratch, a digital composite of data points, engineered to appear as a beautiful woman. The backlash to the Guess ad was swift. Veteran model Felicity Hayward called the move lazy and cheap, warning that it undermines years of work to promote diversity. After all, why hire models of different sizes, ages, and ethnicities when a machine can generate a narrow, market-tested ideal of beauty on demand? I study human-AI collaboration, and my work focuses on how artificial intelligence influences decision-making, trust, and human agency, all of which came into play during the Vogue controversy. This new reality is not a cause for doom. However, now that its becoming much harderif not impossibleto tell whether something is created by a human or a machine, its worth asking whats gained and whats lost from this technology. Most importantly, what does it say about what we truly value in art? The forensic viewer and listener In 1950, computer scientist Alan Turing wondered whether a machine could exhibit intelligent behavior indistinguishable from that of a human. He proposed his famous imitation game. In it, a human judges whether theyre conversing with a person or a computer. If the human cant tell the difference, the computer passes the test. For decades, this remained a theoretical benchmark. But with the recent explosion of powerful chatbots, the original Turing Test for conversation has arguably been passed. This breakthrough raises a new question: If AI can master conversation, can it master art? The evidence suggests it has already passed what might be called an aesthetic Turing Test. AI can generate music, images, and movies so convincingly that people struggle to distinguish them from human creations. In music, platforms like Suno and Udio can produce original songs, complete with vocals and lyrics, in any imaginable genre in seconds. Some are so good theyve gone viral. Meanwhile, photorealistic images are equally deceptive. In 2023, millions believed that the fabricated photo of Pope Francis in a puffer jacket was real, a stunning example of AIs power to create convincing fiction. Why our brains are being fooled So why are we falling for it? First, AI has become an expert forger of human patterns. These models are trained on gigantic libraries of human-made art. They have analyzed more paintings, songs, and photographs than any person ever could. These models may not have a soul, but they have learned the mathematical recipe for what we find beautiful or catchy. Second, AI has bridged the uncanny valley. This is the term for the creepy feeling we get when something looks almost human but not quitelike a humanoid robot or a doll with vacant eyes. That subtle sense of wrongness has been our built-in detector for fakes. But the latest AI is so sophisticated that it has climbed out of the valley. It no longer makes the small mistakes that trigger our alarm bells. Finally, AI does not just copy reality; it creates a perfected version of it. The French philosopher Jean Baudrillard called this a simulacruma copy with no original. The AI model in Vogue is the perfect example. She is not a picture of a real woman. She is a hyperreal ideal that no living person can compete with. Viewers dont flag her as fake because she is, in a sense, more perfect than real. The future of art in a synthetic world When art is this easy to generateand its origin this hard to verifysomething precious risks being lost. The German thinker Walter Benjamin once wrote about the aura of an original artworkthe sense of history and human touch that makes it special. A painting has an aura because you can see the brushstrokes; an old photograph has an aura because it captured a real moment in time. AI-generated art has no such aura. It is infinitely reproducible, has no history, and lacks a human story. This is why, even when it is technically perfect, it can feel hollow. When you become suspicious of a works origins, the act of listening to a song or viewing a photograph is no longer simply about feeling the rhythm or wondering what may have existed outside the frame. It also requires running a mental checklist, searching for the statistical ghost in the machine. And that moment of analytical doubt pulls viewers and listeners out of the works emotional world. To me, the aesthetic Turing Test is not just about whether a machine can fool us; its a challenge that asks us to decide what we really want from art. If a machine creates a song that brings a person to tears, does it matter that the machine felt nothing? Where does the meaning of art truly residein the mind of the creator or in the heart of the observer? We have built a mirror that reflects our own creativity back at us, and now we must decide: Do we prefer perfection without humanity, or imperfection with meaning? Do we choose the flawless, disposable reflection, or the messy, fun house mirror of the human mind? Tamilla Triantoro is an associate professor of business analytics and information systems at Quinnipiac University. This article is republished from The Conversation under a Creative Commons license. Read the original article.
Category:
E-Commerce
All news |
||||||||||||||||||
|
||||||||||||||||||