The Fast Company Impact Council is an invitation-only membership community of leaders, experts, executives, and entrepreneurs who share their insights with our audience. Members pay annual dues for access to peer learning, thought leadership opportunities, events and more. There’s no question that artificial intelligence has taken the world by storm. However, as the initial excitement over the technology fades, we find ourselves in a new phase of thoughtful exploration. There are many innovative AI startups that have captured the world’s attention; however, many organizations still struggle to develop a clear roadmap to take full advantage of this transformative technology.   So, what’s the hold up? And how can business leaders avoid fleeting trends, effectively align their teams, and successfully integrate AI to achieve measurable impact and ROI for their business?  Embrace the journey  AI is already transforming industries, boosting efficiency and automating tasks ranging from data entry and language translation to document processing. And the benefits are clear—recent Accenture research found that the vast majority of organizations are seeing stronger-than-expected returns from their generative AI investments.  Still, it’s important to keep a balanced perspective. While many AI solutions promise substantial benefits, the real challenge is identifying those that add tangible value. With new technologies emerging almost weekly, some leaders may also hesitate to invest because they are unsure if a better option is just around the corner.   AI’s true power comes from practical, enterprise-ready applications. For business leaders wondering where to start, the key is identifying the right challenges to tackle and knowing when and how to implement solutions effectively. Here are seven actionable tips to help you navigate this exciting landscape and build an AI decision-making framework tailored to your organization’s needs.  1. Identify the use case   First, pinpoint your specific needs and business objectives. Start within your organization, identifying pain points AI can address. Think about what AI does well, like spotting patterns, crunching numbers, and making predictions. Could it help with document translation, content creation, or customer insights? With so many potential applications, determining where to start might seem daunting. A focused, purposeful approach ensures you’re investing in AI solutions that deliver real results.  2. Consider specialized models  Over the last two years, we’ve seen much of the excitement around general purpose AI models outpace their value. As you evaluate AI tools for your organization, consider specialized AI models offering tailored solutions for specific industry needs.  General AI models can do many things pretty well, but for higher stakes and more specific demands, specialized models often address complex, industry-specific challenges more effectively. For example, healthcare AI models can help doctors identify diseases more accurately, while banks use credit-scoring AI to determine who’s likely to pay back loans. Language AI tools like DeepL are also specialized to businesses communicating across languages and markets.  Specialized AI offerings are trained on domain-specific data optimized for particular tasks or industries, delivering enhanced quality and accuracy with lower risk of errors. They’re also often designed with built-in compliance features aligned with industry regulations. This makes them more cost-effective, with clearer ROI.   3. Are humans the answer?  When you’re holding a hammer, everything looks like a nail, right? As the founder of an AI company, you might be surprised to hear me say this, but just because AI is the big thing right now doesn’t mean it’s the singular solution for every problem or opportunity. So before diving into the deep end, consider if a human solution might actually be more effective than AI. Weighing what people, supported by AI, do best versus what AI can offer on its own, will help ensure you take the right approach for your organization’s needs.  4. Start with pilot projects   If you’re about to deploy an AI solution for the first time, begin with pilot projects to test your AI integrations in smaller, controlled environments. Starting small with a more limited investment reduces overall risk and can allow you to gather real-world data, monitor performance, and assess alignment with business goals before scaling. Pilot projects can also help build confidence within your teams and among leadership, making way for more successful full-scale AI deployments.  5. Invest in tech (and training)  To truly harness AI’s potential, focus on bringing in new talent and continuously training existing employees. Depending on the implementation’s complexity, you might need new positions like data scientists, machine learning engineers or specialists. Upskilling your existing workforce can be equally essential to ensure employees can adapt and thrive alongside technological advancements.   6. Have a solid data strategy in place  AI requires large volumes of data to perform its best, so it’s essential to have a solid data strategy infrastructure in place. Your plan should address how your organization will collect, securely store and access data; ensure compliance with evolving data privacy regulations, copyright standards and ethical guidelines; and assign responsibility for ongoing data governance and management. Answering these questions up front will save your company stress and problems later.  7. Refresh your ROI framework and adjust it regularly  Most business leaders can recall digital initiatives that didn’t meet expectations, which can lead to concerns that their AI investments might follow a similar path. To enhance your ROI, outline your initiative’s measurable goals, such as efficiency, cost savings, or an enhanced customer experience. Establish baseline metrics to understand current performance; then track improvements directly linked to AI.   It’s important to be adaptable, regularly revisiting goals and metrics to reflect evolving business priorities, market conditions, and technological changes. Unlike standard digital projects, AI initiatives can uncover new opportunities or shift mid-course. Also consider AI’s long-term strategic advantages, which may take time to come to fruition.  From hyperbole to high performance   To make AI work, organizations should shift their focus from what’s trending to enterprise-ready solutions that deliver lasting and specific value. Define your use cases up front, adopt an agile ROI framework, a robust data strategy, and commit to continuous improvement. This will unlock AI’s transformative potential and build a foundation for long-term competitive advantage.  Jarek Kutylowski is CEO and founder of DeepL. 

The Fast Company Impact Council is an invitation-only membership community of leaders, experts, executives, and entrepreneurs who share their insights with our audience. Members pay annual dues for access to peer learning, thought leadership opportunities, events and more. Across industries, a new era of climate innovation is accelerating. The momentum is visible in the data: Global clean energy investment surpassed $2 trillion for the first time in 2024, double the amount invested in fossil fuels.  While solar panels, wind turbines, and grid-connected batteries often grab the headlines, the low carbon economy is growing in far more corners than many realize.  Since founding Supercool last summer to cover proven and scaling climate solutions, I’ve seen needle-moving innovation accelerating across farms, factories, and finance departments.  One sector in particular shows remarkable progress—the built environment, which accounts for 34% of global carbon emissions.  From hard tech and material breakthroughs to AI-powered intelligence to novel business models, here are three approaches to decarbonizing buildings happening now.  1. Hardtech innovation: Build with carbon-negative materials   The engineered materials we use to build our suburbs and cities—primarily timber, concrete, and steel—create a lot of carbon emissions in their manufacture. Concrete and steel account for nearly 18% of global greenhouse gas emissions. Wood-based materials like oriented strand board (OSB), which are commonly installed in new homes, generate most of their manufacturing carbon emissions from burning wood to generate heat during production.  Plantd transforms the built environment using carbon-negative building materials derived from alternative biomass—a hardy, fast-growing grass. Four years ago, I cofounded the company with two former SpaceX engineers. To realize its ambitions, Plantd established a new agricultural supply chain innovating at every step, from building an in-house tissue culture lab to establishing full-scale greenhouse operations to supplying commercial farmers with the company’s proprietary grass.  [Photo: Courtesy of Plantd] Why grass? Because it grows incredibly fast, like bamboo, rapidly removing atmospheric carbon in the process, and possesses the structural characteristics to be transformed into durable engineered building materials.  Yet, the key to sequestering carbon in our materials is Plantd’s manufacturing technology. Our team pioneered a modular, electric-powered production line that turns grass into finished products that replace plywood and OSB in new home construction.  It’s a first-of-its-kind technology that distinguishes a Plantd production facility from every other engineered wood facility in the world; ours is the only one without a smoke stack on top of the building.  This past fall, D.R. Horton, the largest homebuilder in America, which builds about one in every 10 U.S. homes, ordered 10 million Plantd panels, enough to form the walls and roofs of 90,000 new single-family homes.   2. Software innovation: Give buildings brains  An even bigger source of building-related carbon emissions is the energy required to operate them. Globally, this accounts for 26% of all greenhouse gas emissions.   The top culprit: HVAC systems.   The heating, cooling, and ventilation equipment needed to keep us comfortable indoors are responsible for about 35% of all energy used in U.S. buildings.  The challenge is that thermostats, even the smart ones, aren’t very bright. They can track what’s already happened and react to what’s happening right now, but they cannot anticipate changes in weather, occupancy, carbon intensity of the grid, and energy costs.   BrainBox AI can. Using AI-powered intelligence, its cloud-based control system connects to the hundreds, sometimes thousands, of HVAC components in a building and sends them real-time instructions.  The company’s platform provides over 15,000 buildings worldwide—from Nordstrom to Family Dollar—with the intelligence to see six hours into the future with 96% accuracy.   By knowing the future, BrainBox AI cuts energy, costs, and carbon emissions and improves comfort. It’s an easy-to-install solution that works with existing systems and equipment.  The results? HVAC-related emissions reductions of up to 40% and energy savings as high as 25%.   3. Finance innovation: Make efficiency upgrades free  Many buildings are stuck with legacy equipment that gets the job done but consumes far more energy than their more efficient modern counterparts. Yet, new equipment can cost hundreds of thousands of dollars, often placing upgrades out of reach.  Budderfly has built one of the fastest-growing businesses in America by removing the cost barrier. The company identifies energy-intensive businesses like fast food chains and offers them a deal that sounds almost too good to be true: free upgrades to energy-efficient systems, including HVAC, lighting, refrigeration, and security. Budderfly foots the bills and shares the monthly energy savings with its customers.   Scale is key to making this business model work. Budderfly has raised nearly $1 billion to pay for the equipment it installs in customer locations. Its rapid expansion enables it to secure preferential pricing from global equipment suppliers that individual owners and franchisees could never obtain independently.  Budderfly also takes over billing, which is one less thing for customers to worry about, and gives the company a trove of data to drive further energy reductions and cost savings.  From Taco Bell to McDonald’s to Sonic, clients are guaranteed to see savings from day one. In 2024, Budderfly generated $200 million in revenue and now operates in more than 7,000 locations nationwide. Its customers’ collective energy use dropped 43% last year.  The takeaway  Whether it’s growing new materials, giving buildings the ability to think ahead, or reimagining who pays for energy systems, the low carbon economy isn’t just coming someday. It’s already being built.  Josh Dorfman is the CEO and host of Supercool.

Ask a true car or bike enthusiast to drive an EV and they’ll look at you with disdain and disgust. An EV’s quiet whirr doesn’t capture the rawness you get from the sound of a V6 engine roar, or the feeling of the gears turning as you hit the accelerator, or even the smell of gasoline, engine oil, et al. For coffee-lovers, this analogy applies to the art of coffee-brewing too. The idea of pressing a button and getting instant coffee just diminishes the beauty of a cup of golden brew. That’s where the COFFEEJACK V2 comes in, with its gloriously analog brewing method. Designed for a personal hands-on experience, this portable brewer relies entirely on manual brewing rather than automation. Armed with a rotating hand-crank, it looks like a portable coffee grinder, but the COFFEEJACK V2 actually uses that crank to build pressure (up to 10 bars of it). Rotate the crank and the COFFEEJACK V2 creates the right amount of push needed to force hot water through a portafilter filled with a puck of coffee, giving you luxuriously hand-brewed espresso complete with the crema. The best part? A compact, non-electric design that lets you literally brew your coffee anywhere, whether in the comfort of your kitchen or the idyllic wilderness of your campsite. Designer: COFFEEJACK Click Here to Buy Now: $214 $267.5 (20% off) Hurry, only 14 of 750 left! Raised over $347,400. Visually, the COFFEEJACK V2 embodies that rare intersection of industrial utility and considered aesthetics. Its cylindrical form speaks the language of serious outdoor equipment – reminiscent of high-end flashlights or premium binoculars – while the textured silicone grip bands provide both tactile feedback and visual contrast against the machined aluminum body, with inner components made from food-grade stainless steel. The knurled texture on the crank feels delightfully industrial, but the crown jewel of the V2 remains its exposed planetary gear mechanism, designed to evoke a sense of rawness seen under the hood of a vintage car or a Swiss watch. Remember that ‘joy of analog’ emotion I mentioned earlier? You won’t get that from a Nespresso pod – but the COFFEEJACK V2, ah, it feels divinely rugged. Ritualistic almost. What fascinates me about the COFFEEJACK V2 is how it rejects the automation trend dominating home coffee. This device demands engagement. You control the water temperature. You determine the grind. You generate the pressure. The result feels less like a convenience product and more like a specialized tool for the discerning coffee enthusiast who refuses to compromise, regardless of GPS coordinates. This isn’t COFFEEJACK’s first rodeo – their V1 tried to disrupt the Aeropress but came with its set of limitations, the V2 emerges in stainless steel and aluminum glory, absolutely redesigning the entire device altogether with beefed-up seals, superior thermal properties, and a new rotary input rather than the press-input of V1. The upgraded 54mm portafilter basket now accommodates a proper 18g dose, matching standard home espresso dimensions. The brilliance lies in its gear-driven pressure system. Through a series of interlocking cogs that multiply your hand strength (so it feels effortless as you use it), you can generate a legitimate 10 bars of pressure – the sweet spot for proper extraction. An analog pressure gauge provides real-time feedback, while coffee nerds can opt for the Bluetooth digital sensor to track their extraction curves. Crank the COFFEEJACK V2 and it consistently delivers pressure, passing the water through a professional-grade shower screen onto the large 54mm portafilter tamped with coffee. The process, although analog, replicates the exact procedure followed by a barista-grade espresso machine. The COFFEEJACK’s all-metal interior means no skimping or cost-cutting – the coffee is brewed the way it’s supposed to be. The result is an espresso you’d probably never believe came from simple elbow-grease and a tiny tumbler-sized device. Once you’re done savoring your coffee, cleaning up is easy too, as the dry coffee grounds pop out in the form of a solid dense puck, reducing any messy clean-up. At roughly thermos-sized but with the density of professional camera gear, the COFFEEJACK V2 occupies a practical middle ground in the packability spectrum. The insulated water chamber maintains temperature stability – crucial when you’re brewing at altitude where water boils at lower temperatures and heat dissipates quickly. The double-wall stainless steel construction ensures durability in the field, making the COFFEEJACK V2 about as adventure-friendly as you are. The market has responded enthusiastically: the Kickstarter campaign has already secured over £347,400 ($465,680 USD), demolishing its modest £7,000 ($9,282 USD) goal forty times over. Early backers can still claim one for £159 (approximately $210), below the planned £198 retail price. While that positions it as a premium product, it delivers capabilities previously unavailable at any price point. And if you still think you’re better off drinking that instant coffee or something from a Nespresso pod, you’re honestly beyond saving… Does it challenge your thousand-dollar coffee maker? Oh, absolutely – it’s designed to be as technically on-point as automated countertop machines – but with two distinct differences, first, the manual bespoke interaction, and secondly, unmatched portability that gives you the best coffee anywhere you want. Watching steam rise from a perfect crema-laden espresso while perched on some remote overlook might permanently change how you define luxury in the outdoors. The COFFEEJACK V2 reminds us that sometimes the most satisfying technologies aren’t those that remove human effort, but those that honor it in precisely the right ways. Click Here to Buy Now: $214 $267.5 (20% off) Hurry, only 14 of 750 left! Raised over $347,400.The post This Analog Espresso Maker is the Ultimate EDC Coffee Companion – Meet the COFFEEJACK V2 first appeared on Yanko Design.

Scientists have known for decades that certain neurons in the hippocampus are dedicated to remembering specific locations where an animal has been. More useful, though, is remembering where places are relative to each other, and it hasn’t been clear how those mental maps are formed. A study by MIT neuroscientist Matthew Wilson and colleagues sheds light on that question.  The researchers let mice explore mazes freely for about 30 minutes a day for several days. While the animals were wandering and while they were sleeping, the team monitored hundreds of neurons that they had engineered to flash when electrically active. Wilson’s lab has shown that animals essentially refine their memories by dreaming about their experiences. The recordings showed that the “place cells” were equally active for days. But activity in another group of cells, which were only weakly attuned to individual places, gradually changed so that it correlated not with locations, but with activity patterns among other neurons in the network. As this happened, an increasingly accurate cognitive map of the maze took shape. Sleep played a crucial role in this process: When mice explored a new maze twice with a siesta in between, the mental maps of those allowed to sleep during the break showed significant refinement, while those of mice that stayed awake did not.  “On day 1, the brain doesn’t represent the space very well,” says research scientist Wei Guo, the study’s lead author. “Neurons represent individual locations, but together they don’t form a map. But on day 5 they form a map. If you want a map, you need all these neurons to work together.”

Tiny flying robots could perform such useful tasks as pollinating crops inside multilevel warehouses, boosting yields while mitigating some of agriculture’s harmful impacts on the environment. The latest robo-bug from an MIT lab, inspired by the anatomy of the bee, comes closer to matching nature’s performance than ever before.  Led by Kevin Chen, an associate professor in the Department of Electrical Engineering and Computer Science and the senior author of a paper on the work, the team adapted an earlier flying robot composed of four identical two-winged units, combined into a rectangular device about the size of a microcassette. The wings managed to flap like an insect’s, but the bot couldn’t fly for long. One problem is that the wings would blow air into each other when flapping, reducing the lift forces they could generate. In the new design, each of the four units has a single flapping wing pointing away from the robot’s center, stabilizing the wings and boosting their lift forces. The researchers also improved the way the wings are connected to the actuators, or artificial muscles, that flap them. In previous designs, when the actuators’ movements reached the extremely high frequencies needed for flight, the devices often started buckling. That reduced the power and efficiency of the robot. Thanks in part to a new, longer wing hinge, the actuators now experience less mechanical strain and can apply more force, so the bots can fly faster, longer, and in more precise paths. The robots can precisely track a trajectory enough to spell M-I-T.COURTESY OF THE RESEARCHERS Weighing less than a paper clip, the new robotic insect can hover for more than 1,000 seconds—almost 17 minutes—without any degradation of flight precision. “When my student Yi-Hsuan Hsiao was performing that flight, he said it was the slowest 1,000 seconds he had spent in his entire life. The experiment was extremely nerve-racking,” Chen says. The new robot also reached an average speed of 35 centimeters per second, the fastest flight researchers have reported, and was able to perform body rolls and double flips. It can even precisely track a trajectory that spells M-I-T. “At the end of the day, we’ve shown flight that is 100 times longer than anyone else in the field has been able to do, so this is an extremely exciting result,” Chen says. COURTESY OF THE RESEARCHERS From here, he and his students want to see how far they can push this new design, with the goal of achieving flight for longer than 10,000 seconds. They also want to improve the precision of the robots so they could land in and take off from the center of a flower. In the long run, the researchers hope to install tiny batteries and sensors so the robots could fly and navigate outside the lab. The design has more room for those electronics now that they’ve halved the number of wings. The bots still can’t achieve the fine-tuned behavior of a real bee, Chen acknowledges. Still, he says, “with the improved lifespan and precision of this robot, we are getting closer to some very exciting applications, like assisted pollination.” 

The UK's Architects Registration Board (ARB) has advanced to the accreditation phase three new degree programs at the University of Leeds that, for the first time, offer qualifications that meet the body's new Competency Outcomes. They are said to represent the first alternative to the standing Parts 1-3 system and requirement for a minimum of two years of professional practice. Chair Alan Kershaw said he hopes their approval "represents real progress in our work to modernise architectural education and training, and shows how innovative learning providers can make the most of the flexibility opened up by ARB’s education reforms."  Our recent guide to the long road of licensure reform in the UK can be found here.

Researchers from MIT, Yale, McGill University and others found that adapting the Sequential Monte Carlo algorithm can make AI-generated code better.Read More

Training LLMs on trajectories of reasoning and tool use makes them superior at multi-step reasoning tasks.Read More

When Instagram was acquired for $1 billion in 2012, co-founder Kevin Systrom believed that joining Facebook would help Instagram’s “skyrocketing growth” reach even greater heights. In some ways, it did. Instagram now has billions of users and has since “generated many multiples of that price and then some,” Systrom said on Tuesday from a Washington, DC courtroom. But according to him, that success often came in spite of, not because of, Facebook’s help.While testifying in the Federal Trade Commission’s lawsuit to force the spin-off of Instagram and WhatsApp from Meta, Systrom said that CEO Mark Zuckerberg repeatedly withheld critical resources from Instagram and constrained its growth to avoid harming Facebook’s engagement. To the chagrin of Meta’s attorneys, Systrom also made predictions about how, in hindsight, Instagram would have probably still succeeded on its own.Over the course of about six hours, Systrom remained steady and confident on the witness stand. Zuckerberg himself sat in the same seat last week, describing how Instagram would likely not have become the social media powerhouse it is today without his help. In contrast, Systrom’s testimony portrayed Zuckerberg as a withholding and jealous boss. He described how he and Instagram’s other co-founder, Mike Krieger, quit in 2018 after growing increasingly frustrated with Zuckerberg’s meddling in Instagram’s operations. In court, Systrom was presented with an internal chart from that same year detailing the feature integrations Facebook had made with Instagram. With the help of features like notifications promoting Instagram within Facebook and cross-posting between the apps, Instagram experienced growth, while Facebook saw a neutral effect. Systrom said that, shortly before he and Krieger quit, Zuckerberg decided to end the feature integrations because, in Systrom’s view, he didn’t want Instagram to grow at the expense of Facebook. “We were a threat to their growth,” Systrom testified.“If Instagram didn’t grow as quickly, Facebook wouldn’t shrink as quickly, or plateau as quickly,” Systrom said in court. “I don’t think he [Zuckerberg] ever said it out loud that way, but that was the only reason we were having this discussion.”At the time, Instagram had just reached one billion users, which was about half of Facebook’s user base, with a fraction of the employees. Systrom felt that Zuckerberg was “underinvesting” in Instagram and giving it “zero resources,” which Systrom thought was “in stark contrast to the effort I was putting in.”According to Systrom’s telling, ego played a role. Zuckerberg was “very happy to have Instagram in the family,” he testified. “But also, I think as the founder of Facebook, he felt a lot of emotion around which one was better, meaning Instagram or Facebook, and I think there were real human emotional things going on.” “I think there were real human emotional things going on”Systrom recalled other instances where Instagram was denied the resources it needed. When Mark Zuckerberg declared that video would be the next big shift in social networking, Facebook started allocating internal resources towards the push. The company initially allocated 300 employees to making video a prominent part of Facebook, while Instagram received no additional headcount. Following the Cambridge Analytica data scandal that embroiled Facebook in controversy over its privacy practices, Systrom stated that his organization received “zero” of the billions of dollars in trust and safety resources that Zuckerberg had publicly committed to spending. Instead, he said Instagram was given access to a centralized team that was more focused on Facebook. He also described how, years earlier, Zuckerberg suddenly yanked members of the Facebook growth team who had been deployed to help Instagram. During cross-examination, Meta attorney Kevin Huff attempted to discredit Systrom’s testimony. He hardly gave an inch by maintaining that Instagram would have likely been successful as an independent company. “You deal in a world of probabilities,” he said. “You can never be sure. Some things you can be more sure of.”Huff’s questioning of Systrom got tense on several occasions. His stone-faced, one-liner responses prompted rounds of laughter in the courthouse media room, though Judge James Boasberg rarely cracked a smile. When Huff brought up an early email Systrom sent to Zuckerberg crediting an integration with Facebook for much of Instagram’s early growth, Systrom said he was only emphasizing the benefit to appease Zuckerberg. Huff then asked Systrom if he was lying to Zuckerberg in the email. Seemingly irritated, Systrom stared back and simply said, “Sir.”See More:

Solar cells are joined together during production at the SunSpark Technology Inc. manufacturing facility in Riverside, California, U.S., on Tuesday, April 3, 2018. | Photo: Getty Images Solar cells from four Southeast Asian countries that have been major suppliers to the US are facing newly increased tariffs, hiked up as high as 3,521 percent.  The tariffs affect Cambodia, Malaysia, Thailand, and Vietnam, which together accounted for more than three quarters of total module imports last year, according to Bloomberg. The tariffs essentially make the products “unmarketable” in the US, the Wall Street Journal reports. The tariffs essentially make the products “unmarketable” The move comes after long-running Commerce Department investigations into whether Chinese companies were funneling products through Southeast Asia to avoid tariffs and lower prices. It also heightens Donald Trump’s trade war with China after roiling global markets with drastic tariff proposals this month. The president called a 90-day pause on tariffs, excluding China.  US solar companies have been split on what to do about cheap solar cells from Southeast Asia. Domestic manufacturers petitioned the Commerce Department to investigate, while renewable energy project developers are worried about the tariffs raising costs for construction and the manufacturing of panels in the US using imported cells. Cambodia refused to comply with the investigation, and got hit with the highest tariffs at 3,521 percent. Duties for companies in Vietnam reach as high as 395.9 percent, 375.2 in Thailand, and 34.4 for Malaysia.  The US International Trade Commission still has to weigh in on the proposed tariffs in June to finalize them.