 |
|
|
|||||
This drone is so small that it can sneak anywhere. Flying with the stability and agility of a normal quadcopter, its design is unlike anything youve seen before. The tiny aircraft, which could fly comfortably through a Pringles can, also has a built-in camera. Imagine the Death Star’s trench-run-like possibilities. I wanted to build the worlds smallest FPV drone, declares its creator in his how-to video. While there are other commercial drones that are almost as small, I couldn’t find a true first-person-view dronea remote-controlled aircraft you can maneuver with VR glasses onthat could approach the diminutive size of this thingamajig. To create the drone, the inventor, who goes by the YouTube alias Hoarder Sam, needed to redesign traditional drones, defying the accepted wisdom in the drone community that quadcopters with a 2.5-inch (65-millimeter) frame are the absolute minimum for stable flight. His flying critter measures just 0.86 inches between rotors, and yet it flies with perfect precision. Hoarder Sams genius wasn’t in inventing new parts or embracing biomimicry, like experimental robotic insects, but in radically rearranging existing parts. The core idea was inspired by an old community design known as a “bone drone,” which overlaps its propellers to create an extremely narrow profile, allowing it to navigate tight spaces. He started by gutting a popular commercial micro-drone, the BetaFPV Air65, and transplanting its electronics into a completely new, much more compact body. To understand how radical this is, you have to consider how a standard quadcopter works. Drones achieve stability and movement by precisely controlling the speed of four propellers, each placed at the corner of a square or “X” frame. This layout gives it a stable center of gravity and allows for straightforward control logic. A bone drone throws that convention out the window by stacking the motors and propellers closer together, creating an I shape that looks like a comic-book dog bone. This arrangement makes the drone inherently unstable and much harder to control, as the physics of its lift and balance are completely altered. Sams challenge was to make this unstable design work at a micro scale. Using SolidWorks, a 3D modeling program, he designed a new chassis with just 0.86 inches (22 millimeters) between the motorsa nearly 70% reduction from the original Air65. The new 3D-printed skeleton sandwiches the flight controller between two plates and positions the motors on offset mounts. The final structure, fully assembled with its battery and camera, weighs just under an ounce (25 grams). Microsurgery and key components The drones brain is its five-in-one flight controller, which combines five essential componentsthe flight controller itself, an electronic speed controller (ESC) for each motor, and the radio receiveronto a single tiny board. It uses an ICM42688P gyroscopea device that knows the orientation of whatever object it’s attached toand runs on a central processor unit with an unusually high amount of computing power for its size. This processing is crucial, because the drone’s strange layout requires constant, lightning-fast calculations to stay in the air. The gyroscope feeds motion data to the processor thousands of times per second, and the processor adjusts motor speeds independently in real time to keep the drone from tumbling out of the sky. This brain is paired with four motors that spin up to 23,000 times per minute, designed to power quick adjustments for its extreme agility. Power comes from a LiPo battery that’s a bit bigger than a quarter. As with the other core electronics, he reused the camera from the original BetaFPV Air, transplanting it to his reduced micro-chassis. Taming the software With the hardware assembled, the rebuilt drone was a totally new beast, so its original control software couldn’t handle its flight. The “bone” configuration completely confused the system. Using Betaflight, an open-source drone configuration software, Sam rewrote the flight parameters from scratch. He discovered the flight controllers orientation was off by 45 degrees, and the motor configuration was wrong. He had to correct the yaw anglethe drone’s rotation on its vertical axisand then remap the motors one by one in the software until the system understood the new physical layout. After several trials, the micro-drone finally hovered as expected, responding to his controls with surprising stability. Despite its extreme design, the machine is remarkably functional, albeit limited by its batterys size: It achieves a flight time of two and a half minutes. Thats only 30 seconds less than the commercial drone it was born from, but still too little time to be practical for, say, military surveillance missions. It also has to be hand-launched to prevent the propellers from hitting the groundbut once airborne, it demonstrates incredible agility. The latter is not a showstopper. The former could be fixed, perhaps, with wireless power using microwaves. I have no doubt that will happen. We live in the era of the drone. As the war in Ukraine keeps raging on, we are seeing daily iteration and innovation in drones of all sizes and form factors. From Cessna-size aircraft and drones that think they are cruise missiles to hypersonic drone motherships, there is no end to this particular flavor of destructive creativity. Ukraine used radios to jam Russian drone transmissions. So Russia responded by tethering its drones with direct cable connections as long as 25 miles. The idea of swarms of insect-size snooping drones that could be easily deployed by troops to map a terrain, locate enemies, or establish defensive perimeters seems like the kind of application every army will want to have. If only an insect-size flying camera could be limited to creative selfies.
Category:
E-Commerce
When AI wearable company Friend blanketed New York City with ads last month, there was significant backlash. Many of the company’s ads (which included rage-baiting copy like, Ill never bail on our dinner plans) ended up defaced with graffiti that called the product AI trash, surveillance capitalism, and a tool to profit off of loneliness. Despite the campaign running in New York, it struck a national nerve as it became a lightening rod for people’s feelings around AI. It was only a matter of time before the brands got in on the debate. A couple weeks after the campaign’s debut, beer giant Heineken joined the chat, posting on Instagram: The best way to make a friend is over a beer. It touted its own social wearablea bottle openerthat bears a striking resemblance to the AI-powered Friend necklace. [Image: Heineken US] Now, the brand has turned that into a new outdoor ad campaign around New York, adding that the brand has been social networking since 1873. Created with agency Le Pub New York, it is a silly poke at the NYC-centric zeitgeist for Heineken. But its also the latest in a consistent string of work by the brand over the years that has aimed to remind people to put down their phones and log off social media in favor of IRL social interaction. The new ads feature the hashtag #SocialOffSocials, harking back to the Social Off Socials campaign the brand launched in April. Built around the premise that adults spend too much time online, but also feel trapped in a vicious cycle of social media addiction, it starred Joe Jonas, Dude with Sign, Lil Cherry, and Paul Olima. For that campaign, Heineken commissioned a study of 17,000 adults in the U.S., U.K., and seven other international markets and found that more than half of adults feel overwhelmed keeping up-to-date with social media. And nearly two-thirds say they are nostalgic for the 1990s when there were no smartphones. More social, less social media Earlier this year in South Africa, the brand created an installation in a mall so that people watching soccer on their phones alone could actually combine their screens to make one giant, collective viewing experience. The brand also created a limited edition phone case called The Flipper, that would flip your phone over to screen down when it heard the word, Cheers. Meanwhile, last years The Boring Phone tapped into the dumb phone trend among Gen Z. Created with streetwear retail brand Bodega, Heineken made 5,000 Boring Phones to give away. But the message is very much the same: It’s time to ditch the phone for a real social life. I reached out to both Heineken and Le Pub for comment, and to find out if the Friend-like bottle openers will be available to the public. This story will be updated as soon as I hear back.
Category:
E-Commerce
As tech companies shell out millions for top AI talenteven reportedly billionsregular rank-and-file employees are left wondering how to get in on the action and land a job in artificial intelligence. One report found that job postings that mention needing at least one AI skill had salaries 28% higher than other jobs, which translates to $18,000 more. Jobs that required two AI skills had a 43% salary jump. To begin with, its worth considering where the AI jobs are and how this intersects with your interests and existing skills. Many jobs in AI can roughly be divided into five different categories: researchers engineers business strategists domain experts policymakers Researchers bring a deep understanding of neural networks and algorithm design to the table and can push the technology forward, but this is a very small pool and typically requires a PhD. Engineers typically have programming skills that they can use to build AI applications. Business strategists can fold AI into their companys workflows and processes, or spearhead product development. Domain experts understand how to apply AI to their field, while policymakers can craft AI ethics and use guidelines. But what do you do once youve identified where you want to go? Getting experience in AI, and developing skills in it, is a tricky proposition because the field is still so nascent. Plus, things are evolving at breakneck speed; what worked a couple years ago may not be a silver bullet today. But some strategiesbeing scrappy, curious, and adaptablecould prove timeless. We interviewed both HR and recruiting pros, as well as people who have managed to build up their AI skills to land a job in the industry, to learn: What AI industry insiders at LinkedIn and Amazon recommend are the surefire ways to get a hiring manager’s attention How workers are turning their regular jobs into “AI jobs” to get experience Where one talent recruiter looks to see if someone is working on developing AI skills 1. Figure out ways to learn on the job While companies such as Boston Consulting Group (BCG) and Thomson Reuters are rolling out company-wide initiatives to ensure their entire staff gets trained in AI, that isnt true of most companies. Only 2 in 5 employees report receiving AI training on the job. If your company doesnt have AI training, get on projects that do involve AI. Get some experience at your existing company before you try to jump into a truly AI-focused role, says Cheryl Yuran, chief human resources officer at Absorb Software, an AI-powered learning platform provider. Have something on your résumé to talk about from an AI standpoint. Yuran points out that Absorb isnt able to find enough people with AI experience for all of their teams. Thats how few people are out there in the workforce with an actual background in it. Instead, the company makes sure there are one or two members with AI experience on their teams. The remaining jobs go to candidates or insiders who demonstrate they can add value, whether its deep product knowledge or excellent communication skills. If there arent AI projects or initiatives at your job, create them. Or experiment with ways to use AI to help you do your job. Gabriel Harp, a former product manager for multiple companies in academic publishing, oversaw the launch of an AI-powered writing assistant in 2023 at Research Square, an Inc. 5000 company. Although my degree is in English and German, I’ve spent more than a decade building software products, Harp says. For the AI writing assistant, Harp set the initial vision and scope of the project, working on the branding and go-to-market strategy, conducting quality analysis, and much more. Harp wasnt an engineer, yet he still leveraged his background to get great AI experience just before it was popular (or needed) to have any. Since then, hes served as head of product strategy at a startup that uses AI to build privacy tools. When Harp went on the job market, he had plenty to discuss during interviews, although he has a degree in the humanities. Since Id been using AI in the workplace, I was more familiar than the average person with these tools, he says. He recently landed a senior staff product manager job at Mozilla. Were seeing a lot of emerging talent or people who want to shift their career path, says Prashanthi Padmanabhan, VP of engineering at LinkedIn, who regularly hires for AI talent. Nothing beats showing youve actually [used AI] on the job.” 2. Take a course If getting close to an AI project at work isnt an option, you can always take courses. Right before the pandemic, Amanda Caswell was working as a copy lead at Amazon when she became interested in AI. She started listening to podcasts about AI and signed up for courses, including an online prompt engineering class at Arizona State University, an AI boot camp by OpenAI, and a generative AI and prompt engineering master class by LinkedIn. Start at the 101 level, even if you have some experience, she says. That way youll know industry best practices, which can help you teach others. Because who knows? You might have to do a job in AI training. In 2020, Caswell started getting gigs as a prompt engineer at Upwork and has made close to $200,000 on the platform, only working about 20 hours a week. In addition, her knowledge of prompt engineering helped her land a job as an AI journalist at Toms Guide. Similarly, Cesar Sanchez, a full-stack engineer (who is now an AI engineer) became interested in AI in 2023. He immediately signed up for a Coursera course on generative AI with large language models to get an understanding of the fundamentals. It was a great decision. It offered me a strong foundation and helped me understand the theory, Sanchez says. He also signed up for another course that offered im access to a network of AI engineers. While I didnt necessarily learn new things, I was able to connect with other engineers and compare my skills to what else was out there in the market, he adds. Plus, I got lots of free credits for using tools and platforms. 3. Take on a side project However, even if you arent able to fold AI into the job or take a course, recruiters say theres always the trusty side project. Having a side gig is often a privilege thats unavailable to some, but having one can sometimes grow into something that’s more full-time, sustainable, and meaningful, regardless of the field. AI, experts say, may be no different. A lot of candidates will say, I just focus full-time on my current role, says Taylor King, CEO of Foundation Talent, which recruits for top tech startups. But the ones really thriving are the people who dive headfirst into new AI or LLM tools, constantly experimenting and building on the side,” he adds. “An active GitHub tells you theyre genuinely curioussomeone whos growing beyond the boundaries of their job, not defined by it. (A McKinsey report found that people who are adaptable are 24% more likely to be employed.) Nico Jochnick had no background in AI, but managed to land a job as lead engineer at Anara, an AI startup that helps research teams organize and write scientific papers. He says he got a job in AI because of his experience using AI for side projects. I was fascinated with AI and using Cursor to code side projects, and was doing hackathons, he says. [Anaras founder] and I knew these tools were giving us tons of leverage, and we connected over that. While Harp, now at Mozilla, was job searching, he also worked on AI side projects, such as using AI coding tools to create a bingo game for his favorite podcast, as well as a recruiting tool in ChatGPT that allowed recruiters to ask questions about his work experience. I was worried about getting rusty, he says. I needed to continue experimenting with the tools out there. 4. Create your own job Ben Christopher, a screenwriter, taught himself to code in order to keep the lights on. He started experimenting with AI in 2022 and built Speed Read AI, a tool that summarizes scripts and provides business insights, such as budget estimates, for Hollywood executives. I started showing it to some people in the industry, and got enough feedback where people said, Well pay for that, Christopher said. Today, his team is five people strong with a growing customer base. (Christopher is careful to stress the point of Speed Read AI is to help Hollywood executives dig through massive slush piles and find more unique scripts.) Meanwhile, Victoria Lee originally trained as a lawyer but then took a coding boot camp when she felt like she was getting pigeonholed in her career development. She graduated from the boot camp and got her first coding job in 2022, a few months before ChatGPT launched publicly. In her spare time, she had started putting publicly available legal contracts into ChatGPT for analysis and comparing them with her own. She built an understanding of what ChatGPT did well, and where it had gaps. Lee realized the legal industry was embracing AI, and that she was perfectly positioned to fill a gap; she knew what lawyers wanted and also knew how to speak to engineers. She landed a job in product strategy at eBrevia, which uses AI in mergers and acquisitions (M&A) due diligence. However, Lee realized she could add more value by creating her own company. Today, she provides legal services for, as well as works with, mid-market law firms to help them implement AI and craft AI policies. Lee recommends that people who want to go into AI should identify their specialty and build knowledge to understand how it can work better with AI, or where AI currently falls short. Jochnick has since left Anara to found his own AI-powered company, which is still in stealth mode. The people Id hire are already building projects and putting them out in the world, he says. In fact, Jochnick notes the biggest mistake you can make today when experimenting with AI is not trying. Its insane to see how much more powerful you can become in a few months. This is a really fun journey to be on. Everyone should be upskilling themselves.”
Category:
E-Commerce
All news |
||||||||||||||||||
|
||||||||||||||||||