 |
|
|
|||||
In a new bag from Freitag, every partfrom the fabric to the zipper, straps, buckles, and the thread that holds everything togetheris made from a single material. The product is the latest example of a “monomaterial” design approach that some brands are beginning to use for a simple reason: It means that the products can actually be recycled when they eventually wear out. [Image: Freitag] Right now, a typical backpack or bag has components made from multiple different materials. Dismantling everything at the end of its life is too time-consuming and expensive to be feasible. But using one material makes it essentially as simple as recycling a plastic water bottle. “You can basically put that entire product into a shredder instead of taking things apart first,” says Elisabeth Isenegger, who leads international communications at the Switzerland-based, 32-year-old company. Then the material can be melted down and made into pellets to make something new. [Image: Freitag] The company chose a material called polyamide 6, which is commonly used in everything from textiles to toothbrush handles. Because it’s ubiquitous, there’s a market for recycling it. But it was a challenge to find sources for every component that they needed. Freitag also had to find a way to avoid a water-repellent coating that would normally be used on the fabric, but would have made the fabric unrecyclable. To do this, the company worked with one of its suppliers to develop a version of the fabric with three layers, laminated together, that was water repellent on its own. A backpack using the approach came out last year, followed by the smaller new bag, the Musette, that just launched. [Image: Freitag] If something breaks on the bag, consumers can bring it to a Freitag store and then the bag will go to one of the brand’s 10 global repair centers. (The team set up a new repair kit for the backpack and bag, since repairs can also only be made with the same material.) When the bag eventually wears out, consumers can bring it back, and Freitag will work with partners to recycle it; it’s not handled in typical municipal recycling centers. Before launching the first backpack, the company worked with the Switzerland-based Institute for Materials Technology and Plastics Processing to validate the process. Ultimately, Freitag aims to become fully circular in everything that it makes. The brand’s best-known product, messenger bags made from old truck tarps, are already a form of recycling. But the company is now also working with trucking companies to test a new type of tarp material that can also be fully recycled after it’s made into a bag.
Category:
E-Commerce
This article is republished with permission from Wonder Tools, a newsletter that helps you discover the most useful sites and apps. Subscribe here. The AI search landscape is transforming at breakneck speed. New “Deep Research” tools from ChatGPT, Gemini and Perplexity autonomously search and gather information from dozenseven hundredsof sites, then analyze and synthesize it to produce comprehensive reports. While a human might take days or weeks to produce these 30-page citation-backed reports, AI Deep Research reports are ready in minutes. Traditional AI queries deliver isolated answers to specific questions, while Deep Research tools conduct sophisticated investigations with dozens of interconnected searches. Its like the difference between a quick reference check and a thorough research expedition. Nine Practical Ways to Use AI for Deep Research AI research tools shine when you need comprehensive information on complex topics. Here are specific use cases where they excel: 1. Craft Custom Itineraries Create detailed, personalized travel plans by specifying your destination, dates, activity preferences, budget, cultural interests, and whatever else is important to you. These AI-generated itineraries often surface unexpected gems. When planning a family trip recently, my wife and I discovered a fantastic farm stay in Pennsylvania through a Perplexity query. We wouldn’t have found it otherwise. Use the results as a starting point to identify interesting possibilities, then follow up with targeted research. Specify dietary preferences, accessibility needs, and your taste in accommodations, restaurants, and entertainment, for more tailored recommendations. Ive saved a block of text about this to reuse. Use follow-up queries to get more specifics on attractions or activities that appeal to you, or to compare and contrast potential itineraries. Example: Deep Research itinerary for a family vacation in Brookline, MA. Compare results from ChatGPT, Gemini, Perplexity, and Copilot. 2. Compile Organization Reports Get comprehensive backgrounders on companies, non-profits, or any organization in minutes rather than clicking through dozens of search results. Compare similar organizations or competitors Specify format. Maybe youd like a case study format, a topical report, a chronological history, or an industry context analysis. Specify metrics of interest like funding history, revenue growth patterns, leadership changes, media coverage trends, law suits, or anything else you care about. For closely-held private organizations data may be scarce, so read results skeptically. Advanced tip: Copy excerpts from Deep Research reports into Claude to transform them into visual dashboardsincluding charts and interactive elementsusing Claude Artifacts. You can share those with colleagues. Watch: Grace Leungs helpful video illustrates how and why to try this. Example: Deep Research report on Trader Joes. 3. Research Notable People Explore backgrounds of news figures, historical personalities, or even fictional characters. Request specific information like podcast appearances, YouTube videos, or published works to build a well-rounded understanding of the individual. Ask about connections between the person and influential contemporaries Ask for lesser-known background details or contributions often overlooked Specify time periods to focus on particular life phases or epochs Example: Deep Research on Michel de Montaigne. 4. Explore Complex Concepts Learn about complex topics in any fieldfrom botany to venture capitalwith AI-structured explanations tailored to your knowledge level. Ask for real-world examples, analogies, anecdotes, quotes, common misconceptions, and step-by-step explanations. Ask for quiz or discussion questions to test your understanding. After reading the report, generate an AI tutor with a Custom GPT, Gemini Gem, or Poe bot to further strengthen your understanding. Example: Applications for AI in medical diagnosis via Gemini & Perplexity. 5. Discover Places in Depth Investigate a places historical significance, cultural development, architecture, art, music, literature, or economic, social, or political history. I find this richer, personalized context can feel more resonant than a more generic travel guide. Ask about little-known local events, hidden gems, or notable personalities Specify your interest in fashion, architecture, history, sports, or whatever else Products too: Get a backgrounder on a new type of oven youre considering, or pianos that might suit your home. Example: Help me learn about Coolidge Corner in Brookline, Massachusetts. 6. Analyze Debates and Controversies Explore complex controversies from multiple perspectives. Examine international conflicts, ethical debates, or local issues. Deep Research reports can present multiple viewpoints with examples and evidence to deepen your understanding of nuances. You can also ak for notable quotes and an annotated reading list. Ask how the debate has evolved over time and who has been involved Specify that you want evidence-based arguments from multiple disciplines Ask for areas of common ground between opposing viewpoints Example: Is generative art revolutionizing creativityor devaluing it? 7. Decode Cultural Works Gain insights into books, paintings, music, or other creative works by exploring critical analyses, historical context, and expert interpretations. This works particularly well when you’re in the middle of reading a book or have just encountered an intriguing piece of art. At the Metropolitan Museum of Art last week, I saw Tang Dynasty zodiac figurines and asked AI to help me learn about them. (More info & a generated image). Caveat: as a non-expert learning about something new, Im not always in a strong position to assess the merits of AI analysis. When in doubt, its helpful to have cited sources for gauging the quality of the information. Request info about a work’s influence on subsequent artists or movements Ask for analysis of technical innovations or stylistic elements Compare interpretations of the book, play, poem, art, or piece of music from different time periods or cultural contexts Example: Help me deepen my appreciation of Mozarts 5th Violin Concerto. 8. Explore Evolving Trends Investigate linguistic, political, fashion, sports, arts, business, or cultural trends in a particular place or time period, or compare trends across cultures. Ask for predictive insights about how the trend might evolve Include in your prompt questions about counter-trends or critics Request data visualizations if the trend has quantitative aspects, e.g. the Beanie Baby craze Example: What are top AI training programs for journalists around the world? 9. Examine Historical Context Explore historical events through multiple lenses – political, economic, social, and cultural. Direct your AI research assistant to focus on specific date ranges, if relevant. Ask it draw from diverse sources across countries and perspectives. Request primary source recommendations Ask for comparisons across countries, regions or time periods Specify your prior knowledge so the report is tailored for your context Example: Help me learn more about the history of the Dozier School for Boys. When Deep Research isn’t your best option Use other tools when youre not looking for a comprehensive research report, but instead want something quick, or for: Simple factual questions like award winners or sports results are better addressed with basic Google or Perplexity searches. Breaking news where online info is limited. Multimedia searches may work better with specialized search engines, like Listen Notes for finding someones podcast appearances. Paywalls If the open Web lacks relevant info, dont expect miracles. Further Caveats The bottom line: Check the source list before diving into a report. When you know of high quality sources, reference them in your prompt. This Claude thread helped me include high-quality sources for my Trader Joes inquiry. Keep an eye out for errors. Verify info in these reports. The presence of citations doesnt guarantee accuracy. For example, some sources may publish estimates that get treated by an AI search agent as definitive data. Research is only as good as its sources. Some subjects lack extensive source material. AI research reports may, in such cases, rely heavily on publishers with flimsy fact-checking or an axe to grind. How to strengthen Deep Research queries The quality of your prompt significantly impacts your results: Be detailed about your topic, reasons for interest, and how you’ll use the information Unlike standard Google searches where you only provide keywords, deep research queries benefit from detailed direction. Guide your AI research assistant on specific areas of focus, recommended sourcing, prior context, and formatting: how best to present its findings. Specify your preferred tone & format tables, lists, pros/cons, bullets Request tables for comparing options, pro/con lists for debates, or categorized lists for resources like podcasts, videos, and books. You can even customize language complexitygraduate-level analysis vs beginner-friendly simplicity. Provide context about your existing knowledge and audience Mention what you already know. If youll be sharing a report with colleagues, clarify that specific audiences context. If you want something brief, say so. Be patient. Quality research isnt instant. While Gemini and Perplexity typically deliver results within a few minutes, ChatGPT’s deeper analysis can take a half hour. The thoroughness of these results justifies the wait vs. instant but shallower search results. This article is republished with permission from Wonder Tools, a newsletter that helps you discover the most useful sites and apps. Subscribe here.
Category:
E-Commerce
In my decades of working in cybersecurity, I have never seen a threat quite like the one we face today. Anyones image, likeness, and voice can be replicated on a photorealistic level cheaply and quickly. Malicious actors are using this novel technology to weaponize our personhood in attacks against our own organizations, livelihoods, and loved ones. As generative AI technology advances and the line between real and synthetic content blurs even further, so does the potential risk for companies, governments, and everyday people. Businesses are especially vulnerable to the rise of applicant fraudinterviewing or hiring a phony candidate with the intent of breaching an organization for financial gain or even nation-state espionage. Gartner predicts that by 2028, 25% of job candidates globally will be fake, driven largely by AI-generated profiles. Recruiters already encounter this mounting threat by noticing unnatural movements when speaking with candidates via videoconferencing. For many companies, the proverbial front door is wide open to these attacks without adequate protection from deepfake candidates or look-alike candidate swaps in the HR interview process. Its no longer enough to just protect against the vulnerabilities in our tech stacks and internal infrastructures. We must take security a step further to address todays uncharted AI-driven threat landscape, protecting our people and organizations from fraud and extortion before trust erodes and can no longer be restored. Fraud isnt new, but it is taking a new form Heres the thing: Synthetic identity fraud happens in the real world every day, and has for years. Think of the financial industry, where stolen Social Security numbers and other government identifiers allow fraudsters to open and close accounts in other peoples names and ransack savings and retirement funds. The difference now is that hackers no longer have to lurk in the shadows. Instead, a synthetically generated person shows up to a videoconferencing meeting and speaks to you live, and 80% of the time, people will perceive the AI-generated voice as its real counterpart. How do you protect against that? Interview impersonations are not new within HR. There have been cases where an employee’s family member interviews with a company, and a different person shows up on that first day of work. But as it becomes increasingly easier to create deepfakes (taking only about 10 minutes and a web browser), it becomes increasingly more difficult to differentiate between whats real and whats fake across applicants LinkedIn profiles, résumés, and the actual candidates themselves. Preparing our HR departments for a new attack landscape Unfortunately, HR teamsoften understaffed and using outdated techare frequently perceived as the weakest part of the organization by hackers and fraudsters given their lack of security focus (other than perhaps background checks). That makes the HR department the ideal entry point for an adversary. Coming through the front door via the hiring process is often far easier and more fruitful for malicious actors than the back door (i.e., taking advantage of infrastructure vulnerabilities). Further, adversaries could even capture recordings of executives during the interview process for future impersonation attacks or gain access to product road maps or other strategic information that could compromise the company down the road. HR leaders must be aware that fraud at the hiring level can take many different forms, but they cant be the only ones. The C-suite must also recognize these potential dangers to better equip HR teams to combat deepfake and impersonation fraud on the frontlines. For example, real-time deepfake video technology can be used to impersonate someone during virtual interviews, matching facial expressions and lip-syncing. Fraudsters will also use sophisticated voice cloning to simulate accents, intonations, or entire voices. Tools that most people use every day, like ChatGPT and Claude, are being used to fabricate résumés and cover letters, and even code samples or portfolio materials tailored to specific job postings. Information gleaned at any part of the interview process can be weaponized, including an organizations competitive strengths and weaknesses. The individuals who commit applicant fraud can repurpose information to solicit personal or confidential company information that can be used later for more severe extortion. We have already seen nation-states like North Korea leverage these techniques to infiltrate enterprises through their human resources departments. Its time we reassess security at every level and within every process to protect against these threats that show no signs of slowing down. Proper policies and procedures must be in place to navigate and respond to these attacks in real time. From an HR perspective, this involves awareness training on deepfakes, policy development, and implementing solution deployment services throughout to prevent an attack. With sophisticated tools, such as advanced audio and video content authentication and verification platforms that provide alerts if a threat of a deepfake is detected, we can also better detect and mitigate deepfakes, helping our teams understand exactly which aspects of a file are synthetic or manipulated. Its no longer enough to authenticate who is accessing a system from the outside. As we increasingly rely on images, audio, and video for critical decision-making, we now have a vested interest in verifying that every piece of digital content we consume is deemed trustworthy and accurate. If we dont, were putting everyonecolleagues, executives, and ourselvesat risk.
Category:
E-Commerce
All news |
||||||||||||||||||
|
||||||||||||||||||