All Activity

This week, the Out-of-Touch Guide grapples with an eternal question of man vs. beast, sees Gen Z throw a hilarious right cross at millennial decorating trends, contemplates whether refusing to wash your face is a skincare routine or a cry for help, and explains why serving booze at weddings has become a generational flashpoint. What is "100 men vs. 1 Gorilla?"In a video posted three years ago, TikToker yuri5kpt2 was the first online to ask a seemingly simple question: who would win in a fight between 100 men and a single gorilla? It took some time for the general population to catch on to this intriguing hypothetical, but critical mass was achieved this week, and the debate has spread everywhere online. My first thought was 100 men are taking it, without a doubt. I mean, it's a hundred guys! But then I considered the overwhelming power of an enraged gorilla, how it could literally tear off limbs and bite off faces, and the scale started tipping heavily the other way. I mean, look at this thing: On the other hand, actual primatologist Dr. Tara Stoinski gives it to humans: But I don't think she's considering the panic factor. If the question assumes 100 average dudes with no training, I'm guessing about 99 of them run screaming as soon as the most unlucky of the bunch has his face bitten off. I know I would. But the gorilla is going to panic when it sees a mob of 100 men screaming at the top of their lungs! Then there's stamina to consider. According to this (self-proclaimed) animal expert, the gorilla would be gassed after killing 30 or 40 dudes, so the men would ultimately triumph, but only after taking heavy casualties. Right now, I'm leaning slightly toward humans, based on the combined stamina of 100 people and our natural survival instinct. This is the kind of question that won't be answered until someone does it for real, and good luck making that happen with all the "laws" we have. Because this is 2025, the question couldn't be left to quietly drift away like previous internet imponderables like "what would you do if a super intelligent, deadly snail was following you?" Instead, it's become a part of the tedious "gender wars" with people hijacking the format to ask questions like, "Can we get 100 women vs. accountability?" Responses include "how about 100 men vs. shutting the fuck up?" Can we just have one fight between a gorilla and 100 men without people ruining it? What is "caveman skincare?"Moving up the evolutionary ladder from gorillas, we have "caveman skincare," a minimalist approach to personal hygiene that involves not washing your face, or at least not using soap. The idea has been kicking around since at least the 1960s, but TikToker Tia Zakher brought it to the world's attention in April by documenting her experiment of not applying any products to her face, or even washing it at all. The idea is to "reset" your skin to a more natural, and presumably healthier, state. The result is not exactly pretty. But this is temporary, at least according to Zakher. She says the uneven, bumpy skin on her face is a build-up of dead skin cells that would normally be washed away, and underneath is healthy, clear skin. Most commenters don't seem to agree, and suggest that it's actually a fungal infection, or maybe she's rage-baiting, faking it for the attention, or it's just gross. There's a lot going on here culturally. Widespread interest in the technique seems like a backlash to the elaborate, expensive skincare routines many people feel they need, and the online trashing of Zakher is depressingly predictable, given that she's a young woman who is doing something other than presenting herself as alluringly as possible. But all that aside, is this actually a good idea? According to dermatologists, not really, but a little, maybe, kind of. "It could temporarily help some people whose skin is extremely irritated from overuse of products, as it reduces the chances of chemical irritation," dermatologist Dr. Angela J. Lamb told Teen Vogue, but she added that if you're still wearing makeup and working out, not washing will likely clog your pores and could make acne worse or lead to fungal infections. According to another dermatologist quoted by Teen Vogue, Dr. Asmi Berry, the caveman method is not backed by scientific evidence, and there's a better approach to dealing with strengthening your skin. "Stick to a gentle cleanser, a hydrating moisturizer, and a mineral sunscreen," Dr. Berry suggests. What is millennial green and why is it so cringe?Maybe it's my inner mean girl, but I can't get enough of members of Generations Z noticing lame details about millennial culture and savagely roasting them for it. First it was being overly into Harry Potter, then eating at millennial burger joints, and now it's "millennial green." Sometimes called "sage" or "forest green," Millennial green is an evolution of the "sad beige" aesthetic from a few years ago. It hovers around here on the Pantone scale: Credit: Pantone Once you notice it, you notice it everywhere—just as evolution eventually leads to crabs, all decorating eventually leads to millennial green. I love this trend because millennials are still young enough to care what kids think of their decorating choices, and they're posting funny/poignant videos about discovering that they're not cool: Wait, I just realized that Lifehacker's color is kind of a millennial green. We're not super cringe, are we? Oh my god. I'm going to rethink my life now. Gen Z and cash bars at weddingsOK, I'm back. Our website may be millennial green, but at least Lifehacker understands the importance of free booze at weddings. With June marriage season approaching, many young people are defending the practice of charging their guests for drinks at their weddings. This is against the order of nature and should not be allowed. Check out the discussion in this X comment section: This Tweet is currently unavailable. It might be loading or has been removed. On the other side of the debate are Gen Z members calling people "alcoholics" for expecting free drinks. Seriously, people should be able to have whatever kind of wedding they'd like, but this debate highlights another cultural divide between Gen Z and We Who Have Come Before: Youngsters aren't drinking as much as previous generations, proving that the real generation gap isn’t about napkin colors; it’s whether you think vodka should come with a price tag. Viral video of the week: I Hid in Viral YouTube Videos and Nobody Noticed… This weeks' viral video is about viral videos. So meta. YouTuber Airrack has made a name for himself with challenge and prank videos, often involving sneaking into and hiding in unlikely places, like the Superbowl or Disneyland. This week, he decided to sneak into other YouTubers' videos. The idea is that Airrack would disguise himself and appear in the background of videos from online celebrities like cooking YouTuber Nick Digiovanni, fitness influencer Jesse James West, filmmaker Darman, gadget-maker Mark Rober, and car destroyer Whistlin' Diesel. If the commenters spot him, they get a point. If they don't, Airrack does. I won't spoil the ending, but Airrack has promised that if he loses, he'll do whatever the most upvoted comment says. That comment is "Legally change your name to diddy," which really raises the stakes. Whether the entire project is a delightful sitcom-crossover-style experiment or an annoying influencer stunt depends on your point of view, but the kids, as they say, love it. View the full article

PPC helps brands expand into new markets, but real success depends on everything around it, from measurement and localization to trust, cultural differences, and more. The post Breaking Into New Markets With PPC: Key Considerations appeared first on Search Engine Journal. View the full article

Advisers who work on acquisition deals emphasize the importance of being open and transparent with employees about potential changes and differences in a new workplace. View the full article

Most marketers know the abbreviation ABC – “always be closing.” In SEO, there’s an even more important one – ABT, or “always be testing.” The more you test your theories about what works for your users, the more you’ll learn new strategies to optimize and improve your site. And testing is always valuable, even if your tests aren’t always winners (and they definitely won’t all end up the way you predict). One of the most interesting tests I’ve ever run was actually a failure. Our testing team developed some new CTA button copy, backed by user testing and keyword research, that we hypothesized would improve conversion rate. After testing our new button copy, our results showed that: The original version, the control, was actually the winner, and by a huge margin. The test pages’ conversion rate underperformed the control by about 40%. That failed test taught us that keeping our default CTA text was the best option, so we rolled back the tests to our original copy. But to put it another way, we could have lost 40% of our revenue if we hadn’t stopped to test it first. We learned that button copy can be one of the biggest factors impacting conversion rate, compared to past tests around CTA color, design, and headers. It was a good lesson in the value of testing hypotheses, measuring results, and sharing learnings. Lay the groundwork: How to test SEO the right way Before you start testing, set up a testing process that will make it easy to measure and define success or failure. Establish your hypothesis: What are you expecting to happen with each test and why? Define key metrics: Which stats do you think will be impacted the most? Can you accurately measure all of those numbers? Choose appropriate content: Which pages are good candidates based on those metrics? Set the time frame: How much time do you need to run the test to reach a high degree of confidence or statistical significance? Plan ahead: If the test works, what will you do next? What happens if you need to roll back your test? For example, if your theory is that too-similar pages are cannibalizing keyword rankings and that merging them will improve rank, start by collecting baseline data on the keywords, rank, and traffic for both pages. After merging content and setting up your 301 redirect from one page to the other, you might monitor the page for a few weeks to see if the page’s traffic and rank have improved above your baseline. 1. Learn from pages that are already performing well If you’re struggling to come up with new test theories, start with what you know already works. Find your pages that are already driving traffic, engagement, and conversions, then create theories about why they’re working. Use Google Analytics, Search Console, and other SEO tools to find your top-performing pages. You can also look for specific pages that are outperforming similar ones based on intent, topic, or relevance. Compare the pages side by side to identify what’s working about your star performer. Now, take those elements and turn them into tests to improve similar pages. Consider the most important SEO factors like: Page layout. Content structure. Headers. Readability. UX. Keyword optimization. Visuals. Links. CTAs. For example, you might find one of your key landing pages has a higher conversion rate than ones with a similar intent. You might decide to test the specific CTA language and placement from that page on similar ones. 2. Get ideas from competitive analysis Look at what your competitors are doing well and where they’re dropping the ball. This will help you find easy opportunities to beat your competitors’ strategies. Start by auditing competitors’ top pages by traffic and rank, then comparing them with your similar pages. Consider elements like: Content depth. User experience. Expertise. Schema. Internal linking. Page speed. Mobile experience. Information quality. Keyword optimization. Multimedia and visuals. Rank your test ideas based on the biggest impact opportunities. For example, if your competitors are all optimizing for video, how much estimated traffic are they getting from video SERPs? Does that mean users expect that info in video format? Even if you decide to test into videos, that doesn’t mean you need to copy your competitors’ video format exactly. Consider ways you could make your video stand out with its format, schema, content, style, etc. Even more importantly, think about whether there are things none of your competitors are doing well or at all. Could you build a better calculator? Or test a unique CTA that lets users customize their options? Or speak to a niche set of users your competitors aren’t targeting at all? 3. Add interactive features to encourage and track engagement Embedding interactive elements on your site can improve time on site, engagement rate, and bounce rate, which can also indirectly impact rank. Some examples of interactive features could include: Survey. Quiz. Calculator. Poll. Map. Dropdown. Video. Infographics. Animations. Sliders. Add tracking to anything you add or build so you can measure how often users interact with or click each element. Even a basic heatmap tool should be able to provide some rough stats on interaction rates. Then, use this data to continue to test optimal variations of each element. Depending on how your website is built, you may already have features like this that you can test on your site in new ways. If not, consider how you can test features before investing time in fully building them out, like with user testing or A/B testing. For example, a very simple way to test whether you should invest time in creating a video on a page could be to add a clickable image of a fake video screenshot with a play button. You can measure how many users click it to estimate the potential interest in actually creating a video. Optimize visual elements for search by including descriptive alt text, structured data, and fast-loading formats. Also, make sure nothing you add hurts page speed or Core Web Vitals. Get the newsletter search marketers rely on. Business email address Sign me up! Processing... See terms. 5. Run regular CTA tests across the site Your CTAs (calls to action) are what get users to take the next step in their search journey. Spend time on CRO (conversion rate optimization) by continuously testing and building different types of CTAs on your site. Think about all the actions your users might want to take on each page and encourage them to move forward with their decision by staying on your site. Who: Customize copy for the specific users coming to each page as much as you can. What: Use the right type or format of CTA for the right intent and impact. An obtrusive CTA that looks too much like an ad could cause users to leave your site, while a CTA that’s buried or not easy to understand could have the same effect. Where: Consider all the places a user might expect CTAs, such as headers, footers, sidebars, in content, etc. Also consider how you’re drawing users’ focus to each element. When: Encourage users to take action only after the user has gotten the main answer to their question. Consider placement of secondary CTAs as users get into the details or might think of other questions. Why: The copy in and around the CTA should show users how taking this action will get them to what they need. 6. Focus your testing on your top content If you’re just starting out with SEO testing, it’s a good idea to prioritize your time wisely. Focus your first tests on the high-impact content: Top traffic/engagement pages. Top of page content. The more views or clicks, the quicker the results. Just don’t test on any high-risk pages until you’re confident in your results first (meaning don’t run brand new tests on your top revenue-driving pages). Look at page heatmaps and you’ll probably find most of your users don’t get very far down most pages. The cutoff point is often much higher than you’d expect: maybe the 1st or 2nd main sections of the page, or ⅓ of the way through the page. That’s why testing the above-the-fold experience will have the biggest impact on metrics like conversion rate, bounce, and time on page. You can use heatmap data to find which features or sections users spend more time viewing or engaging with, then optimize based on your findings. You might also find users are jumping down to specific sections of the page to find what they want, which can give you insights about what content to lead with. Use this to your advantage, too. If you want to add a new feature to the site and see how it performs, add it to the top of the most relevant page that gets the most traffic. That will give you more click data to work with to measure the success of the new feature. This applies to on-page testing only. For example, if you want to test adding star rating schema to improve SERP CTR, you can add the schema but show the stars on the page in whatever format or location you like. If you want to add an FAQ to a page to try to rank in a PAA, you can put the FAQ content anywhere on the page, and it can still get pulled into SERP features. 7. Optimize content for user intent One of the simplest SEO tests involves optimizing your content for intent. Aligning your page’s content hierarchy with user intent means users get the right info at the right stage in their journey. Order: Users like getting answers to their questions as quickly and efficiently as possible. Structure: Users prefer a well-ordered page that makes it clear what the page is about. Headers: Users want to scan headlines and bold content to hunt for the info they need. Format: Users tend to prefer lists and graphics that are easy to understand over long prose. Language: Users need to understand the words and concepts you’re describing without much prior knowledge. Authoritative: Users want a solution to their problems that makes them feel confident about their decision. For example, if users are searching for “best RPGs on Switch,” they’re probably expecting to see a page with a list or table comparison of games, plus more details on why each game is the best. But for someone landing on a page about “best RPGs on Switch for 12-year-olds,” they might be shopping for a gift. Including info or filtering with the maturity rating and price of each game may be more important to feature than details about the game’s plot or art style. A good exercise to get started Pick a page with a poor bounce rate. Write down the target keyword and headline, then draft what you think would be a good outline for the page. Look at the actual headline order and copy, and see how well they match your outline. You can also try giving the page’s current section outline to someone on your team who has never read it before and asking them to rank each section in the order they’d expect for that topic. Get started with your own ABT method Start with a clear testing process, then run smart, focused experiments on high-impact content. Roll out what works, roll back what doesn’t, and document everything. Share wins, failures, and surprises with your team because every test is a chance to get smarter. ABT isn’t just a tactic – it’s how great SEO teams stay ahead. View the full article

The narrative that women entrepreneurs receive less than 2% of venture capital (VC) funding has been widely circulated. It stems from data provided by Pitchbook, a respected research firm that delivers insights on global capital markets. However, a closer examination of their data reveals a more nuanced perspective. Pitchbook only studies investments funded by VC firms, which is a big part of the market but does not include the very substantial investments made by angel investors. Significant progress has been made in these early stages of the venture market. Twenty years ago, a mere 3% of angel-funded startups were led by women. Fast-forward to today, and women now account for well over 30% of angel-funded companies. Researching and monitoring these shifts have been a critical part of our own investing journey as the co-CEOs of Golden Seeds, an organization that invests exclusively in early-stage women-led U.S. companies. Back in 2004, when Golden Seeds was founded, there was a data void. Insights on women entrepreneurs and women’s leadership were rarely collected because they simply weren’t considered important. Thankfully there is now more research into these topics. Pitchbook has been a leader in this effort, particularly as it relates to later-stage venture capital funding. But in the process of deciphering the data around female founders, a misleading narrative has inadvertently been created. Understanding the methodology behind these—and any venture—statistics is crucial to appreciating the strides made by women in the startup ecosystem. Here are the two biggest misrepresentations worth clarifying about women entrepreneurs and their ability to secure capital. Misinterpretation #1: Women get less than 2% of capital Countless articles, books, and panels have cited that only 2% of VC funding goes to women entrepreneurs. However, this interpretation is incorrect because Pitchbook collects data only on company founders, which excludes women in executive leadership roles or those who might move into executive roles and hold substantial equity. These women are not included in these calculations but they play important roles in the success of their companies. Furthermore, when including funding received by gender-diverse founding teams, the numbers reveal a more encouraging trend. In 2024, companies founded by at least one woman secured 23% of total VC capital, a considerable increase from just 9% in 2008. Additionally, the percentage of deals VCs invest in that include at least one female founder has more than doubled in the same time frame, from 12.2% to 25.4%. That number may in fact be even higher when you consider that this data excludes companies that have non-founder women in key roles who hold substantial equity. The growing presence of gender-diverse teams signals a positive shift in the funding landscape, as startup investors increasingly recognize the benefits of diverse leadership in driving business success. It also more accurately reflects the full universe of startups seeking capital today. Of course, the reality remains that 75–80% of VC funding goes to companies with all-male founding teams (although we do not have information on the gender diversity of the management teams of these companies at the time of funding). More work is needed to get the venture industry closer to realizing gender parity. And it’s critical that we analyze the current data beyond just the single point of the initial founders. Misinterpretation #2: Female founders need a male cofounder to successfully raise VC capital One notable trend in the VC space is the increased success of mixed-gender founding teams in securing investment. While some may interpret this as a disadvantage for all-female teams, this conclusion is misleading. Founding teams form in many different ways, and most investors prioritize skill, determination, and the strength of the business over gender composition. Research consistently highlights the advantages of diverse teams, including broader skill sets, varied perspectives, and enhanced problem-solving capabilities. These benefits may make mixed-gender teams attractive to investors seeking to maximize their returns. And in our experience, founding teams are increasingly more likely to include both women and men. Systemic biases, however, still persist within the VC industry. Historically, VCs have favored investing in industries such as software and AI—sectors where women have been underrepresented. Additionally, many investors prefer to back serial entrepreneurs with prior successes, a criterion that disproportionately benefits male founders due to historical inequalities in startup funding. Addressing these biases is essential to ensuring that innovative ideas from women and underrepresented founders receive the recognition and investment they deserve—and that the progress made at the earliest stages of the market continues into later stages. Women are doing well raising capital from angel investors. They are receiving funding at a comparable yield to other entrepreneurs. (The yield is the rate at which companies seeking funding receive funding.) This trend has been growing for a long time now, as more women are actively pitching their businesses to angel investors. Forty-six percent of all companies seeking funding in 2023 were women-owned, up from 5% in 2004. In addition the growth of women angels, now over 40% of angel investors in the country, would seem to have played a significant role in increasing the share of funding. It’s worth noting that the pace of VCs funding women-led companies has also been steadily improving, as described above, albeit slower than the progress in the angel market. And remember, VCs aren’t the only path to later-stage capital. Many entrepreneurs, both women and men, successfully seek subsequent funding from family offices, corporate ventures, and other high-net-worth individuals. Embracing a more empowered narrative Transforming an industry is challenging and oftentimes frustratingly slow. For 20 years, we’ve tracked the trends, educated investors, and rallied the support, both financial and otherwise, that women entrepreneurs need to grow their businesses. The world is recognizing the contributions and value of women-led companies and that progress shows up in the numbers when you look closely. But progress is a continuum, and the work isn’t finished. Encouraging greater diversity among investors, expanding funding opportunities in traditionally male-dominated industries, and addressing biases in investment decision-making will be crucial in leveling the playing field. Moreover, continued advocacy and accurate representation of data are essential in shaping mindsets and initiatives that support women-led businesses. Perpetuating a narrative that overlooks the resilience and ingenuity of women entrepreneurs over the past 20 years discredits the progress we’ve made and subtly signals defeat. What’s needed is a nuanced understanding of the ecosystem so that a clearer picture of the barriers, progress, and opportunities of women-led companies is continually embraced and acted upon. View the full article

Google may be testing replacing local pack at the top of the Google Search results for "near me" related queries with an AI Overview. So instead of showing a map and the three or so local listings, Google is showing an AI Overview with some local information.View the full article

For a while, the comforting narrative went like this: AI won’t take your job. But someone using AI will. So, all you had to do was to use AI, and even if you lost your job you could take someone else’s? The idea that you only needed to worry about AI secondhand—via another human—is in fact somewhat naive. AI is coming for your job directly. Not with fanfare or grand announcements, but through silent, pervasive creep: software agents booking meetings, writing reports, sending personalized emails, making decisions. There are even tools to send your digital clone to videoconference meetings, without people even noticing it’s not the real you—yes, an AI deepfake of your professional self capable of intervening exactly as you would, if not more cleverly. Soon, fully autonomous agents will do entire workflows without human hand-holding. So, if you are an ambitious knowledge worker the question is no longer whether AI will automate aspects of your job. It’s whether you’ll have the initiative and creativity to out-evolve the automation. The more you use AI, the more vulnerable you become Here’s the paradox you need to internalize: the more you leverage AI to become hyperproductive, the more you expose yourself to being replaced by it. It’s no different from making your memory or spatial awareness redundant by relying too much on Google Maps or Waze, or abandoning any hopes of memorizing anything because you can always reach for your smartphone. In an age where AI can handle the bulk of our cognitive labor, we risk intellectual atrophy. When Scott Galloway called AI “corporate Ozempic” he was onto something: a tool that suppresses the need to think, even as it sharpens our output. Our ancestors didn’t need gyms or Pilates classes to stay fit; survival took care of that. But we might soon require the cognitive equivalent—structured, even artificial, forms of mental exertion—just to keep our brains from becoming intellectually obese. Efficiency is a trap. If your value to an organization is framed entirely in how quickly and predictably you can produce outputs, congratulations—you’ve just turned yourself into a template. And templates are easy to automate. Does this mean you shouldn’t use AI? Absolutely not. It means you have to reinvest your newfound time intelligently. Most organizations haven’t yet figured out what to do with the massive time savings AI is generating—largely because managers, bless their quarterly obsessed hearts, lack the imagination to redesign jobs beyond output metrics. A recent survey by Deloitte found that while 94% of executives believe AI will dramatically shift work models, only 17% have a clear plan for what that shift actually looks like. Which brings us to the golden opportunity: you don’t need to wait for your manager to reimagine your job. You can start now. Indeed, here are 10 strategies to de-risk being automated by AI: 10 Ways to Avoid Being Automated by AI Reinvest time saved by AI into higher-value, human-centric tasks. Use automation to eliminate drudgery, but spend that freed-up time deepening client relationships, mentoring colleagues, or solving problems that require empathy or judgment. Bridge communication gaps. Act as the translator between technical and nontechnical teams. AI still struggles with nuance, humor, and reading the emotional temperature of a room. Combine skills in unique and strategic ways. Be a generalist with spikes—someone who blends multiple competencies across fields, forming a professional fingerprint that’s hard to replicate. Make yourself unpredictable. Routine and predictability are blueprints for automation. Engineer variability into your tasks. Experiment. Cross disciplines. Add complexity that AI can’t model easily. Strengthen emotional intelligence. Cultivate empathy, persuasion, adaptability, and the ability to resolve conflicts—core human capabilities that are still well outside AI’s reach. Own niche domain knowledge. Carve out expertise in verticals where context and nuance matter—areas where even the best AI stumbles due to lack of real-world grounding. Invest in your personal brand. Write, speak, and share your thinking. Visibility creates optionality. People hire (and retain) people they know, not templates they can replace. Master AI tools in your domain. Don’t compete with AI—promote it. Be the go-to person for AI literacy in your field. People who understand the tools are less likely to be replaced by them. Be the human-in-the-loop. AI often needs human oversight—editing, refining, validating. These judgment calls are increasingly valuable. Stay curious and adaptable. Treat this era not as a tech shift but a cognitive revolution. Your ability to unlearn and relearn will be more important than any static skill set. Evolve faster than your environment You can’t sit this out. You can’t wait and see. The dodo bird strategy—stay passive, hope predators ignore you—didn’t work out great for the dodo. Nor will it for the knowledge worker who thinks “AI-proofing” means hiding behind corporate inertia. You need to evolve faster than your environment. That means embracing AI as a tool, even as you actively cultivate the parts of yourself AI can’t touch. Learn to become a less predictable, more creative version of yourself or be ready to face automation. The choice, for now, is still yours. So, where does that leave you? Somewhere between irreplaceable and obsolete, depending on what you choose to do next. View the full article

Google is testing a new label or green text on its shopping ads to tell you when the price of that item is low. The label says "Lowest in 30 days," which says the price is the lowest it has been in the past month.View the full article

Google's AI crawler, Google-Extended, the crawler it uses for Gemini and other related AI services, does render itself and can render JavaScript, just like Googlebot, Google's main web search crawler.View the full article

Google Ads search themes for Performance Max campaign has a limit of 25 search themes per asset group. Well, according to Thomas Eccel, Google is testing doubling that limit to 50 search themes per asset group.View the full article

Since ChatGPT sparked the generative AI revolution in November 2022, interacting with AI has felt like using a digital confession booth—private, intimate, and shielded from public view (unless you choose to share). That’s about to change dramatically with Meta’s rollout of social features in its stand-alone AI app, released last week. Those quiet queries—“What’s this embarrassing rash?” or “How can I tell my wife I don’t love her anymore?”—could soon be visible to anyone scrolling through the app’s Discover tab. If society is still grappling with how to navigate artificial intelligence, Meta’s changes risk throwing even more confusion into the mix. For tech-savvy users, the shift from private to public might be manageable—they’ll at least be aware it’s happening. But most people aren’t monitoring every policy tweak from Big Tech, and may have no idea that what once felt like a private conversation with AI could now become public fodder, ripe for ridicule. (Meta did not respond to Fast Company’s request for comment.) AI has quickly become a hybrid of search engine and digital confidant. Remember the embarrassment of accidentally posting a private message publicly? Now imagine that happening on a massive scale, as millions unknowingly expose deeply personal questions and experiences. This isn’t a hypothetical concern. Posts from Meta AI users are already surfacing in the app’s social feed, including verbal queries asked via voice mode, like one user’s question about folic acid, which also revealed her age and postmenopausal status. The Discover feed is shaping up to be a slow-motion privacy disaster, as users unintentionally share raw, unfiltered pieces of their lives—far from the curated, polished image we’ve grown used to displaying on social media. Meta said in a press release that its AI app aims to “connect you with the people and things you care about,” and calls the Discover feed “a place to share and explore how others are using AI.” While the company insists that “nothing is shared unless you choose to post it,” the app nonetheless nudges people to share—and overshare—whether they fully realize it or not. View the full article

Google announced a number of new tag, analytics and data strategies on Friday including a new Google tag gateway for advertisers. Google tag gateway for advertisers allows you to run your Google tags, be it client-side or server-side through your own website infrastructure, such as a content delivery network (CDN), load balancer, or web server.View the full article

Google AdSense will remove the ability to block the video games category of ads. So if you previously told AdSense not to show ads related to video games on your site or in your apps, that block will soon not be respected.View the full article

Get B2B buying groups talking with this new playbook: persona building, buyer-led content, continuous optimization, and empowering sales teams to nurture buyer relationships. The post B2B Buyer Behavior Has Changed: Proven Strategies For Sustainable Relationships appeared first on Search Engine Journal. View the full article

You're spoiled for choice when it comes to Facebook ads — from images and videos to carousels and collections, there’s a format (and placement) for just about every goal. Advertising on Facebook also gives you the chance to experiment and find just the right approach for your brand. But with so many options, it’s important to keep up with the latest Facebook ad specs. Facebook offers a world of creative opportunities, and getting the specs right is your first step to making the most of them. If you’ve ever found yourself Googling “Facebook ad sizes” or “Meta ad specs” five minutes before launching a campaign, we’ve got you covered. In this fully updated guide, you’ll find all the current Facebook ad dimensions, file types, and technical details you need to create standout ads in 2025. Quick Tip: Add this article as a bookmark so that you never have to go spec-hunting again. 💡New to advertising on Facebook? Here’s everything you need to know about Facebook Ads to get your campaigns up and running. Jump to a section: Technical terms you need to know The anatomy of a Facebook ad Create the perfect Facebook ad: design specs for different ad types Choose the right Facebook ad format Facebook ad specs and sizes at a glance 5 quick tips for creating standout Facebook ads Keep these guidelines handy Facebook ad specs FAQ More Facebook resources Technical terms you need to knowIf all the talk about Facebook ad specs, sizes, and file types feels a little overwhelming, I hear you. But take it from a writer: the good news is that you don’t need to be a designer or a tech whiz to create Facebook ads that work. To help you cut through the jargon, here’s a quick rundown of the most common terms you’ll see when reading this article and creating your ads: Audience network: Allows your Facebook ads to appear in apps and websites outside of Facebook and Instagram that are part of Meta’s ad network.Aspect ratio: The shape of your image or video (for example, 1:1 is a square)Aspect ratio tolerance: How much your ad’s shape can be slightly off from the ideal shape and still be OK.GIF: A short, looping animation made from several images that’s good for quick, fun visuals (Read our Ultimate Guide to GIFS to learn how to create and use them).JPEG: A common type of image file that works well for photos with lots of colours.MOV: A type of video file, mostly used by Apple devices. Facebook accepts it, too.MP4: A popular video file type that Facebook likes because it looks good and loads quickly.Pixels (px): Tiny dots that make up your image or video. More pixels usually means better quality.PNG: Another image file type that’s good for pictures with text, sharp edges, or transparent backgrounds.Resolution: The clarity of your image or video. Higher resolution means it looks sharper and more professional.💡Ads are one part of your Facebook strategy, but don’t forget about optimizing your profile. Click here to find out how to create a perfect cover photo. The anatomy of a Facebook adEvery Facebook ad is built around your objective, your conversion location, and your call to action (CTA). Think of them as your ad’s purpose, destination, and invitation, which all work together to move people toward your goal. Objective: This is what you want your ad to do. Do you want more people to visit your website? Watch your video? Download your app? Facebook offers a wide range of ad objectives so you can choose what fits your goal best. Conversion location: This is where you want the action to happen. Depending on your objective, it could be your website, a mobile app, Instagram profile, Facebook Messenger, WhatsApp, or even a physical store. Note: You need to have a Facebook Page to run ads using the Meta Ads manager. You can then link your Page and Instagram account to run ads on Instagram. You can only link one Instagram account to one Facebook Page. Call to action (CTA): This is the button or prompt that encourages people to take the next step. Options include things like “Shop Now,” “Learn More,” “Send Message,” or “Download.” Picking the right one helps set clear expectations and gives your audience a helpful nudge. Once you’ve chosen an objective, you’ll be able to choose a conversion location and a call to action: Here's a table showing relevant conversion locations and call-to-action (CTA) options based on each Facebook ad objective: Ad Objective Conversion Locations Call-to-Action Options Brand Awareness Facebook Instagram Audience network Learn More Reach Facebook Instagram Messenger Audience network Learn More Send Message Call Now Traffic Website App Messenger WhatsApp Learn More Shop Now Sign Up Contact Us Book Now Get Offer Download Engagement Post Page Event Like Page RSVP Comment Share Learn More App Installs App Store (iOS/Android) Install Now Download Video Views Facebook video player Instagram Watch More Learn More Lead Generation Instant form Website Messenger Instagram direct Sign Up Get Quote Subscribe Apply Now Learn More Messages Messenger WhatsApp Instagram direct Send Message Send WhatsApp Message Conversions Website App Shop Now Sign Up Subscribe Contact Us Learn More Get Offer Catalogue Sales Website (linked to product catalogue) Shop Now View Collection Store Traffic Physical store (with map directions or store finder feature) Get Directions Call Now Learn More Visit Us Create the perfect Facebook ad: design specs for different ad types I once uploaded an ad that looked great on desktop but got weirdly cropped in Facebook Stories; lesson learned. Now I always double-check my dimensions. Using the correct Facebook ad sizes and specs is so important. When your creative meets the recommended guidelines, your ads are more likely to look great, perform better, and help you reach your goals. Quick Tip: I like using Canva to get my visuals sized just right, fast. Just set up a custom image or video using the pixel dimensions below. It makes the whole process so much easier. As an added bonus, when you’re creating organic posts, Buffer’s Canva integration allows you to create scroll-stopping images without leaving the platform. Choose the right Facebook ad formatOnce you select an objective for your ad, you’ll be guided through the process of creating your ad and get to choose one of four ad types to serve your target audience. Meta offers different types of ads for different products, objectives, and goals: Single image ad: Shows one picture with a message to grab attention.Single video ad: Plays one video to tell a story or show a product.Carousel ad: Several images or videos you can swipe through.Collection ad: Shows a big image or video with smaller pictures underneath to explore more.1. Single image ad specsWant to drive traffic to your website or app? Start with eye-catching, high-quality visuals. You can use your own photos or tap into stock images to help tell your brand’s story in a way that grabs attention and encourages clicks. Design recommendations File type: JPG or PNGRatio: 1.91:1 to 4:5Resolution:1:1 ratio: 1440 x 1440 pixels4:5 ratio: 1440 x 1800 pixelsText recommendations Primary text: 50-150 charactersHeadline: 27 charactersTechnical requirements Maximum file size: 30MBMinimum width: 600 pixelsMinimum height:1:1 ratio: 600 pixels.4:5 ratio: 750 pixels.Aspect ratio tolerance: 3%2. Single video ad specsShow off your product’s best features and capture attention with the power of sound and motion. You can upload a video you’ve already created, or use Facebook’s built-in video tools in Ads Manager to put something together. Design recommendations File type: MP4, MOV, or GIF. (Here’s a complete list of supported video formats)Ratio: 1:1 (for desktop or mobile) or 4:5 (for mobile only) Video settings: H.264 compression, square pixels, fixed frame rate, progressive scan and stereo AAC audio compression at 128kbps+Resolution: 1:1 ratio: 1440 x 1440 pixels4:5 ratio: 1440 x 1800 pixelsVideo captions: optional, but recommendedVideo sound: optional, but recommendedText recommendations Primary text: 50-150 characters Headline: 27 characters Technical requirements Video duration: 1 second to 241 minutesMaximum file size: 4GBMinimum width: 120 pixelsMinimum height: 120 pixels360 videos With some objectives, you can use a 360 video. When people see this type of ad, they can turn their device or drag their finger to move around within the video and explore every angle. 3. Carousel ad specsCarousel ads let you showcase up to ten images or videos in a single ad, each with its own link. They're perfect for highlighting different features of the same product, or even creating one long, swipeable image to tell a visual story. Design recommendations Image file type: JPG or PNG Video file type: MP4, MOV or GIF Ratio: 1:1 or 4:5Resolution: At least 1080 x 1080 pixels Text recommendations Primary text: 80 characters Headline: 45 characters Description: 18 characters Landing page URL: RequiredTechnical requirements Number of carousel cards: 2 to 10Image maximum file size: 30MBVideo maximum file size: 4GBVideo duration: 1 second to 240 minutesAspect ratio tolerance: 3%💡If you’re interested in using video ads across different placements, I highly recommend checking out this video guide provided by Facebook.4. Collection ad specsMake it easier for people to shop by displaying items from your product catalog, automatically tailored to each person who sees your ad. When people tap on the collection ad, it opens up and they’re taken into an immersive, full-screen experience (known as instant experience) where they can interact with your branded content — with the option to exit out of the ad at any time, of course. Design recommendations Image type: JPG or PNG Video file type: MP4, MOV or GIF Ratio: 1:1 Resolution: at least 1080 x 1080 pixels Text recommendations Primary text: 125 characters Headline: 40 characters Landing page URL: RequiredTechnical requirements Instant experience: RequiredImage maximum file size: 30MBVideo maximum file size: 4GB⚡Plan, schedule, and analyze your posts to Facebook Pages and Groups with Buffer's Facebook scheduling and analytics tools.Facebook ad specs and sizes at a glance Ad Type Design Recommendations Text Recommendations Technical Requirements Single Image Ad File type: JPG or PNG Ratio: 1.91:1 to 4:5 Resolution: 1:1 — 1440 × 1440 px 4:5 — 1440 × 1800 px Primary text: 50–150 characters Headline: 27 characters Max file size: 30MB Min width: 600 px Min height: 1:1 — 600 px, 4:5 — 750 px Aspect ratio tolerance: 3% Single Video Ad File type: MP4, MOV, or GIF Ratio: 1:1 (desktop & mobile) or 4:5 (mobile) Resolution: same as image Captions & sound: Optional but recommended Thumbnail: From selected video thumbnail Primary text: 50–150 characters Headline: 27 characters Duration: 1 sec to 241 min Max file size: 4GB Min size: 120 × 120 px Carousel Ad Image type: JPG or PNG Video type: MP4, MOV, or GIF Ratio: 1:1 or 4:5 Resolution: At least 1080 × 1080 px Primary text: 80 characters Headline: 45 characters Description: 18 characters Cards: 2 to 10 Image max file size: 30MB Video max file size: 4GB Video duration: 1 sec to 240 min Aspect ratio tolerance: 3% Collection Ad Image type: JPG or PNG Video type: MP4, MOV, or GIF Ratio: 1:1 Resolution: At least 1080 × 1080 px Cover: First media in instant experience Primary text: 125 characters Headline: 40 characters Instant experience: Required Image max file size: 30MB Video max file size: 4GB 💡So, what about the specs and sizes for other social media platforms? Good news: we have an updated guide to social media image sizes on 9 major networks. 5 quick tips for creating standout Facebook adsStart with a clear goal: Before you even open Ads Manager, ask yourself: What do I want this ad to achieve? This will help you select the right format, objective, conversion location, and CTA.Keep the text to a minimum: Facebook recommends using images with little or no text overlay, as these tend to perform better and reach more people.Use music legally: Facebook is strict about copyright, and using unlicensed music can get your ad taken down. Watch out for restricted content: Facebook has clear rules about what can and can’t be advertised. I once tried to run ads for a weight loss coach using before-and-after photos, and Facebook disallowed them. All that hard work wasted. Learn the rules and save yourself the headache.Make use of the Meta Ad Library: It’s a great tool for researching trends, checking out competitors, or just gathering ideas for your next campaign. Here’s a guide on using this tool effectively.Keep these guidelines handyHaving these image size guidelines on hand will help you create awesome ads that stop the scroll and get you closer to your goals. And when it comes to organic content planning, scheduling, and analytics, Buffer makes getting in front of your audience a breeze. Sign up for Buffer (for free!) to start scheduling posts to Facebook and other platforms, storing content ideas, using AI to make your life easier, and so much more. Facebook ad specs FAQWhat’s the best size for Facebook ads in 2025?The sweet spot for single image ads is 1080 x 1080 pixels — that’s a perfect square and works great in most places. But if you're going for Facebook Stories or Reels, 1080 x 1920 (vertical) is your go-to. Pro tip: No matter the objective or the ad types, always aim for high quality and a clear image. Can I use the same creative across all placements?I get it; repurposing is a time-saver. While you can use the same creative across all placements, it’s not always ideal. A square image might look great in the feed, but not so much in Stories. It’s best to tweak your ad for each placement. What file types work best for Facebook video ads?Stick with MP4 or MOV; they’re the most reliable and Facebook-friendly. Keep the video short, clear, and under 4GB, and you’re good to go. What is the format of a Facebook ad?A Facebook ad usually has a visual (image or video), headline, primary text, description, and a call-to-action button (like “Shop Now” or “Learn More”). You can also choose from different styles like single image, video, carousel, or collection, depending on your goals. More Facebook resources How to Run Facebook Ads: Beginner's Guide to Advertising on FacebookBest Time to Post on Facebook: We Analyzed 1 Million PostsMeta Ad Library 101: 7 Ways to Use the Facebook Ad Library to Improve Your AdsThe Ideal Facebook Cover Photo Size and How To Make Yours Stand Out (+ 11 Ideas and Examples)How to Schedule Facebook Posts in 3 Easy Ways (+ Save Hours Every Week)Inside the Facebook Algorithm in 2025: All the Updates You Need to KnowView the full article

Performance reviews are often arduous, but they don’t have to be. AI tools can enhance the process for both managers and employees, offering new possibilities for efficiency and fairness. From streamlining data analysis to eliminating bias, here’s how AI is transforming performance evaluations and employee development across various industries. AI Connects Dots for Comprehensive Reviews AI has significantly improved our performance review process by providing managers with a clearer, more comprehensive view of their teams. Previously, we had vast amounts of data buried across various productivity tools—including meeting notes, shared documents, messages, and task updates—but none of it was truly actionable. Let’s face it: No manager with a team of 10 can realistically remember everything that happened over the last quarter for each person. Today, the way we work—how we communicate, collaborate, and deliver—leaves behind valuable signals. AI helps connect the dots across that information to highlight key trends, surface individual contributions, and flag potential blind spots. For employees, it means their impact is more accurately recognized, even if they’re not the most vocal. For managers, it creates a more holistic, data-informed foundation for conversations around performance and development. We also believe this approach can save a tremendous amount of time during review season, when so much energy is typically wasted trying to gather feedback and recall details. Equally important, it helps managers make fairer, more balanced assessments by surfacing the full scope of each person’s contribution. Simon De Baene, CEO and Cofounder, Workleap Streamline Reviews with AI-Generated Questions I used AI to take a client’s company values, create performance questions around them, and then tier the reviews so they were applicable to entry-level employees, individual contributors, managers, leaders, and senior executives. It produced those products for me in minutes. HR professionals or managers who aren’t using AI are wasting time and missing out on major enhancements to their leadership. Kerri Roberts, Founder and CEO, Salt & Light Advisors AI Tools Enhance Review Quality and Satisfaction We have found that managers dread the performance review process as much as employees do. Both struggle with effectively articulating KPIs [key performance indicators], achievements, and challenges in the required documents and during the review itself. This contributes to the second major shared complaint regarding the “paperwork” and workload to complete the process. We encourage managers and employees alike to utilize AI tools to analyze KPI trends, provide tables and charts, and even draft the performance review to save time and reduce anxiety. Additionally, AI tools can suggest appropriate SMART [specific, measurable, achievable, relevant, and time-bound] goals for the next period and/or recommend learning and development opportunities for the employee. As always, useful output from AI requires good input. Furthermore, the employee and manager must carefully review and edit all AI-generated information to accurately and clearly represent reality. However, we have found AI tools have greatly decreased the workload of the performance review process, while at the same time increasing the quality and satisfaction with the results for everyone involved. Joe Palmer, Managing Partner, Prosperity Partners Consulting Balance AI Objectivity with Human Touch One specific example from our organization involved the marketing team, where managers had long struggled with bias and inconsistency in performance reviews. To improve the process, we introduced an AI tool that aggregated peer feedback, performance metrics, and goal progress into clear, objective drafts. It flagged subjective language and suggested more neutral alternatives, reducing bias and saving managers valuable time. However, a new challenge emerged: Employees described the AI-generated feedback as sterile—accurate but impersonal. This concern became especially clear during a departmental feedback session. To address it, we encouraged managers in marketing to use the AI drafts as a foundation, then add personal insights, context, and specific examples to restore a sense of authenticity. This balance between AI-driven objectivity and a human touch made a noticeable difference. Employees received clearer, fairer, and more meaningful feedback, while managers gained a tool that streamlined the process without losing the connection that makes reviews truly valuable. Michael Ferrara, Information Technology Specialist, Conceptual Technology AI Creates Personalized Development Plans Post-Review As part of my current doctoral research in learning and organizational change, I’ve been studying how HR leaders are actively using AI to enhance human-centric leadership practices—and performance reviews have definitely come up. One high-level HR executive I interviewed shared how they used AI to create a personalized learning and development plan immediately after a performance review. The AI helped analyze feedback and skill gaps, then recommended tailored next steps—what the employee could do now, next, and later to grow in a specific area. The employee later thanked their manager for recommendations that were on that plan, suggesting they felt supported. Another HR executive at a global automotive company used AI-enabled project management tools to analyze team metrics that correlated with performance. She felt this helped her make more objective, data-informed decisions rather than relying solely on instinct. In both cases, AI didn’t replace the human side of leadership—it amplified it by making conversations more personalized, fair, and focused on growth. Bailey Parnell, Founder and CEO, SkillsCamp Voice Notes Capture Nuanced Performance Feedback One thing that has surprised us was how well an AI-powered voice note tool worked during performance reviews not as a replacement for feedback, but as a way to capture tone, nuance, and real-time reflection. In our own staffing agency, where many of our clients rely on private staff-like housekeepers, chefs, and estate managers, soft skills matter just as much as task completion. Managers started using short voice notes to highlight specific interactions, such as how a nanny handled an unexpected visitor at the door or how a housekeeper went above routine to solve a problem without being asked. These moments used to get lost between checklists. On top of everything else, rather than treating reviews like a checklist, the voice notes created a space where real appreciation could be felt. A personal chef once told us that hearing the emotion behind the words made all the difference—it felt honest, not formal. These notes turned routine evaluations into conversations that captured what often goes unseen. In our world, where intuition and quiet consistency define excellence, giving those qualities a voice brought something far more meaningful than numbers or written summaries ever could. Brooke Barousse, CEO, Lexington Executive and Household Staffing AI Builds Objective Benchmarks for Fair Reviews We’re starting to use AI to build objective performance benchmarks to make our reviews more fair and impartial. Essentially, the AI analyzes key metrics and skill feedback from our own internal, anonymized data across similar roles, comparing performance among our project managers, engineers, or CNC [computer numerical control] machinists, for example. It helps our managers get a better grasp on ratings and performance discussions, as they can use the data to more easily identify if someone is truly excelling in their specific job or spot an area where the entire group could benefit from improvement. Our employees gain a much clearer understanding of the expectations for their role and can see how they’re performing compared to others in similar positions, which can be motivating or help pinpoint areas for development. The AI might highlight that one of our project managers consistently achieves client satisfaction scores that are 10 points higher on average than other PMs performing similar jobs, for instance. It provides solid evidence supporting positive feedback about their client skills, allowing us to go beyond mere gut feelings. Since implementing this data-driven approach, we’ve noticed that our manager calibration meetings for reviews run more smoothly and efficiently, reducing subjective debate time by 30%, because everyone is working from the same baseline comparisons to initiate the conversation. Leon Huang, CEO, RapidDirect AI Analysis Improves Review Conversations We implemented an AI feedback tool that analyzes communication patterns during performance reviews. Managers upload meeting recordings, and the AI provides insights on speaking time balance, interruption frequency, and sentiment analysis. This improved our reviews in several ways: Managers now receive data showing they dominated 70% of conversations (previously unaware), and they adjusted to achieve better balance. Employees report 40% higher satisfaction with review fairness. The AI also flags emotional responses, revealing when discussions trigger defensiveness. Most importantly, the AI tool summarizes action items and creates trackable goals, increasing follow-through by 65%. What surprised us was how the AI revealed that our female team members were interrupted twice as often as male counterparts—an insight that led to meaningful cultural change. The technology doesn’t replace human judgment, but it makes our performance conversations more balanced, actionable, and fair. Kirti Poonia, Founder, Caimera Creative Performance Profile Tracks Progress We’ve always found it challenging to review the performance of roles that aren’t tied directly to strategic goals, like our graphic designer. They don’t set quarterly targets or lead major initiatives. Their work is reactive, based on tasks assigned to them, which makes it hard to define clear goals or track measurable progress. Feedback often felt generic, and improvement was tough to gauge other than informal “good jobs.” To change that, we set up an AI-enhanced performance tracker using tools we already had access to. We connected Asana to Google Sheets through Zapier, which allowed us to automatically track things like task volume, turnaround time, and revision frequency. We also pulled in feedback from Slack, where a lot of real-time collaboration was happening. Using OpenAI, we ran sentiment analysis on both task comments and relevant Slack messages, which described how work was being received and the tone of the day-to-day communication. Together, this gave us a monthly snapshot we called the Creative Performance Profile. It helped spot progress over time and gave our designer real insights they could reflect on during their review, without needing a complex dashboard. In one case, we saw our designer’s average turnaround time improve by 22% over the quarter, while revision rates dropped by 35%. That led to a great discussion around how they were proactively clarifying briefs earlier in the process, something we wouldn’t have uncovered from the numbers alone. What’s been most valuable is how this gave us a new way to talk about progress in roles where goal-setting has always felt forced. It’s not about ranking team members against each other, but helping them see how their efforts translate into measurable growth. For the first time, our designer walked into their review with stats that reflected their day-to-day work and was able to explain where they could show improvement over the coming year. Not only did this help them grow their individual performance, but oddly, they expressed that it made them feel more part of the team in our planning and goal-setting discussions. It was just an overall win. Kyle Senger, Founder and Lead Strategist, Unalike Marketing AI Triggers Timely Check-ins Between Reviews AI is a powerful tool not only for performance reviews themselves, but also for pre-review and post-review check-ins. Instead of just standard calendar pings, we’ve experimented with systems that trigger automated reminders based on actual work data. For example, a manager could get a nudge to schedule a check-in if an employee’s key goal from the review isn’t progressing on track or if feedback indicates a challenge or bottleneck arising. This way, managers can intervene early and potentially prevent problems rather than waiting months for the formal review. From the employees’ perspective, it means they receive more regular support and feedback throughout the year. When the formal review time does arrive, it feels less like a big reveal because progress and any issues have already been discussed along the way. Traditionally, employees’ biggest complaint about reviews is that they feel like a pointless, arbitrary exercise. However, with AI reminders, it’s easier to take real action and create an ongoing conversation rather than forgetting about reviews a few weeks later until the following year. This approach is more supportive and more productive. Fineas Tatar, Co-CEO, Viva Automated Tools Spot Patterns and Reduce Bias AI has really changed performance reviews for the better. It’s made a huge difference in how managers view the work of their teams. Two tools that I absolutely love are Lattice Analytics and Betterworks. Lattice is useful because it tracks all the performance data automatically and spots patterns that might be overlooked. It has cut down prep time and helps craft feedback without bias. Betterworks, on the other hand, is useful for picking up wins that people usually forget to mention themselves by analyzing project work and communication. These tools can be game changers for efficiency when implemented, since they focus on actual data instead of just opinions. I know there are a lot of tools out there, but I think it’s best to find one or two that align with your organizational goals and leverage them for maximum benefits. Jacqueline Twillie, Leadership Officer, ZeroGap.co AI Promotes Equitable and Actionable Feedback As a former senior HR leader at a global tech company, I have observed how performance reviews can either foster growth or reinforce inequity. The thoughtful use of AI tools has begun to shift that balance when used intentionally. One impactful example: For a recent client in Big Tech, we introduced AI to support managers in writing more objective, bias-aware feedback. Performance reviews often contain vague or personality-driven comments, especially for women, people of color, and LGBTQ+ professionals. Research from Stanford and McKinsey confirms this disparity. We asked managers to run their draft feedback through an AI tool trained to flag vague, nonactionable phrases and suggest more equitable alternatives. For example, “Indira is a pleasure to work with” might prompt: “Consider elaborating on Indira’s specific contributions or business impact.” This helped leaders offer fairer, more actionable reviews and also created powerful learning moments around unconscious bias. Crucially, we do not see AI as a replacement for human leadership, but as a collaborator. Tools like ChatGPT or Gemini cannot grasp context or individual nuance, and they reflect the bias in their training data. However, they can help standardize fairness, sharpen awareness, and prompt better conversations. Used well, conversational AI can encourage leaders to ask, “Am I being fair? Am I being specific? Am I giving everyone the same chance to grow?” In a system where performance reviews shape careers and compensation, those questions matter. And AI, used wisely, can help us answer them better. Manuel Schlothauer, Founder, HeyManuel.com View the full article

The Positionless Marketer is the new marketing professional who is a triple threat with data, creative and optimization power. They blow up the traditional marketing assembly line, where roles are rigidly defined. Instead, they have agility, intelligence and execution speed in defining success. A blueprint for the way Positionless Marketers achieve this level of independence and mastery is Stephen R. Covey’s “7 Habits of Highly Effective People.” Covey’s principles, trusted by millions as an excellent foundation for personal and business growth, align with the principles of Positionless Marketing. Covey’s seven habits are divided into three basic groups. The first three emphasize personal mastery (or private victory); habits 4–6 foster collaboration and synergize with others, public victory; the final habit emphasizes the need for continual improvement. All seven are embodied in the Positionless Marketer. The 7 habits and how they align with Positionless Marketing The whole concept behind Positionless Marketing is that all members do anything and be everything — a triple threat. However, a Positionless Marketer needs to focus first on self-development, including self-discipline, strategic thinking, and time management before being able to master cross-functional execution. Habit No. 1: Be proactive Covey’s first habit is being proactive. There are three types of marketers – those who make things happen, those who watch what happens and those who wonder what happened. Positionless Marketers act, while traditional marketers often watch. The traditionalists can watch what happens and wonder what happens after campaigns fail. Positionless Marketers proactively leverage AI tools to extract insights, create campaigns and execute without waiting. Instead of asking an analyst for this data, they self-start to use AI-driven analytics to find the answers on their own. This proactive approach enables them to react immediately as the insights uncover opportunities. Reactive marketers – those who watch what happens often wait on creative, data or engineering teams before moving forward, meaning they usually react once the opportunity has already passed, and do so with outdated messaging and campaigns. Habit No. 2: Begin with the end in mind Covey points out that you need to determine where you want to go before determining how to get there. You don’t get on a jet, train or in a car before determining where you are going. Marketing is no different. It’s the outcome that matters. The messaging or campaign is how you get there. Start every initiative with a clear objective. That means defining KPIs, customer journey impact, and personalization goals before determining the messaging or marketing campaign to get you there. Remember that while AI can speed execution, strategy is still driven by humans. Positionless Marketers ensure technology serves the vision, rather than the other way around. Habit No. 3: Put first things first Covey discusses four quadrants of time management: Urgent and important (quadrant 1); urgent and not important (quadrant 2); not urgent and important (quadrant 3); and not urgent and not important (quadrant 4). Positionless Marketers understand the importance of spending their time on those efforts that truly move the needle, not on busy work. Today more than ever, AI can handle those urgent but low-level tasks that don’t require human input, such as tweaking email subject lines endlessly. So Positionless Marketers focus on strategy, creativity and decision-making while allowing AI to handle the repetitive tasks. The bottom line is Positionless Marketers spend more time in Quadrant 2 — strategy, experimentation and automation — rather than reacting to constant executional requests. Public victory: Collaboration and cross-disciplinary mastery Covey discusses a win-win philosophy. This is a foundational element of not only personal growth and business, but also of sales and marketing. A Positionless Marketer must move beyond silos and embrace collaboration, adaptability, and customer-first marketing. Habit No. 4: Think win-win Positionless Marketing isn’t about AI replacing humans—it’s about AI empowering marketers. Instead of competing with AI — a lose-lose proposition; Positionless Marketers follow a win-win concept, collaborating with AI tools to execute faster and smarter. AI helps marketers become more agile, creative and strategic, eliminating repetitive tasks and allowing them to focus on high-impact work. This also means a win-win for the brand and customers – the brand delivers effective marketing and campaigns and customers receive messaging and campaigns that truly resonate with them. Positionless Marketers leverage AI for personalization at scale, journey orchestration and content generation — then focus on refining strategy. Habit No. 5: Seek first to understand, then to be understood Covey discusses understanding the audience first, then determining communications. Positionless Marketers follow this theory by relying on customer intelligence, not gut feelings for campaigns and messaging. Instead of assuming what customers want, Positionless Marketers rely on data to understand customers. They listen to customer behavior through AI-driven insights, then act accordingly. Positionless Marketers don’t do “Ready. Fire. Aim.” marketing. They analyze, then tailor marketing efforts accordingly. Habit No. 6: Synergize Synergize is about working together to find new solutions to challenging problems, according to Covey. The best Positionless Marketers synergize by blending creativity, analytics and execution. Synergy is non-existent for traditional marketing, which has specialists in separate silos, such as email marketers, copywriters, and data analysts. The Positionless Marketer integrates skills, using AI-powered platforms to execute seamlessly across multiple functions. Positionless Marketers have cross-functional mastery as their superpower and optimize without relying on multiple departments. Renewal Covey discusses that you can’t be buried in work all the time, such as sawing down trees – all of the time. Otherwise, you (the saw) become dull and can no longer do the job at hand. Habit No. 7: Sharpen the saw Technology is evolving fast — Positionless Marketers evolve faster. The tools that define marketing automation today will be different in a year. AI is advancing exponentially, and marketers who don’t evolve will be left behind. Even AI has evolved: concepts like generative AI and agentic AI were only discussed by those who were most technically savvy a couple of years ago. Now they are Positionless Marketers. Positionless Marketers invest in upskilling They learn new AI capabilities, experiment with emerging platforms, and continually refine their marketing expertise. For them, continuous learning is non-negotiable. Positionless Marketers have the right habits Stephen Covey’s seven habits provide a logical roadmap for mastering Positionless Marketing that frees marketing teams from the limitations of fixed roles, giving every marketer the power to execute any marketing task instantly and independently. Positionless Marketing provides three transformative powers: data power, creative power and optimization power. For any marketer to realize their full capabilities as a Positionless Marketer — they have to have the right habits. View the full article

The mayhem of Mar-a-Lago is less effective than the boredom of BrusselsView the full article

A/B testing compares two versions of an element to see which performs best. Learn how & see examples. View the full article

Since before he took office, President Donald The President made his disdain for the Federal Emergency Management Agency (FEMA) clear. Now, he’s leaving survivors of severe thunderstorms and tornadoes in Arkansas without any federal aid. After large swaths of the South and Midwest were hit by deadly thunderstorms and tornadoes in March and April, Arkansas’ Republican Governor Sarah Huckabee Sanders—a frequent supporter of the president—repeatedly wrote to FEMA asking for support in her state. “The sheer magnitude of this event created disastrous amounts of debris, extensive destruction to homes and businesses, and resulted in the death of three citizens, and caused injuries to countless others,” Sanders wrote in her initial request. (Since that letter was sent, 40 people in the path of the storms were killed.) After reviewing Sanders’ pleas, which went on to describe the extent of the hazardous weather and hundreds of thousands of dollars in damage, the The President administration ultimately wrote back that it had “determined that the damage from this event was not of such severity and magnitude as to be beyond the capabilities of the state, affected local governments, and voluntary agencies,” and that it therefore would not provide supplemental federal assistance. In contrast, in 2023, former President Biden granted Arkansas’ disaster declaration request following a deadly tornado within 48 hours. Given that Arkansas is a red state that voted for The President in the 2024 election, many were shocked that the president denied Sanders’ request for aid. But this isn’t the only time that The President has turned down appeals for federal help after severe weather events—and, while disappointing, the administration’s insistence that states should help themselves during times of crisis is in line with its larger efforts to dismantle federal disaster mitigation infrastructure. Several states are denied support from FEMA Since January, The President has denied several other FEMA requests that have surprised state lawmakers. In March, North Carolina’s Democratic Governor Josh Stein wrote to ask for 180 days of extended FEMA support for recovery costs related to Hurricane Helene, which was denied by the The President administration in April. That same month, Washington’s Democratic Governor Bob Ferguson requested FEMA support for repairs after a “bomb cyclone” windstorm last November that caused an estimated $34 million in damages. His appeal was also denied. “There are very clear criteria to qualify for these emergency relief funds. Washington’s application met all of them,” Ferguson said in a statement on April 14. “This is another troubling example of the federal government withholding funding.” Most recently, in early April, The President did approve a FEMA disaster declaration in Virginia to help the state recover after severe flooding. However, he refused Republican Governor Glenn Youngkin’s request for hazard-mitigation money as part of the disaster-aid package—a step that no president has taken in nearly 30 years. The Hazard Mitigation and Grant Program (HMGP) is overseen by FEMA and allocates funds to help communities protect infrastructure from future damage after severe weather, like by elevating flood-prone homes or strengthening buildings in earthquake zones. According to Politico, the program has allocated nearly $18 billion to states to safeguard 185,000 properties. “It’s an extremely important program for hazard mitigation,” Anna Weber, senior policy analyst for climate adaptation at the Natural Resources Defense Council, told Politico. “Instead of just rebuilding, we’re building resilience so we’re preventing future damages, deaths and injuries.” Historically, presidents have paired HMGP funds with FEMA’s overall recovery efforts, accounting for about 15% of overall costs for any given disaster response. But, since early April, The President has stopped approving allocations from the program. A larger plan to dismantle federal disaster response infrastructure This scaling back of HMGP runs parallel to a larger effort within the The President administration to potentially shut down FEMA altogether. On the 2024 campaign trail, The President repeatedly spread misinformation about FEMA’s response to Hurricane Helene. In office, he’s already cut hundreds of staff from the agency, leaving its remaining staffers concerned about their ability to handle upcoming severe weather, like tornado and hurricane seasons. The administration has also withheld FEMA aid to migrant shelters, suggesting that they may have violated a law used to prosecute smugglers. Funding reductions have further resulted in FEMA canceling programs like federal fire training academy courses. In March, Kristi Noem, secretary of Homeland Security, reportedly said that her department planned to “eliminate” FEMA—a notion that The President has also echoed. And last Monday, The President named 13 members to a council tasked with recommending potential overhauls at the agency, though it’s still unclear how significant those overhauls might be. Experts have repeatedly warned that scrapping FEMA would result in a dark future for disaster relief. Now, several states—including some that voted for the president—are getting a first glimpse at that future. View the full article

Understand which SEO tasks will effectively, efficiently, and strategically optimize your website. View the full article

Johann Pauwen and Michaele Simmering founded their furniture design business, Kalon, in Los Angeles in 2007. At the time, the U.S. was entering a major recession with many industries headed for total implosion. Pauwen and Simmering, committed themselves to finding local manufacturing relationships and logged countless hours looking for factories that could deliver on their solid wood designs within the United States. It wasn’t an easy process, and the founders had to write their own playbook as they went. “We really had to beat the streets and find these places on our own,” says Simmering. “Sometimes literally you’d drive past an open roller door, see certain machines or materials, and say, ‘Oh my God, they’re making X, Y, or Z and that’s how we’d find them.” Now, nearly 20 years in, all of Kalon’s products, except for its baby crib, are made in the U.S. The profitable business supports their family as well as those of their five employees. From the outside, it might appear that Kalon is entirely insulated from the roller-coaster tariff storyline unfolding every day here in the U.S. And to some degree they are: Simmering and Pauwen say their supply chain is strong and reliable and they have few doubts about their ability to deliver their product to customers as expected. Still, the pair is pretty stressed. They’ve noticed that many of their peers in the industry are losing business and, in some cases, carrying out layoffs. Kalon itself marked its worst sales month in history in April, on the heels of The President tariff news. “I canʻt believe we built this healthy business out of nothing in a really inhospitable industry: two collapses, a pandemic, and multiple wars,” says Pauwen. “And, a move to domestic manufacturing freaks out the consumer so much, no one will spend money. Maybe that will kill us, even though we’re U.S.-produced.” Kalon “It’s been an emotional roller coaster” Pauwen and Simmering represent an ecosystem of founders who’ve invested the time and money to make and sell things in the U.S. They’ve cultivated relationships with mom-and-pop manufacturing outfits. They’ve created jobs in the local economy. They’ve made it work in the name of sustainability and community. And now, as the The President administration’s wildly shifting tariff policy has shaken the foundation of how so many small and midsize designers do business both abroad and at home, these founders of American-made brands don’t feel any more at ease than their counterparts who sit at the helm of globally produced supply chains. Kalon What’s coming next is truly anybody’s guess, and many designers in positions similar to Pauwen and Simmering say they’re just bracing for the next jolt, whether that’s due to consumer insecurity, price swings in raw materials, a dearth of manufacturing options, or something else they’ve not yet considered or experienced. “It’s been an emotional roller coaster,” says Clare Vivier, founder, CEO and creative director at leather handbag brand Clare V. Vivier’s company, which is based in Los Angeles, sits at the nexus of The President’s tariffs. She works with five separate manufacturers in L.A., along with 17 manufacturing partners across India, Europe and Asia. The leather and hardware used to make Clare V. bags, says Vivier, come from Italy and Asia respectively. “We’re a great case study of what’s going on,” she says. “Seventeen years into this company, we have 14 stores and are sold in close to 200 shops around the world. Fifty percent of our product is made in L.A. and the other 50 overseas.” Clare V. Vivier says she’s structured her business this way out of necessity—to tap into different forms of workmanship. “We don’t have the options to make woven leathers and basket bags here in the U.S.” she says. “Those artisans aren’t here.” If they were, says Vivier, she’d already be using them. These types of skills and jobs, she says, went away years ago, as the industry was retooled for less hands-on, more mechanized manufacturing methods. In other countries, though, artisans (and the infrastructure to train new talent) are still a part of local economies. “These are not widget-producing jobs,” says Vivier. “These are artisans who are trained for many years.” Clare V. Vivier has considered bringing more of her manufacturing in-house. One of the manufacturing partners she works with in Burbank, California, is family-owned and run, and the owners are looking for a succession plan as retirement nears. But for Vivier, it’s just not in the cards. “We aren’t in the position to be a manufacturing business,” she says, likening the endeavor to the knowledge jump a writer would have to make in order to suddenly buy and run a printing press. “This is a highly specialized industry you can’t expect companies to just jump into. . . . My husband is French and we have a place in France. Vuitton has opened a huge training facility outside of our town there—to train artisans. I think, wow. We just aren’t doing that in the U.S. It would be amazing.” Clare V. For Simmering and Pauwen, they’ve decided to relocate their crib manufacturing to the United States. And while the decision aligns with their ethos to manufacture in their own communities, it presents a tough balance and some hard decisions around quality and cost. “Producing in Germany is roughly on par with the U.S. in terms of material and labor costs, but the level of craft and know-how is significantly higher there, which means the end product is often of superior quality—a failure of America’s industrial policy,” says Simmering. “The U.S. partner we’re working with [on the crib] was surprised by the quality of our Eastern European production and acknowledged that matching it would be a challenge—and at a much higher cost, at least 150% more.” The long game of factory building Some businesses, like East Fork Pottery in Asheville, North Carolina, have built out a manufacturing arm to their business from the start, which has helped hedge the pile-on effect happening with tariffs. Cofounder and potter Alex Matisse says that East Fork makes more than 650,000 pieces of pottery per year in its two factories. “We are relatively insulated,” says Matisse. “Our material supply chain is domestic. Clay isn’t expensive, but we put value into it. Our greatest fear is that if we do slide into a recession, it will impact us all. Building factories takes a long time. It’s hard to think about when confidence is so unsure.” Tyler Hays, artist and founder at furniture maker BDDW, which owns two of its own manufacturing facilities, says he’s grateful he made the decision to keep all pieces of his business under one umbrella so many years ago. “We have always had the slow business approach,” he says. “And we are patting ourselves on the back a little bit. But the way this is happening is bananas, with no plan. This should have been a five-year-plan. There should have been funding for small businesses; it’s reckless.” One way Hays has been able to thrive during this time is via an auction platform that’s allowed BDDW to circumvent traditional retail altogether, offering up pieces at a discount. Hays says that’s kept consumers engaged and buying: “It’s becoming more popular, but we have seen a 5% reduction in closing price at auction.” Still, even with the confusion and chaos around tariffs, many of these founders remain deeply passionate about being American-made and revel in the spirit of community and localization it can foster. For CEO Bill Banta at Decked, being American-made is just baked into his company’s brand. Decked designs, makes and sells organizational systems that fit on the beds of pickup trucks. “There’s nothing more American than a pickup truck,” says Banta. “It’s core to the customer and there’s a lot of expectation that comes with an American-made product. Plus what we make is big and heavy and hard to ship.” Decked Banta says some of the machines used to make Decked products are as heavy as 737 aircraft, and that as the company has grown, so has its manufacturing capabilities. The business, which is based in Idaho, Utah, and Ohio, now accounts for close to 400,000 square feet of manufacturing space and tens of millions of investment in injection molding. “We are seeing volatility in raw materials—steel, resin,” he says. “They’ve been all over the place for four or five weeks.” Banta’s focus has been working with suppliers to stabilize pricing as best as possible so the price for a Decked system is the same when customers initially consider it, as when they actually buy it a month later. Additionally, as consumer insecurity dips, so do truck sales, which is directly tied to the Decked value prop. “If that binds up and the automotive supply chain gets whacked by tariffs, we’ll feel that, too.” Decked A wholesale shedding of small businesses For now, says Simmering, it’s too soon to guess what any of this means. “It comes down to the mindset of the consumer,” she says. “Will consumers, at the end of the day, feel it’s more valuable to invest in American-made products? Will the tariffs last? There isn’t clarity. Industrial retooling is expensive and a lot of independent businesses won’t be able to hang in there to see how it shakes out.” One pivot the Kalon founders have made is to offer consulting services to other American businesses looking to make things here, too. Their goal is to help other founders navigate “the complexities of local sourcing, supply chain restructuring, and sustainability-first practices with insight grounded in our two decades of experience,” says Simmering. “From the beginning, part of Kalon’s mission has been to model a different way of doing things—to build a values-based business that responds to the realities of our time: the global environmental crisis, mass overconsumption, and wasteful production. “This feels like a natural extension of that original intent—taking this as an opportunity to help others navigate this shift and continue working toward transformation from within the industry.” And while the dust of tariff swings begins to settle, says Pauwen, larger, big box businesses have the resources to relocate their operations to the U.S., pushing smaller companies out of their manufacturing relationships in one swift movement, able to promise bigger manufacturing runs and longer contracts. “At first blush, when the government is saying, ‘We’re in this for the Americans,’ that’s a great impulse,” says Simmering. “I see that we can’t all be titans of industry. We want to have national resources and jobs with integrity and meaning. But the way this seems to be executed, it’s a land grab and happening at the highest levels. There is a wholesale shedding of independent business.” View the full article

​​Recent breakthroughs in generative AI have centered largely on language and imagery—from chatbots that compose sonnets and analyze text to voice models that mimic human speech and tools that transform prompts into vivid artwork. But global chip giant Nvidia is now making a bolder claim: the next chapter of AI is about systems that take action in high-stakes, real-world scenarios. At the recent International Conference on Learning Representations (ICLR 2025) in Singapore, Nvidia unveiled more than 70 research papers showcasing advances in AI systems designed to perform complex tasks beyond the digital realm. Driving this shift are agentic and foundational AI models. Nvidia’s latest research highlights how combining these models can influence the physical world—spanning adaptive robotics, protein design, and real-time reconstruction of dynamic environments for autonomous vehicles. As demand for AI grows across industries, Nvidia is positioning itself as a core infrastructure provider powering this new era of intelligent action. Bryan Catanzaro, vice president of applied deep learning research at Nvidia, described the company’s new direction as a full-stack AI initiative. “We aim to accelerate every level of the computing stack to amplify the impact and utility of AI across industries,” he tells Fast Company. “For AI to be truly useful, it must evolve beyond traditional applications and engage meaningfully with real-world use cases. That means building systems capable of reasoning, decision-making, and interacting with the real-world environment to solve practical problems.” Among the research presented, four models stood out—one of the most promising being Skill Reuse via Skill Adaptation (SRSA). This AI framework enables robots to handle unfamiliar tasks without retraining from scratch—a longstanding hurdle in robotics. While most robotic AI systems have focused on basic tasks like picking up objects, more complex jobs such as precision assembly on factory lines remain difficult. Nvidia’s SRSA model aims to overcome that challenge by leveraging a library of previously learned skills to help robots adapt more quickly. “When faced with a new challenge, the SRSA approach analyzes which existing skill is most similar to the new task, then adapts and extends it as a foundation for learning,” Catanzaro says. “This brings us a significant step closer to achieving generalization across tasks, something that’s crucial for making robots more flexible and useful in the real world.” To make accurate predictions, the system considers object shapes, movements, and expert strategies for similar tasks. According to one research paper, SRSA improved success rates on unseen tasks by 19% and required 2.4 times fewer training samples than existing methods. “Over time, we expect this kind of self-reflective, adaptive learning to be transformative for industries like manufacturing, logistics, and disaster response—fields where environments are dynamic and robots need to quickly adapt without extensive retraining,” Catanzaro says. Biotech breakthroughs The biotech sector has traditionally lagged in adopting cutting-edge AI, hindered by data scarcity and the opaque nature of many algorithms. Protein design, essential to drug development, is often hampered by proprietary data silos that slow progress and stifle innovation. To address this, Nvidia introduced Proteína—a large-scale generative model for designing entirely new protein backbones. Built using a powerful class of generative models, it can produce longer, more diverse, and functional proteins—up to 800 amino acids in length. Nvidia claims it outperforms models like Google DeepMind’s Genie 2 and Generate Biomedicines’ Chroma, especially in generating large-chain proteins. According to a paper on Proteína, the team trained the model using 21 million high-quality synthetic protein structures and improved learning thanks to new guidance strategies that ensure realistic outputs during generation. This breakthrough could transform enzyme engineering (and, by extension, vaccine development) by enabling researchers to design novel molecules beyond what occurs in nature. “What makes it especially powerful is its ability to generate proteins with specific shapes and properties, guided by structural labels,” Catanzaro says. “This gives scientists an unprecedented level of control over the design process—allowing them to create entirely new molecules tailored for specific purposes, like new medicines or advanced materials.” A new AI tool for autonomous vehicles Another standout from ICLR 2025 is Spatio-Temporal Occupancy Reconstruction Machine (STORM), an AI model capable of reconstructing dynamic 3D environments—like city streets or forest trails—in under 200 milliseconds. With minimal video input, it produces detailed, real-time spatial maps that can inform rapid machine decision-making. Nvidia sees STORM as a tool for autonomous vehicles, drones, and augmented reality systems navigating complex, moving environments. “One of the biggest backlogs in current models is that they often rely heavily on optimization—an iterative process that takes time to refine and produce accurate 3D reconstructions,” says Catanzaro. “STORM tackles this by achieving high-accuracy results in a single pass, significantly speeding up the process without sacrificing quality.” STORM’s potential extends beyond vehicles. Catanzaro envisions applications in consumer tech, such as AR glasses capable of mapping a live sports game in real time—allowing viewers to experience the event as if they were on the field. “STORM’s real-time environmental intelligence moves us closer to a future where machines and devices can perceive, understand, and interact with the physical world as fluidly as humans do,” he says. While STORM is built to help machines understand the physical world in real time, Nvidia is also pushing the boundaries of how large language models reason—through a project called Nemotron-MIND. This 138-billion-token synthetic pretraining data set is designed to enhance both mathematical and general reasoning. At its core is MIND, a new framework that turns raw math-heavy web documents into rich, multi-turn conversations that mirror how humans work through problems together. By turning dense math documents into conversations between people with different levels of understanding, MIND helps AI models break down problems step by step and explain them naturally. This method doesn’t just teach models what the right answer is—it helps them learn how to think through problems like a person would. According to its research paper, a seven-billion-parameter model trained on just four billion tokens of MIND-style dialogue outperformed much larger models trained on traditional data sets. It showed significant gains on key reasoning benchmarks like GSM8K (grade school math), MATH, and MMLU (massive multitask language understanding), and achieved a 2.5 percent boost in general reasoning when integrated into an LLM. Can startups and researchers keep up? Training and deploying advanced AI models requires substantial GPU resources, often out of reach for smaller players. To close this gap, Nvidia is rolling out its next-gen AI models through Nvidia Inference Microservices (NIMs), a suite of containerized, cloud-native tools designed to simplify deployment across different infrastructures. NIM includes prebuilt inference engines for a wide array of models, helping organizations integrate and scale AI with fewer computing resources. “Improving efficiency has always been a major focus for us,” Catanzaro says. “Ultimately, our goal is to democratize access to AI capabilities and make deployment practical at every scale, regardless of their computing resources, to harness the power of AI.” As agentic and foundational AI becomes more capable and more embodied, the future of tech may hinge on how effectively it works with humans. “It’s critical to identify and support use cases across diverse fields,” Catanzaro says. View the full article

Promise is similar to a controversial enticement to University of California to halt a flood of redemptionsView the full article

A little over two years ago, AI avatars took the internet by storm as people flocked to apps like Lensa, which generated idealized, often fantastical portraits of themselves. But in the ever-elusive offline world, another, quieter trend has been bubbling up: real portraits, made by real people. Portrait commissions have been on the rise. In 2024, the Royal Society of Portrait Painters, headquartered in London, saw a 40% increase in portrait commissions from American clients now make up roughly 20% of their total. “The U.S. has a fascination with the Royal Family more than we do sometimes,” says Martina Merelli, fine art commissions manager at the Royal Society of Portrait Painters. “It’s an acknowledgment of the quality of work.” It’s no wonder Americans are fascinated. Since its founding in 1891, the Royal Society of Portrait Painters, also known as RP, has been the society of choice for the British Royal Family’s public and private commissions. Its members have famously painted the late Queen Elizabeth II, Princess Diana, the Duke of Edinburgh, the Prince and Princess of Wales, and Prince Harry. Notable figures like Dame Judy Dench, Sir David Attenborough, and Stephen Hawking have also been captured on canvas. the Royal Society of Portrait Painters But the commission service isn’t limited to the elite. As long as you have disposable income (a head and shoulder begins at around $6,500) anyone can commission a portrait. At a time where AI is squashing many artists, this particular art form is enduring—perhaps as a symbol of our need for tangible human connection. A brief history of portraiture Portraits, like art more broadly, have long been seen as a mirror to society. Before the camera was invented, the only way to record someone’s likeness was to paint, or sculpt, a portrait of them. But portraits were never just a record—they were signifiers of wealth, taste, and power. In ancient Egypt, painted portraits were placed over mummies to guide them into the afterlife. In Ancient Rome, they were used to commemorate the dead and assert lineage. Emperors used them to reinforce authority. Dictators turned them into propaganda tools. One of the first portraits to depict a merchant couple from the middle ranks of society appeared during the Renaissance, when the focus expanded from rulers, nobility, and clergy, to wealthy merchants, bankers, and scholars. Today, portraiture remains intertwined with global politics and economic tides. “It’s no secret that many of our clients are brokers, bankers, hedge fund managers—people whose decisions are deeply affected by how the market is going,” Merelli says. In 2024, the U.K. saw two major elections. These ushered in a transition from a conservative to a Labour government that directly impacted the tax structures around private schooling. Merelli recalls one acquaintance with three daughters in private schools remarking that taxation money used to be their art money. The new faces of portraiture the Royal Society of Portrait Painters Despite its exclusive history, over the past few decades, the art of portraiture has become more accessible. Frances Bell, an RP member who has been painting portraits for over 20 years, says her clientele now includes newlyweds, young professionals, and parents wanting to leave behind a tangible heirloom. “It’s a time stamp,” she says. “Something important they will carry on.” Institutional portraits of CEOs, lawyers, chancellors, and the like still account for a big portion of the market. (Bell has also painted members of the royal family but these are cloaked in NDAs.) She believes the impulse behind a portrait commission often goes deeper than vanity. “I’m not saying it’s not there, I think it’s there for all of us, but I get people who want a little thrum of the life force to be put on into the canvas to last forever,” she says. “It’s that feeling of posterity, and permanence.” Unsurprisingly, that kind of posterity doesn’t come cheap. Merelli—who often acts as “cupid” between prospective clients and painters at the RP—says the average price for a portrait in 2025 has decreased from what it used to be, but it still hovers around $13,000–$20,000. “You can go up to $130,000 depending who the artist is, what brief you have, but a comfortable number is probably $66,000 to $80,000 if you want a full length of yourself with your house in the background and the dogs.” (Frances, who was trained at the prestigious Charles. H. Cecil Studios in Florence, charges $10,000 and upwards for a head and shoulders painting—or about $4,000 for a charcoal drawing.) A proud antithesis to AI That portraiture remains popular is both a rejection of the zeitgeist and, paradoxically, a natural extension of it. It is a slow process that can take countless hours over many sittings, and that is precisely why it is appealing. “It’s quite confessional,” says Bell, who places great importance on the in-person sittings. “I have their secrets coming out of my ears.” Everyone interviewed for this story emphasized the intimacy of the sitting process. Something about two people breathing the same air, in the same room, and looking at each other for hours. For Anthony Connolly, president of the Royal Society of Portrait Painters, this dynamic even shapes the vocabulary painters use. While photographs shoot, painters find a presence, come to a lightness. “You’re there, with your model and it’s like a triangular conversation, where the third point of the triangle is the thing you’re making,” says Connolly. The connection goes both ways. For the painter, it’s an act of seeing. For the sitter, it’s an act of being seen. It’s a bonding experience—an art form—that no algorithm can ever replicate. An investment piece Claudia Fisher, an American who moved to the UK around the beginning of the pandemic, was not allowed to divulge the cost of her painting—a head and shoulder by painter Paul Brason. Having never owned a piece of art before, the cost was “one giant gulp,” she says. But she has no regrets. the Royal Society of Portrait Painters Fisher, now 69, was reading a book about the social history of tiaras when I called her. After a multifaceted career as an opera singer and a classical architecture designer, she has turned to fashion and today runs a fashion label called Belle Brummell, which makes luxury jackets inspired by 18th and 19th century British couture. Fisher wanted her portrait to act as a marketing tool for her designs. She had just wrapped up the first prototype of her jacket, when it dawned on her: what better way to evoke the historical spirit of her brand than to be portrayed in one of her own designs, in a composition reminiscent of the era? “I’ve always loved the idea of getting a portrait done because I had vision of myself being in a gorgeous dress,” she says with a laugh. “It wasn’t about immortalizing me, I just wanted a pretty dress.” She got a pretty jacket instead. Fisher made four separate trips to Bath, where Brason lives, on four separate occasions. Brason also traveled to her and her husband’s house in Brighton to get a sense of her personality at home, take reference photos, and do a pencil sketch. The two are still in touch. “If we’re in the area I’ll call and see if he’s around,” she says. “These relationships continue.” View the full article