himala AI

Sound familiar? Whether it’s juggling projects, jumping between emails, or tackling last-minute deadlines, sometimes life feels like it’s running on a browser with way too many tabs open. 😅 And let’s be honest—some of them haven’t been clicked in weeks.But here’s the fun part: Even in the chaos, there’s creativity. In the noise, there’s opportunity. And if we can figure out where that music is coming from, we’ll be just fine. So if you’re also navigating through a screen full of tabs and thoughts - take a breath, find your focus, and remember that sometimes the best ideas come from the tab you almost forgot about. 🧠💡 💬 What’s your secret to handle all your 'open tabs' (aka tasks)? #WorkLifeBalance #MentalWellbeing #ProductivityHacks #Mindfulness #ModernWork #MentalTabs #TabOverload #Multitasking #himalaAI

🎉  Official Reveal – Meet himala 🎉 Today, we say goodbye to #Oxolo and hello to himala. This is more than a name change—it’s a step forward into what we’ve been building together: a digital partner designed to bring balance, clarity, and focus to your life. We created himala to help you cut through the chaos and focus on what truly matters. From streamlining your daily tasks to helping you achieve long-term goals, himala is here to make life simpler, clearer, and more intentional. This is just the beginning, and we’re so excited to share this journey with you. Welcome to himala. 🌿 #LaunchDay #hellohimala #GoodbyeOxolo #NewEra

🔄 The Myth of Multitasking 🔄 We’ve all been there—juggling a dozen tabs, flipping between emails, calls, and to-do lists, thinking we’re being productive. But here’s the truth: multitasking often just adds noise to the day, scattering focus instead of sharpening it. We’ve been rethinking how to solve this challenge and create something that doesn’t just help you manage your tasks but helps you manage your focus. Stay tuned, something new is on the horizon… 🌱 #Focus #ComingSoon #Productivity

🌿 Imagine Life, Seamlessly Managed 🌿 What if you had a human-like AI companion, assistant or co-pilot that truly understood what matters most to you? Picture fewer interruptions, more focus, and a sense of balance that keeps you grounded, no matter how busy life gets. With something exciting on the horizon, we’re thrilled to introduce a solution designed to simplify your day—your way. ✨ Stay tuned for a smarter, more intentional way to manage work and life. ✨ #Teaser #FutureOfWork #AI

🌅 Big Changes Are on the Horizon 🌅 After weeks of intense brainstorming, planning, and more coffee, matcha, energy drinks… than we care to admit, our team is excited to share that a transformation is underway.  But here’s the truth — it hasn’t all been easy nor perfect. This pivot meant making tough decisions, and we even had to say goodbye to some incredible people along the way. It was hard, and it made us pause and reflect. Yet through it all, we turned those challenges into a positive force, driving us to reimagine our path with even more purpose. The past three months have been a journey of growth, grit, and gratitude. We’ve shaped an idea we believe will elevate how we all interact with technology and productivity. 🌱  Stay tuned... Something new is coming, and we can’t wait to introduce you to what’s next. #ComingSoon #NewBeginnings #PivotingForward

𝗜𝘀 𝗮𝗻 𝗔𝗜 𝗴𝗲𝗻𝗲𝗿𝗮𝘁𝗲𝗱 𝗶𝗺𝗮𝗴𝗲 𝗱𝗲𝗰𝗶𝗱𝗶𝗻𝗴 𝘁𝗵𝗲 𝗨.𝗦. 𝗽𝗿𝗲𝘀𝗶𝗱𝗲𝗻𝘁𝗶𝗮𝗹 𝗲𝗹𝗲𝗰𝘁𝗶𝗼𝗻? 🗳️🤷🏼♀️ History is happening right in front of our eyes. Taylor Swift is endorsing Harris and had Trump not posted the AI generated Image of Taylor Swift falsely endorsing him, she probably would not have publicly supported Harris. 📣 "𝘙𝘦𝘤𝘦𝘯𝘵𝘭𝘺 𝘐 𝘸𝘢𝘴 𝘮𝘢𝘥𝘦 𝘢𝘸𝘢𝘳𝘦 𝘵𝘩𝘢𝘵 𝘈𝘭 𝘰𝘧 '𝘮𝘦' 𝘧𝘢𝘭𝘴𝘦𝘭𝘺 𝘦𝘯𝘥𝘰𝘳𝘴𝘪𝘯𝘨 𝘋𝘰𝘯𝘢𝘭𝘥 𝘛𝘳𝘶𝘮𝘱'𝘴 𝘱𝘳𝘦𝘴𝘪𝘥𝘦𝘯𝘵𝘪𝘢𝘭 𝘳𝘶𝘯 𝘸𝘢𝘴 𝘱𝘰𝘴𝘵𝘦𝘥 𝘵𝘰 𝘩𝘪𝘴 𝘴𝘪𝘵𝘦. 𝘐𝘵 𝘳𝘦𝘢𝘭𝘭𝘺 𝘤𝘰𝘯𝘫𝘶𝘳𝘦𝘥 𝘶𝘱 𝘮𝘺 𝘧𝘦𝘢𝘳𝘴 𝘢𝘳𝘰𝘶𝘯𝘥 𝘈𝘭, 𝘢𝘯𝘥 𝘵𝘩𝘦 𝘥𝘢𝘯𝘨𝘦𝘳𝘴 𝘰𝘧 𝘴𝘱𝘳𝘦𝘢𝘥𝘪𝘯𝘨 𝘮𝘪𝘴𝘪𝘯𝘧𝘰𝘳𝘮𝘢𝘵𝘪𝘰𝘯. 𝘐𝘵 𝘣𝘳𝘰𝘶𝘨𝘩𝘵 𝘮𝘦 𝘵𝘰 𝘵𝘩𝘦 𝘤𝘰𝘯𝘤𝘭𝘶𝘴𝘪𝘰𝘯 𝘵𝘩𝘢𝘵 𝘐 𝘯𝘦𝘦𝘥 𝘵𝘰 𝘣𝘦 𝘷𝘦𝘳𝘺 𝘵𝘳𝘢𝘯𝘴𝘱𝘢𝘳𝘦𝘯𝘵 𝘢𝘣𝘰𝘶𝘵 𝘮𝘺 𝘢𝘤𝘵𝘶𝘢𝘭 𝘱𝘭𝘢𝘯𝘴 𝘧𝘰𝘳 𝘵𝘩𝘪𝘴 𝘦𝘭𝘦𝘤𝘵𝘪𝘰𝘯 𝘢𝘴 𝘢 𝘷𝘰𝘵𝘦𝘳. 𝘛𝘩𝘦 𝘴𝘪𝘮𝘱𝘭𝘦𝘴𝘵 𝘸𝘢𝘺 𝘵𝘰 𝘤𝘰𝘮𝘣𝘢𝘵 𝘮𝘪𝘴𝘪𝘯𝘧𝘰𝘳𝘮𝘢𝘵𝘪𝘰𝘯 𝘪𝘴 𝘸𝘪𝘵𝘩 𝘵𝘩𝘦 𝘵𝘳𝘶𝘵𝘩. 𝘐 𝘸𝘪𝘭𝘭 𝘣𝘦 𝘤𝘢𝘴𝘵𝘪𝘯𝘨 𝘮𝘺 𝘷𝘰𝘵𝘦 𝘧𝘰𝘳 𝘒𝘢𝘮𝘢𝘭𝘢 𝘏𝘢𝘳𝘳𝘪𝘴 𝘢𝘯𝘥 𝘛𝘪𝘮 𝘞𝘢𝘭𝘻 𝘪𝘯 𝘵𝘩𝘦 2024 𝘗𝘳𝘦𝘴𝘪𝘥𝘦𝘯𝘵𝘪𝘢𝘭 𝘌𝘭𝘦𝘤𝘵𝘪𝘰𝘯. 𝘐'𝘮 𝘷𝘰𝘵𝘪𝘯𝘨 𝘧𝘰𝘳 @𝘬𝘢𝘮𝘢𝘭𝘢𝘩𝘢𝘳𝘳𝘪𝘴 𝘣𝘦𝘤𝘢𝘶𝘴𝘦 𝘴𝘩𝘦 𝘧𝘪𝘨𝘩𝘵𝘴 𝘧𝘰𝘳 𝘵𝘩𝘦 𝘳𝘪𝘨𝘩𝘵𝘴 𝘢𝘯𝘥 𝘤𝘢𝘶𝘴𝘦𝘴 𝘐 𝘣𝘦𝘭𝘪𝘦𝘷𝘦 𝘯𝘦𝘦𝘥 𝘢 𝘸𝘢𝘳𝘳𝘪𝘰𝘳 𝘵𝘰 𝘤𝘩𝘢𝘮𝘱𝘪𝘰𝘯 𝘵𝘩𝘦𝘮." There is a certain asymmetry in their relationship- Trump clearly likes Swift, she's everything he admires: rich, beautiful, successful and good at business. Yet, he's everything she resents ("those aren't your dad's republicans") but she still would probably not have jeopardized losing fans over politics- it seems like it takes a lot to irritate her- but Trump's post on X got the job done and now she might decide the election. Her predominantly female fan base called Swifties are regarded by journalists as one of the largest, most devoted, and influential fan bases and are very likely to listen to their idol. And female voters will play a pivotal role in determining the outcome of the election. Women comprise the largest group of registered voters, and they tend to turn out at high levels to vote, as 68.4% of them did in the 2020 election. ❓The fact that an AI generated image triggered this cascade of decisions worries me though. What do you think? 𝘚𝘰𝘶𝘳𝘤𝘦𝘴: 𝘛𝘢𝘺𝘭𝘰𝘳 𝘚𝘸𝘪𝘧𝘵 𝘐𝘯𝘴𝘵𝘢𝘨𝘳𝘢𝘮 𝘢𝘯𝘥 𝘟 #uselection #genai #deepfake #catlady Dr. Jan Philipp Burgard

𝗛𝗼𝘄 𝗺𝘂𝗰𝗵 𝗴𝗲𝗼𝗽𝗼𝗹𝗶𝘁𝗶𝗰𝗮𝗹 𝗿𝗲𝘀𝗽𝗼𝗻𝘀𝗶𝗯𝗶𝗹𝗶𝘁𝘆 𝗰𝗮𝗿𝗿𝘆 𝗰𝗼𝗺𝗽𝗮𝗻𝗶𝗲𝘀 𝗱𝗲𝘃𝗲𝗹𝗼𝗽𝗶𝗻𝗴 𝗔𝗜?💄🔫 💰Ilya Sutskever just raised a billion USD to fund his start-up SSI focused on safety. Clearly people are worried about AI being in the wrong hands and misinformation, manipulation, and distortion of truth. This is especially concerning when deployed in political contexts, where the stakes are incredibly high. 🇺🇸 As we approach the next U.S. elections, the role of AI-generated images and deepfake technology becomes even more relevant and with great innovation comes great responsibility. But how much responsibility should be carried by the enterprises? How much should the regulator cover? Where should it be regulated? Do the individuals writing the code carry responsibility? How high should the guardrails be? If all sides have equal access to the same tools is it still a weapon? I don't have answers yet but would love to hear your thoughts. 🥐♥️ 𝘊𝘳𝘦𝘥𝘪𝘵𝘴: 𝘵𝘩𝘦 𝘥𝘰𝘳 𝘣𝘳𝘰𝘵𝘩𝘦𝘳𝘴 via Jan Andresen #AI #Deepfakes #Geopolitics #EthicsInAI #US2024 #ElectionIntegrity

𝗧𝗵𝗶𝘀 𝗶𝘀 𝘄𝗵𝘆 𝘄𝗲 𝗽𝗶𝘃𝗼𝘁 ⬇️ At Oxolo we have been building Gen AI long before it was fashionable. We have built software for people to easily generate videos featuring humans. For e-commerce, for training and HR purposes. We have done it successfully, we’ve raised Series A from top tier VCs, we have over 600k users, have built a high caliber team and grew daily. ❌ However, we decided to discontinue the product oxolo as it is. Why? ❌ ⚖️ 𝗖𝗼𝗺𝗽𝗹𝗶𝗮𝗻𝗰𝗲 𝘄𝗶𝘁𝗵 𝗮 𝗻𝗲𝘄 𝗽𝗶𝗲𝗰𝗲 𝗼𝗳 𝗹𝗲𝗴𝗶𝘀𝗹𝗮𝘁𝗶𝗼𝗻 𝗯𝗲𝗰𝗼𝗺𝗲𝘀 𝗶𝗺𝗽𝗼𝘀𝘀𝗶𝗯𝗹𝗲 The AI Act came into place on August 1. Our avatars very likely fall under the “deepfake category” as they are a “AI manipulated video resembling an existing person”. Providers of AI systems generating synthetic content have to ensure that the outputs of the AI system are marked in a machine-readable format and detectable as artificially generated or manipulated. So imagine marketing videos with labels saying “AI GENERATED” on it. That’s not going to work. Further imagine the legal work behind trying to comply with an obscure piece of regulation. 🐌 𝗦𝗹𝗼𝘄 𝗮𝗱𝗼𝗽𝘁𝗶𝗼𝗻 𝗿𝗮𝘁𝗲 𝗼𝗳 𝗔𝗜 𝘁𝗼𝗼𝗹𝘀 𝗶𝗻 𝗕𝟮𝗕 𝗦𝗮𝗮𝗦 There is a lot of insecurity about how to implement AI and also regulatory uncertainty. It will take a while until this will change. A recent study showed that 77% of employees report that AI has increased their workloads and hampered productivity (link in comments)- this of course will change and is only a matter of time. However, as a start-up we don’t want to wait it out. 🫦 𝗜𝗻 𝘁𝗵𝗲 𝗳𝘂𝘁𝘂𝗿𝗲 𝘃𝗶𝗿𝘁𝘂𝗮𝗹 𝗵𝘂𝗺𝗮𝗻𝘀 𝘄𝗶𝗹𝗹 𝗯𝗲 𝗴𝗲𝗻𝗲𝗿𝗮𝘁𝗲𝗱- 𝗻𝗼𝘁 𝗹𝗶𝗽-𝘀𝘆𝗻𝗰𝗵𝗲𝗱 Even though we believe there is a large market for lip-synched generated content - take one look at the video below (Source: Flux/Leo Kadieff). It’s only a matter of months until hyper realistic humans with the ability to speak will simply be generated. Without IP restrictions and AI Act compliance issues. 🔮 𝗔𝗻𝗱 𝗻𝗼𝘄? We didn’t take this decision lightly but timely. We will now build a product using our existing team and tech - most likely B2C. It’s been hard on us - letting go of people and a product and all the hard work that went into it is not easy. I want to thank everyone for holding up the fort - you guys rock. Thanks to our loyal customers - you were the best and thanks to our investors for your trust. 🚄 Onwards and upwards! 🌊 Heiko Hubertz 👩🏻🚀 Victoria Busse Gülsah Wilke #ai #genai #product #growth

𝗟𝗟𝗠 𝗢𝗹𝘆𝗺𝗽𝗶𝗰𝘀 𝗮𝗻𝗱 𝗠𝗼𝗱𝗲𝗹-𝗠𝗼𝗻𝗼𝗽𝗼𝗹𝘆 𝗮𝗸𝗮 𝘁𝗵𝗲 𝗼𝗽𝗲𝗻 𝘀𝗼𝘂𝗿𝗰𝗲 𝗱𝗲𝗯𝗮𝘁𝗲 🥇🏁 🦙 Meta's Lama 3.1 405B launched and every new model release opens the same discussion: 𝘞𝘩𝘰 𝘸𝘪𝘭𝘭 𝘸𝘪𝘯 𝘵𝘩𝘪𝘴 𝘳𝘢𝘤𝘦 𝘢𝘯𝘥 𝘪𝘯 𝘸𝘩𝘢𝘵 𝘧𝘰𝘳𝘮 𝘢𝘯𝘥 𝘴𝘩𝘢𝘱𝘦? 🅿️ 𝗞𝗲𝘆 𝗣𝗿𝗼𝗯𝗹𝗲𝗺 If we ignorantly assume that the quality of all models will more or less be the same (some feel a little anodyne but that will improve)- from a consumer perspective open source is the better option: ⭕️ 𝗦𝗲𝗿𝘃𝗶𝗻𝗴 𝗲𝗻𝘁𝗲𝗿𝗽𝗿𝗶𝘀𝗲 𝗻𝗲𝗲𝗱𝘀 Enterprises don’t want to depend on models they cannot run and control themselves. Many organizations handle proprietary data they need to secure and can’t send over to closed models through cloud APIs. Other organizations simply don’t trust the closed model providers with their data. 𝘌𝘴𝘱𝘦𝘤𝘪𝘢𝘭𝘭𝘺 𝘪𝘯 𝘎𝘦𝘳𝘮𝘢𝘯𝘺. 🥃 𝗗𝗶𝘀𝘁𝗶𝗹𝗹𝗶𝗻𝗴 𝘀𝗺𝗮𝗹𝗹𝗲𝗿 𝗺𝗼𝗱𝗲𝗹𝘀 Enterprises also like to customize. They can continue training open source with their data and then distill them down to a model of their optimal size – without anyone else seeing your data. 💰 𝗟𝗲𝘀𝘀 𝗲𝘅𝗽𝗲𝗻𝘀𝗶𝘃𝗲 Developers can run inference on open source Models on their infrastructure at roughly half the cost of using closed models, for both user-facing and offline inference tasks. 💡𝗖𝗼𝗻𝗰𝗹𝘂𝗱𝗶𝗻𝗴 From a societal perspective, AI has more potential than any other technology to improve our lives and an open AI ecosystem would allow access to a critical infrastructure and can be amended as needed. It's unlikely AI companies with only closed source will survive. 🐲 𝗖𝗼𝗻𝘀𝗶𝗱𝗲𝗿𝗮𝘁𝗶𝗼𝗻𝘀 However, is releasing the model without releasing the training data set (and code base, data methodology, and the weighting/normalization process etc.) ever open source? The move to release an open source model by a company 𝘄𝗵𝗼𝘀𝗲 𝗰𝗼𝗿𝗲 𝗯𝘂𝘀𝗶𝗻𝗲𝘀𝘀 𝗺𝗼𝗱𝗲𝗹 𝗶𝘀 𝗻𝗼𝘁 𝘀𝗲𝗹𝗹𝗶𝗻𝗴 𝗔𝗜 𝗺𝗼𝗱𝗲𝗹𝘀 -while others are raising billions- is a statement which will be very hard to replicate. Meta has access to the biggest data pools which keep growing - 30% of the global population is using a meta product daily. it will cost Meta 10x to train the next model- Lama 4. It looks like Meta is loading up on capex to release more best-in-class open source models to flatten its competition. With similar data and distribution access, Google will probably be the only one to survive in the long run. So hopefully some competition will remain to prevent model-monopoly. No wonder OpenAI keeps panicky "pre-releasing" a whole variety of products (SORA, SearchGPT) to find streams of revenue they can compete with in the future. 𝙒𝙝𝙖𝙩 𝙙𝙤 𝙮𝙤𝙪 𝙩𝙝𝙞𝙣𝙠? Felix Schlenther Tristan Post Rik van der Woerdt Jan Andresen Oxolo #genai #tech #datascience PS: The prompt for the video was Mark Zuckerberg, Sam Altman and Sundar Pitchai competing in the swimming Olympics on Kling. Not 100% satisfactory.

𝗖𝗵𝗶𝗹𝗱𝗿𝗲𝗻 𝗮𝗻𝗱 𝗔𝗜 👧🏼👾 AI has the potential to completely disrupt how kids will learn and according to UNICEF kids are coming in contact with AI through their devices daily. Screen time and AI are deeply intertwined and woven into the fabric of daily life. 🚨 AI is in every application. Most interactive toys for younger kids depend on AI and depending on their age they might already have a digital footprint: search history, downloads, viewing, listening, AI-powered smart speakers, and the list goes on. But there isn't any comprehensive regulatory framework or even consideration of how AI will affect the well-being of our children. 𝗜𝗳 𝘄𝗲 𝗷𝘂𝘀𝘁 𝗹𝗼𝗼𝗸 𝗮𝘁 𝘀𝗼𝗰𝗶𝗮𝗹 𝗺𝗲𝗱𝗶𝗮 𝗮𝗻𝗱 𝗰𝗵𝗮𝘁𝗯𝗼𝘁𝘀 𝗰𝗵𝗶𝗹𝗱𝗿𝗲𝗻 𝗮𝗿𝗲 𝗮𝗻 𝗲𝗮𝘀𝘆 𝘁𝗮𝗿𝗴𝗲𝘁. 👍🏽 Social media gets smarter and more sophisticated with AI. Systems are being optimized to max out the dopamine release in their growing brains by individualizing content through self-learning systems that understand the mechanics and habits of the viewer. Children's brains are particularly vulnerable as their brains release more Dopamine than grown-ups- which means that even little spikes in the neurotransmitter from a piece of candy or a little “dingding” at Candy Crush feels really really good to them. 💡In a recent study by the University of Edinburgh, children between 6 and 11 were surveyed about smart speakers, essentially chatbots with a voice. 75% thought that the systems might think like humans, most believed the systems were smarter than themselves, and 30% believed that AI could think for itself. 70% said it is wrong to be rude to the devices. Most three to six-year-olds believe that chatbots have thoughts, feelings, and social abilities. 🙌🏻 Kids are easy victims. They share personal information, they trust AI more than they trust humans. We have to educate them early on the perils, we have to educate ourselves (only 1 in 4 parents knows if their child uses gen AI), we have to adapt to “anti-dopamine parenting” and we have to vote politicians in office who intend to set up meaningful regulations. What do you think can we do to make sure AI is not harmful? #ai #genai #coppa #socialmedia  Oxolo Fredrik Harkort Jan Marius Marquardt