George Kyriakopoulos

A very good read with solid punditry about Apples Developer Conference.

5

1 Kommentar

Webby Winners Rewired: Doxy.me's VR Mental Health App Forget flashy websites and dancing cats – this year's Webby Awards saw a revolution brewing beneath the surface. Doxy.me's VR mental health app wasn't just an award winner, it was a paradigm shift. Imagine ditching the sterile office for a calming virtual space – that's the magic doxy.me brings to therapy. Partnering with Lucid Reality Labs, they're not just making therapy more accessible. They're making it better. This is a game-changer for folks who struggle with exposure therapy in traditional settings. Winning Silver at The Webby Awards, among competition from Ford Motors and Universal Pictures, was just the beginning for us. Doxy.me's VR app is a peek into a future where mental health ditches the physical realm and embraces the boundless possibilities of VR. We're talking about a future where therapy is engaging, effective, and finally, accessible to everyone. Buckle up, because mental healthcare is about to get a serious upgrade. #VRMentalHealth #BreakthroughTherapy #FutureofTherapy #TelehealthRevolution #MindfulnessVR #MentalHealthMatters #VRExposureTherapy #AnxietyReliefVR #VRforDepression #DigitalHealth https://bit.ly/3vHIsBS

21

1 Kommentar

Yann LeCun, chief AI scientist at Meta and one of the founding fathers of AI, declares that LLMs that power GAI will never be able to reason like humans. Despite many’s assumptions that it’s only a matter of big models and processing power before the likes of ChatGPT catch up with us humans,  LeCun believes they are fundamentally superficial and limited – and he’s not alone in this. They have ‘very limited understanding of logic . . . do not understand the physical world, do not have persistent memory, cannot reason in any reasonable definition of the term and cannot plan . . . hierarchically’. “It certainly appears to most people as reasoning — but mostly it’s exploiting accumulated knowledge from lots of training data,” LeCun said, but added: “[LLMs] are very useful despite their limitations.” Admittedly, he has an axe to grind as he’s leading a team of 500 to develop a new model of human-level General AI that he calls SuperIntellignece, which takes an approach called ‘world modelling’. GAI is a massive step forward with plenty of runway for further improvements and practical applications, but we can safely put to bed the 2023 talk of God-level AI being around the corner. #GAI #AGI #LLMs https://lnkd.in/eYrr74yR

8

Adobe Researchers have unveiled VideoGigaGAN, a revolutionary model that can upscale videos by up to 8× while preserving intricate details and ensuring smooth temporal consistency. 🎬✨ 🔍 Under the Hood: VideoGigaGAN builds upon the success of GigaGAN, a large-scale image upsampler, and takes it to the next level by addressing the challenges faced by previous video super-resolution methods. 💪 The team has implemented several innovative techniques, such as: 🧩 Temporal Attention Layers: These special layers help the AI pay attention to how the video changes over time, ensuring that the upscaled video looks consistent from frame to frame. 🌊 Flow-Guided Propagation: This feature allows the AI to better understand the motion in the video, resulting in smoother and more accurate upscaling. 🎨 Anti-Aliasing Blocks: Just like how a painter blends colors to create a smooth picture, these blocks help the AI blend pixels together to avoid jagged edges in the upscaled video. 🎢 Direct Shuttling of High-Frequency Features: This technique helps preserve the fine details in the video, making sure that the upscaled version looks sharp and clear. These improvements enable VideoGigaGAN to generate videos with stunning high-frequency details and impressive temporal consistency. We are almost at 'Zoom and Enhance' kind of upscaling levels. » https://lnkd.in/dEshEGh2 This won't be open-sourced. I'm sure of it. Please, Adobe, prove me wrong. Source: https://lnkd.in/dB3K_CyX ↓ ☝️ Follow for more AI-News » https://lnkd.in/esb3PH9W ✌️ Check out our AI Workshops and supercharge your AI Game (so far 90+ successful individual workshops done for companies and corporations) https://lnkd.in/dXGDz58E" 🤘 Test our Gen AI platform creamAI: http://www.creamai.de #ai #genai #videoupscaling #machinelearning #videoproduction #upscale

293

31 Kommentare

#figma disabled it’s ‘make design’ feature released earlier this month at #config Figma CEO Dylan Field says the company will temporarily disable its “Make Design” AI feature that was said to be ripping off the designs of Apple’s own Weather app. Interesting this- AI generated images and Ui can’t ever be ‘original’ simply because it is pulling data from a range of sources to create an ‘original copy’. So whatever will be generated using this feature will by default be trained on existing components. The only thing that can be ensured is that the model generates greater variables creating only a perception of the outputs not being a direct ‘rip off’. Basically it’s always going a rip off - directly (in this case) or indirectly. Thoughts? #design #figma #ai

5

2 Kommentare

Bill Gates said: ​"This is such a cool book. The range of Jono's knowledge is astounding, and so is his ability to digest complex ideas into deceptively simple drawings. You'll learn something on every page―and be entertained too." I am excited to talk with Jono Hey (Product & Design leader), creator of the very popular Sketchplanations, and now author of the new book "Big Ideas, Little Pictures: Explaining the world one sketch at a time". We will: ​​1) Dissect specific pictures from the book. ​2) Discuss how Jono decides what to visualize. ​3) Explore Jono's creative process. ​​​4) Take live audience Q&A. Live attendance is for Level Up Newsletter & Community paid members. The video recording will be later shared via our newsletter to all subscribers. Members, RSVP here: https://lu.ma/2ta6xi3j

7

4 Kommentare

Luma Labs: New AI Text to Video 👇👇👇👇 You can try it free. OpenAI definitely has some competitors in the text to video space. AI video is rapidly improving. My prediction is by the end of this year over 50% of content being produced will use some type of AI tools. #ai #aivideo #luma https://lnkd.in/eKB3nHRw

3

🚨 𝗔 𝗠𝗮𝗷𝗼𝗿 𝗦𝗵𝗶𝗳𝘁 𝗮𝘁 𝗢𝗽𝗲𝗻𝗔𝗜: 𝗧𝗵𝗲 𝗗𝗲𝗽𝗮𝗿𝘁𝘂𝗿𝗲 𝗼𝗳 𝗜𝗹𝘆𝗮 𝗦𝘂𝘁𝘀𝗸𝗲𝘃𝗲𝗿 🚨 In a significant turn of events, OpenAI co-founder and 𝘊𝘩𝘪𝘦𝘧 𝘚𝘤𝘪𝘦𝘯𝘵𝘪𝘴𝘵 𝘐𝘭𝘺𝘢 𝘚𝘶𝘵𝘴𝘬𝘦𝘷𝘦𝘳 has left the company. This marks the end of an era for OpenAI, a company that has been at the forefront of the AI boom. The departure of Sutskever, a staunch advocate for safe AI, underscores a growing tension within the organization. The power struggle between Sutskever's vision of responsible AI governance and Sam Altman's focus on commercial success and market competition came to a head in autumn 2023. With tensions now resolved in favor of commercial interests, Sutskever's exit serves as the 𝘭𝘢𝘴𝘵 𝘯𝘢𝘪𝘭 𝘪𝘯 𝘵𝘩𝘦 𝘤𝘰𝘧𝘧𝘪𝘯 𝘧𝘰𝘳 𝘖𝘱𝘦𝘯𝘈𝘐'𝘴 𝘰𝘳𝘪𝘨𝘪𝘯𝘢𝘭 𝘮𝘪𝘴𝘴𝘪𝘰𝘯. Altman, who once promised to self-govern AI, now leads a company where the lone voice for strict self-regulation has exited. Sutskever's departure raises questions about the future direction of OpenAI. With no prominent advocate left for the initial non-profit mission, there's a palpable vacuum. Will OpenAI maintain its commitment to ethical AI, or will commercial interests dominate? This is an 𝘪𝘮𝘱𝘰𝘳𝘵𝘢𝘯𝘵 𝘮𝘰𝘮𝘦𝘯𝘵 𝘧𝘰𝘳 𝘵𝘩𝘦 𝘈𝘐 𝘪𝘯𝘥𝘶𝘴𝘵𝘳𝘺, reflecting broader challenges in balancing innovation with responsibility. As we witness this shift, it's crucial for industry peers to consider the implications for the future of AI development. What are your thoughts on this shift at OpenAI? How do you think it will impact the AI landscape? 💬 https://lnkd.in/eW_UQNDu #ArtificialIntelligence #Ethics #AI #OpenAI #

7

5 Kommentare

Deadlines - An Important Aspect of Any Delivery Day 5 of 60 - PM Roadmap by Carl Vellotti Study Material Link: https://lnkd.in/ezKfseWA Creator: https://lnkd.in/eBgmqDYC Q1. How can deadlines help teams combat human nature and ensure progress towards project goals? A1. For me, deadlines are important just as to do lists are important. And the simple fact is most folks don’t understand the real reason as to why they should have both. It’s the process behind creating them - not the end result. Good deadlines can be used effectively to direct efforts with a goal in mind. Bad deadlines can be used to punish the team and defeat morale. In my opinion, when it comes to project goals, deadlines should loosely be related to the unit of work done (somewhat agile methodology). As long as the work is being done and delivered roughly around where we expect - deadlines are working. The reality is, time, money, and effort go hand in hand - defining all three is an important aspect towards progress. Q2. Do you agree that deadlines should be used even when the date is arbitrary? Are there any potential negative consequences to using deadlines in this way? A2. Yes and No - making the date arbitrary and making it obvious will ensure the team never takes deadlines seriously. On the other hand punishing a team for missing deadlines can also lead to morale and issues with estimation. A lot of team get in the habit of falsifying estimations to avoid backlash of missed deadlines. Deadlines should be flexible, BUT that knowledge should be closely guarded between a select few. If deadlines need to be moved, an obvious and believable reason needs to be provided. And when the deadlines are met, as the article suggests, there should be praise for the team that actually did within the time they set. It's a complex matter, as with most constraints, deadlines too can empower your progress or derail it. Q3. How can you use the insights from this article to effectively implement deadlines in your own product management process? What steps can you take to ensure that deadlines are used in service of good products, rather than in priority to them? A3. The summary at the end does a good job of how deadlines should be treated. The biggest lesson to learn is that deadlines are by nature aspirational a lot of the time. Unless there are external factors - like a hard event date, anything internal can always be delayed given good enough reasons. To ensure that deadlines are used as an enabler for good product development, I believe they need to be created with balance in mind. Too many or too little deadlines often make no impact. But having the right amount, tied especially to milestones of the roadmap, can be an effective way to use deadlines to communicate prioritization with good direction. _______________________ Context behind this post: https://lnkd.in/e7ZQefjS

2

UX designers design for tomorrow, but we needa think responsibly from today. Several thoughts after reading Professor Philip Alston and Professor Joanna Redden's articles ✨ : - In 2019, concerns were raised about automated decision-making systems (ADS) leading to potential injustices and inequalities. In 2024, these concerns persist as the use of AI in the public sector continues to raise questions about accountability and transparency. - The challenges in AI's explainability make it difficult to hold AI-powered decisions accountable, highlighting the need for human oversight in these systems. - Human intelligence enables us to understand and adapt to real-time information, including reading body language and nuances in communication, aspects that machines are not yet capable of until we reach a state of AGI/ASI. Thus, human oversight remains essential in ensuring fair and just outcomes in AI-powered systems. How is it related to UX? As the impact of AI technology expands, it's really important to think carefully when we design products. We need to look at the bigger picture, considering not just how the user flow works from start to finish, but also thinking about any potential harm or problems it might cause. This helps us create products that are safe and ethical for people to use. #ResponsibleAI #EthicalAI #UXDesign #ArtificialIntelligence https://lnkd.in/eAjakeVG

7

#OpenMetaverse: Meta's Vision vs. Apple's #WalledGarden The Open Metaverse is not a new concept; rather, it's an evolving vision of a digital landscape where connectivity knows no boundaries. The Metaverse, as conceptualised by Matthew Ball, "represents a massive, interoperable network of real-time rendered 3D virtual worlds, which can be persistently experienced by a theoretically unlimited number of users." This vision aligns with the principles of an open and connected digital ecosystem, promoting broad and seamless integration across different platforms and digital environments (Read more below in the first comment ). Mark Zuckerberg's vision for Meta closely adheres to this concept of the Metaverse. Meta's approach is centred on creating an immersive, "embodied internet" where users are not just observers but participants within the digital space. This initiative aims to foster a platform where various forms of digital content and social interactions can occur within a unified, continuous digital space, suggesting a more open, integrative approach that allows for greater user agency and content portability across different virtual environments In stark contrast, Apple's approach has often been described as a "walled garden," a strategy that typically emphasises user experience and ecosystem control over openness. Apple's ecosystem is designed to keep users within its own suite of services and hardware, which, while providing an integrated and streamlined user experience, limits interoperability with systems outside its control. This approach could potentially conflict with the Metaverse ideal of a vast, interoperable network, as it imposes restrictions on the extent to which Apple users can interact with or migrate to other ecosystems. Thus, the fundamental divergence between Meta's and Apple's approaches to digital ecosystems mirrors broader debates in the tech industry about open systems versus controlled environments. Meta's vision, aligned with Ball's definition, advocates for a Metaverse that is as inclusive and expansive as the internet itself, potentially leading to a more interconnected and less siloed digital future. Apple's more insular strategy may provide a seamless user experience but at the cost of broader compatibility and openness #OpenMetaverse #DigitalEcosystems #MetaVision #AppleVsMeta #TechDebates #Interoperability #DigitalFuture https://lnkd.in/eazF-uv3

2

1 Kommentar

Design Principles for Generative AI Applications A team of researchers from IBM has published six valuable design principles for designing visual generative AI tools. For each principle, they have selected several implementation strategies and successful examples. Principles: - Design responsibly. Design for the user by understanding their needs and pain points rather than prioritising the technology or its capabilities - Design with mental models in mind. Generative AI presents new challenges for users, so it's important to communicate to users through mental models how the system works and how individual actions affect it. - Calibrate user trust. It is important that people check AI outputs for inaccuracies or unacceptable results and treat them critically. Be honest about the limitations and weaknesses of the system. - Take into account the variability of generative AI results. Help users select what they need or adjust a query - Provide opportunities for human and AI co-creation. Add more controls during generation so the user can enhance the results and create exactly what they need - Take into account the imperfections of the AI. Help users improve the result of generation, point out errors and collect feedback. https://lnkd.in/eDZ7V8js #design #ai #principles

34

2 Kommentare

Last fall, rumors circulated that OpenAI’s Sam Altman was partnering with legendary Apple designer, Jony Ive, on a new device for the age of #AI. The duo was in the early stages of sketching out a device that would, according to The New York Times, “succeed the smartphone and deliver the benefits of AI in a new form factor, unconstrained by the rectangular screen that has been the dominant computing tool of the past decade.”   Eight months on, it appears rectangular screens will hold sway a bit longer. In a month rife with AI announcements from the biggest names in tech, #smartphones, #laptops, and #tablets were the focus. For now, it seems, we’ll need little else to get our AI fix. It makes sense. These devices have advanced sensors, chips, internet connectivity—and nearly everyone has one. You can tack on the latest AI with a simple download or update.   “Just as #chatbots have promised to condense the internet into a single program, #generativeAI now promises to condense all of a smartphone’s functions into a single app, and to add a whole host of new ones,” The Atlantic’s Matteo Wong wrote.   But it’s not just smartphones (or tablets). The humble PC will get an AI makeover too. This month, OpenAI, Google, and Microsoft stumbled over one another to remake how we interact with the devices and digital services already used by billions. Even Apple, long quiet on the AI front, is rumored to be close to unveiling its own AI vision.

13

RECKLESS. This is the word I would use for the current Wild West of AI. So many grifters are attaching their wagon to something they fundamentally do not understand. Thank you for your work Charles! Human factors experts are the key to unlocking how technology can work in concert with humans. My prediction, human centered designers with deep respect for anthropology will rise to the occasion and correct the ship. The sentiment from the Humane founder of “We just want to build”, is reckless and self serving. It lacks the WHY. #ai is powerful and important, but please do your homework before going for that cash grab, it will bite you in the end, just like Humane. Wake up and let’s goooo! Want a real team on the right side of humanity, one who asks the why’s not just the how’s? Reach out to us at Nova, we can bring your dream to reality. #design #leadership #humanfactors

13

5 Kommentare

Mega prompt eh? All of my prompts are this long and detailed - maybe this explains why I get such good results with LLMs.

3

**DECISION DESIGN – Setting the Opportunity Space (#GG 2)** I’ve mapped decision design’s opportunity space based on the two main design parameters to enable broader discussions on the topic. This exercise is a starting point to nuance arguments around what decision design can become. Design Parameter I – FUTURES EXPOSURE: The image below shows the x-axis representing a decision-maker's possible exposure to a certain future, ranging from dystopian to utopian. The decision-maker can be an individual, a group, an organization, or any system with agency. While the spectrum of possible futures is infinite, the focus here is on plausible futures. The two normal distributions (bell curves) represent two settings: GREY: The probability of exposure in today's setting without any decision and intervention by the decision-maker, likely skewed towards dystopia by current systemic trajectories. WHITE: A setting where the designed decision flattens the curve towards a more positive futures exposure for the decision-maker (note: not necessarily for the world at large). Design Parameter II – FUTURES DESIRABILITY: The y-axis reflects the decision-maker’s attitude towards a certain future. Desirability ranges from low (dystopian) to high (utopian). Human perception is mapped with two logistic growth curves (S-curves) derived from the science of risk perception and Prospect Theory (Slovic, 1987; Kahneman & Tversky, 1979): DOTTED LINE: Without any intervention, the probability of a more dystopian future is high, and desirability at the distribution’s mean is considerably low. SOLID LINE: With decision design applied, desirability of basically any possible future can be significantly increased by, for example, shaping robust core narratives, resilient identity design, and transformational hero journey framings. THE ROLE OF DECISION DESIGN: Discover, define, develop, and deliver decision setups, guidelines, and hacks to actively shape and co-create the decision-maker’s futures exposure and futures desirability. For most decisions a reasonable probability-desirability equilibrium is the sweet spot of decision-becoming. However, design towards a more utopian individual future is possible too. At Gravity & Grandeur, the world’s first decision design studio, we aim to amplify our client’s opportunity space by systematically steer exposure towards antifragility and by elevating desirability into full-spectrum futures resonance. We are the pioneers of multi-parametric decision design. Our mission: BECOME YOUR BEST DECISION --- How to become your best decision: Check out the Gravity & Grandeur website and book a free GG Check for your next important decision (link in comments). Also, don't forget to follow us on LinkedIn for more insights on decision design and decision-becoming.

17

17 Kommentare

Hello World: I'm swinging for the fences on this one; digging into past Human Factors thinking to form a better choice framework for integrating AI. If you're at all interesting in AI, society, psychology and human meaning I'd suggest taking a look. 🤖 Our rapid decisions about the place of AI in our society feel like foreclosures on the inevitable fate of humanity; abandoning human labour for AI. After all, these robotic tools are faster, tireless, and unrelenting in the face of uncertainty. 📣 But neither unconditional advocacy or complete avoidance of AI are the right approach. We need better frameworks to understand when and how AI should support human behaviour. ⚙ To find a better model I’ve dug into thirty-year-old theories of automation, explored the psychology of supervisory control, considered automation accidents and tried to find the essence of what it means to make choices about automation. Join me in this first part of a multi-part exploration of making better and more rational choices about AI automation. https://lnkd.in/gq7PnGkb #HCD #humanetechnology #peoplethentechnology #AI #automation #humanfactors

9

9 Kommentare

The Institute of Design's 2024 Report "Taking Responsibility in the Age of AI" provides valuable insights into the current state of the design community. As we navigate through 2024, these insights are only accelerating. https://lnkd.in/gzbpSstt Key takeaways: 1. Focus on value creation: Design should emphasize discovering meaningful customer outcomes and demonstrating how our work drives business success. This is a challenge for entire product teams, not just designers. 2. Be at the origin point: Design needs to be part of defining success and product strategy from the inception. Our ability to move faster with prototyping platforms allows us to influence roadmaps and inform thinking quickly. 3. Drive organizational transformation: Designers can contribute to company success by partnering with engineering to reduce complexity, collaborating with PM to prioritize high-impact outcomes, and supporting customer success teams in improving user onboarding. 4. Measure impact in subscription-based products: Design research can work backwards from subscription renewals to understand successful and less successful customer cohorts and the product experience impacts. 5. Address AI within the context of design: Designers working on AI-empowered products can help organizations understand user trust, adoption, context, and appropriate implementation. Simply bolting on AI is not enough. To improve the perceived value of designers, we must better describe the impact we make and help our cross-functional colleagues understand the value we create. #instituteofdesign #design #designleadership p.s. I used an AI writing assistant to help me refine and condense my thoughts for this post.

35

2 Kommentare