#139 AI Updates for Designers

Hello and welcome back to the future of UX or better. Welcome to the future of AI. My name is Patricia, I'm your host and in this podcast we talk about how design, technology and AI are reshaping the way we build products, experiences and even our roles as designers. In the last episode we looked at the big trends and today we are doing something a little more practical. We talk about the actual updates because to become completely honest I think that nobody has the time to read every release note, follow every announcement, test every new tool and then keep up with five different platforms changing weekly. I did that for you and in this episode I'm sharing, yeah, five AI updates from the last quarter that I think designers should know right now and for each one I will break down what changed, why it matters and where it shows up in real world and what you should take away as a designer. But before we jump in, one quick announcement. I want to quickly share something that might be interesting if you want to get a little bit more hands on with AI. Next Monday, today is the first day when the episode goes live. So next Monday my free AI briefing challenge starts. You can still sign up, it's completely for free and from Monday to Friday you will receive one email per day. Each email walks you through a concrete step of creating your first AI project brief and the goal is super simple. You will learn how to think about AI problems from a product and your ex perspective, how to structure in AI idea, define constraints and use case and then turn a vague idea into a clear project, briefing that you can work with. Because I think that yeah, experimentation, trying things out and really getting your hands on AI is super important. This is why I created this challenge. As I mentioned, it's completely for free. You can sign up, you can find everything that you need to know in the show notes, in the description box, you get five emails, all for free and it hopefully helps you to kick start this year the right way. So this sounds interesting, you will find the link to sign up in the info box. Okay, now let's get into today's episode. All right, so as I said, I'm going to talk about for me the most interesting AI updates, the five most interesting AI updates. And I would say let's get started with the first one, which is Claude, Claude code and co-working with agents and yeah, Claude code in general. Claude has been pretty popular for a while. It's a large language model for writing, for doing research and coding. But the recent shift, I think is a little bigger than yeah, the model just got better because I'm seeing a real shift in Claude is being positioned more like a coworker. You can give work to another chat that you ask questions in. So first of all, what changed? There are two things that I find personally pretty interesting. The first is the agent style workflow feature. So planning before acting, reusable skills, structured execution, they announced a co-working skill basically where Claude can access to your tools, your apps, your documents, and then basically you chat with the prompts and the agent acts on your behalf. I think this is the first thing that I find pretty exciting. It's new, so it's still not available to everyone, but I will link it in the show notes in case you want to check it out. I think this is super interesting and a huge shift. The second thing that I find fascinating and I think many probably will agree is Claude. Claude is a more, I would say, a serious way to collaborate on code and projects, especially if you want the AI to understand your repo, follow conventions and make changes in a controlled way. I personally use Claude now for my projects either with terminal or with cursor and I think it's wonderful. Claude code is absolutely fascinating. The capabilities are fascinating and it's pretty easy to get started. So I would highly recommend actually to get your hands on it. So why does this matter for designers? Because this changes what your role looks like when you build AI. You spend less time doing all the micro prompts, write this button copy, now rewrite this, now rewrite again and more time doing really setting the constraint and let the agents handle this, reviewing the plan, making judgment calls and shaping maybe output quality. And that's a very designer mode, I would say. So what I'm definitely seeing is that I'm tropic with Claude is moving towards this co-worker mode, co-working mode, planning, other usable skills, adding different agents, prompting your own agents, creating your own multi-agent system. This is pretty strong actually, what Claude is capable of. So I think this is, I think for me, the biggest and most interesting update, where I would recommend you to check out. The second really interesting update and I would also really love to hear your thoughts on it. Are the open AI health or the open AI health features? Yes, open AI health. This is completely new, it's, I would say it's pretty important because healthcare is one of the strictest environments where I am. And if AI can work here responsibly, it's a signal that AI products are growing up. First of all, what is open AI's health? It's a consumer-facing health experience in chat GPT, so it's health-oriented workflows, and it can also be used for enterprise healthcare usage. And the product is pretty interesting, you basically share your health data with chat GPT, and then it gives you feedback on it. And for me, this is as many updates in AI on the one hand fascinating and super interesting. I think health, healthcare, living more sustainable, becoming fit, living a better life, longevity, are future topics that a lot of people care about and should care about. So I think there is a huge market. On the other hand, I think this is truly dangerous to share your data. There are many cases, and also when we have a look at the legal situation, we don't know what happens with the data that you share with open AI. There is just basically one law change away from sharing this data with other companies, and the things can maybe it's used against you at some point. That would be very cautious when it comes to sharing your health information, or I will say like in general sharing very personal information with a large language model, because everything that gets in there can be, you can get it out. This is the way how these large language models work. So I think fascinating, super interesting project makes sense to have a look and see also like seeing the direction in which the company is moving. But on the other hand, I think for us as designers, something that we need to be very cautious about, a lot of private, very sensitive data, and we don't really know if this is safe. So I like from your perspective, this again shows how important it is to rethink ethical topics and privacy and yeah, everything like that. Topping number three, which I also personally find pretty interesting. We are seeing a lot of updates. We talked about updates from open AI, we talked about Claude, and also Google is not sleeping. I think it's pretty impressive what they presented in the last six months pretty much, especially with Gemini. So if yeah, Claude is co-worker with structured tasks, Google is really pushing Gemini towards multimodal builder and the visual explainer, especially with all their creative tools. We have multi-model everywhere. So Gemini is really built to handle text, images, video, audio, and code. We have generative UI. I talked about everything two, three episodes ago. As an interesting experiment that I'm working on, as well instead of paragraph and having fixed layout, they have these moving layouts and modules and interactive pieces. And they have a lot of prompt to app workflows. So tools like the AI studio with Opal where you can create your own agents and other Google experiments make it easier to turn prompts into working UI. And they just announced in YouTube, unfortunately, I forgot the name where you can basically create different visuals based on examples. So a little bit like Nanobanana, but on the mirror board pretty much. So fascinating, they are definitely going in the direction where you find feel as a pure graphic designer, you probably can do a lot of things that a non-bographic designer could actually do. And I think this is pretty interesting, are moving away from just like being a search tool to being an interactive studio suite where you have all the different tools that help you via prompt with anything that you can think of building your own agents and creating your own visuals with nano-banana, which is a wonderful and pretty fascinating image generation tool, which is still better than what you can find in TattuPT. It's not as good as mid-roney, but still I think super super interesting to work with and again you can try it for free. So what we are seeing with Google and also with Gemini 3, that it's pushing AI towards a visual and interactive output and fast prototype creation, which means that designers need to think in systems like consistency and control and transparency. I think for me, super interesting also to integrate these tools in my workflow more and more. And for the next point number four, we are moving again to TattuPT to open AI. Maybe some of you have heard about the news, but TattuPT allows apps. So basically AI as a platform inside your tools. This is yeah, this is all about AI becoming a layer basically across your workflow and not the destination that you visit. What changed is that TattuPT can integrate with tools like Slack, with Figma, with Kanwa, with Photoshop, with Adobe Acrobat, with Figma and others. It can work with your files, it can generate artifacts, it can pull context from connected services inside one conversation. So how does that look? For example, a mountain that you would like to, I don't know, your chatting with TattuPT, your thinking about different features, that you would like to build for the product that you're working on, thinking about the prioritize them and then you think, okay, let's go deep into one feature. Like what, like how could the flow look like and then you say, okay, add Figma, please also create the flow chart. And then you say, oh, wonderful. And now at Acrobat, or Adobe Acrobat, now create a PDF and share that with my boss. So you see, you can use the different tools, which makes it so much easier also for people who don't know how to use the tools to use the tools. Thinking about tools like Figma, especially, yeah. And for me, I think what's most interesting here is that I think one of the biggest productivity killers is context switching. And if you are, I can read the context, summarize the right info, draft right thing, and then hand you an artifact in the tool that you already use. That's a real workflow improvement. So you don't switch to Figma and then to another tool. You're staying in, in the TattuPT window. That's wonderful for OpenAI as well. But I think also for us and our productivity, I think it's pretty interesting that we are not switching between all the different tools and all the different windows and all the different tabs that we have open. I think like also UX wise, I'm seeing some problems because for people, it's a bit unclear of which tools they have access to, but these tools actually can do. So I think there's something you ask for is that OpenAI probably needs to solve in the pretty near future. And also, yeah, a problem could also be like over automation. Graphs need to review and actions need confirmation. I think it's super important. So to summarize, TattuPT apps turn AI into an integration layer across the tools, which makes transparency, mission UX and user control the real design of the battleground. Now let's move to our last update. And yeah, I think that we have all seen the CS 2026 updates. And I think this is a perfect way to close because this is one that really ties everything together. If I had to sum up CSE as this year in one sentence, it will probably be that AI is no longer something that you open, like a window or tool, but it's something that's already there everywhere. The CS is one of the biggest consumer fairs basically in the world. It happens every year in Las Vegas and it just happened last week. So the CS has always been a consumer tech show, but consumer tech usually becomes enterprise tech a few years later. And this year the message was really clear, at least what I have seen and heard AI is no longer bolted onto product as a feature, but it's embedded into devices and environments and workflows as this permanent background layer that's everywhere and that's coming with you, or something that you click or something that you activate, but something that's just like running all the time. And the big themes that matter for designers is first of all that AI as an ambient coordination layer. One of the strongest signals came from companies like Samsung. They showed AI not chatbot, but as a system that quietly coordinates your home, your devices, your routines. So basically we are working with smart homes. We have connected all our devices. We have Alexa. We have, I don't know, different light bulbs. It's all smart home, but now things are finally coming together with a smart AI assistant that helps us to remake sense of the whole smart home thing. And the same thing applies to the work environment. We have AI coordination schedules, prioritizing tasks, syncing tools, and also optimizing systems without explicit commands. And the design implications are that we are no longer designing the interfaces for commands, but that we are actually designing a relationship with systems that act on their own. And I think this is a huge shift. We are also seeing, I think when AI got, or in the beginning of the AI hype, everyone build a co-pilot, like a mini chat interface, basically where you can prompt something and then the co-pilot to work for you. And now we are seeing that the co-pilot disappears a little bit and the autopilot appears. Let's talk a little bit about why this is exciting, but also a little bit scary. So a lot of the language at CS moves from assistive AI to autonomous AI. So robotics in the clearest example is that the Boston Dynamics showed Khuman-Rit robots designed to work continuously alongside humans. I'm also super curious to hear your thoughts about robots and robotics, especially helping you in the household. Here is to hear what you think. They're not waiting for commands, not executing scripts, but they are operating in shared environments on their own. And if you're not designing robots, the pattern applies to software as well. So the AI systems anticipate the needs, they trigger actions, and then they operate in the background. So if AI acts without asking first, consent, visibility, and override become core UX problems. You users need to know what the system is doing, why it's doing it, and how to stop the correct it. Another really big theme was edge AI. Companies like ARM, Samsung, Dell, showed a hardware design to run AI locally. Or faster responses for lower latency and also privacy. And that matters because always on AI can't depend entirely on the cloud. This means more real-time interactions and more contextual systems and bigger expectations, rounded privacy and trust. So to sum this up, what is important for us as designs, there are a million different tools and gadgets out there. AI is everywhere. We need to think of how to integrate AI, not as a co-pilot, not as a chatbot, but to have this layer basically across everything that's around the product. And I think this is an interesting challenge and also opens new questions when it comes to privacy. I think this is super important. One product I already talked about it in one of the last episodes. The new AIR glasses that build a memory out of everything they see. Sounds, again, pretty fascinating. But on the other hand, I think it opens the way of living where we don't want to live, basically. Having a device that knows everything about you, with your dog's food or your kid's favorite toy and what's your favorite coffee. So a lot of information that a very person on private privacy is such an important topic. Data is the new oil. Every data is about fuels AI and AI knows so much about the user, about us, about me, about you. That this is something that really scares me thinking about the future, thinking ahead, and something that we need to be very cautious about. Also, not only what we personally share, but also how we treat the data or the private information from our users, we're using the product that we are building. Okay, so let's zoom out for a quick second. I would say across all five updates, if you patterns, keep, AI is becoming more agentic. Interfaces are becoming more visual and more dynamic and designers are moving closer to building real working system and not just the screens. I would say that it's about understanding how all role changes when AI becomes the infrastructure. And in case you want to dive a little bit deeper on AI and start the year with a bang, you you are welcome to join AI for a designer the five week bootcamp. cohort number five is about to start. Enrollment opens on January 26. So yeah, an almost one. It's only open for a few days before we start with the cohort into five weeks. It's an online bootcamp focused on AI from a design perspective. Each week includes a live workshop, an assignment and across the program. We build an AI related project step by step with things that you learn with your assignments, something that you can put in your portfolio later on. And also we have the pre-recorded videos in the community where you can ask questions, share, work and learn together. I think this is one of the most underrated things when it comes to learning and growing. It's community learning together in group. I think like-minded people who push you having an opportunity to ask questions. This is I think something that's truly magical and super important because you can watch YouTube videos, you can read an article. But what you can do is you can share questions, get detailed feedback on it, really have an exchange of others and learn. Basically. So if you want to learn AI in a structured way, check out the info box for the details you can sign up for the waiting list. Then you will get notified when enrollment opens. The spots are limited and when they, yeah, when they're, when everything falls and it's full, so they tend to fill up super fast. Yes. All right. Let's, yeah, I would say thank you so much for listening and yeah, I will see you next week and thank you so much for listening. Yeah, you in the future. [Music]