Rupert Breheny

"Uncertainty-aware language models require significantly more computation than today's approach, as they must evaluate multiple possible responses and estimate confidence levels. For a system processing millions of queries daily, this translates to dramatically higher operational costs." So, it's cheaper and faster to just make stuff up. What are we even doing here? There are costs to flooding the world with lies and fabrications. This is a business model just stacked with heinous externalities. Read the whole thing. https://lnkd.in/gibuxMmB

186

23 Kommentare

There's a shift happening in how we design software. We're moving from UX to AX (Agentic Experience). Traditional software treats every interaction like meeting a stranger. Open app → click through the interface to do what you want → get result → software forgets everything. Repeat again. This made sense when computers had no intelligence and memory was expensive. Now it's just lazy design. Working on my product taught me that just adding AI to existing interfaces is not the solution. It's about rethinking the relationship between human and software entirely. So I’m designing publyc with the following core principles in mind: > Dialogue over commands. No clicking through menus when you can just say what you want. And publyc responds like a real partner: asks clarifying questions, pushes back on weak ideas, suggests better approaches. The conversation goes both ways. > Memory that compounds. Every interaction adds to a permanent understanding of you - your style, your patterns, your insights. Not session-based memory that resets, but real context that builds over time. The software becomes more valuable with every use. > Invisible intelligence. The best interface is no interface. Every decision publyc removes, every click it eliminates, every time it just knows what you need – that's the product. Trust builds not through features you can see, but through friction that disappears. That's the shift from UX to AX. And once you experience it, everything else feels broken.

137

42 Kommentare

Top UX Research Products of 2025! 🏆 UserTesting – Enabling rapid feedback through real-time video insights and user interviews 🏆 Dovetail – Organizing qualitative data at scale with collaborative, research-friendly workflows 🏆 Maze – Empowering teams to test ideas early and often with rapid, unmoderated testing 🏆 Dscout – Capturing in-context user experiences with powerful diary studies and remote research tools These tools are redefining how product teams understand user behavior. They help leaders validate ideas, uncover deep insights, and build with empathy. See all the 2025 #ProddyAwards winners and discover the tools transforming the future of #productmanagement: https://lnkd.in/gfuvb7Zc

268

7 Kommentare

Gary refactored two functions, color-coded a Notion doc, and now he’s on a panel about engineering excellence. You can build the most elegant system in the world... But if no one knows why it matters? It’ll vanish behind a ticket marked “done” and a PM saying “great job, team!” Leave breadcrumbs. Leave war stories. Leave receipts. → Why you didn’t use that shiny new library. → What that edge case cost you. → How that “simple fix” dodged a meltdown. You’re not trying to look impressive. You’re making sure Gary doesn’t become the hero in your story. If your best work keeps disappearing behind “done,” I built something to help you change that. Start here ↓ https://lnkd.in/gMGfmVas

43

26 Kommentare

I think this post is largely aligned with my view on the impact AI will have on the future of design. The big challenge for designers is that taste is a hard thing to articulate/demonstrate, and it's something most people (including your PM and exec) think they have in spades. https://lnkd.in/eNX4cYae

66

30 Kommentare

Intelligent interfaces make real-time design choices. For designers, sharing design decisions with robots can be… uncomfortable. But delegation ≠ abdication. The new work for designers is to give context and guidance to help the system make good choices. I made a guide, demo, and video for designers (link in the comments) about how to do this and keep the results on the rails. Done right, the result is a radically adaptive experience that responds to user context and intent. Layouts that rearrange themselves. Forms that choose smart defaults. Chat that “speaks” with well chosen GUI elements instead of text. It’s easier and more reliable than you might expect. The guide includes a simple, directional pattern library for giving the LLM its marching orders. For designers, sketching in simple plain-language system prompts becomes part of the design process, at least as important as drawing interfaces in your design tool. Instead of designing every interaction, you’re designing the *physics* of your application’s tiny universe. You define the behavior and constraints for making design decisions in the interface. It’s design system work for real-time decision-making. The basic recipe for wiring interface to intent: 1. Provide a constrained set of UI outputs. 2. Map those outputs to intent (“use this pattern to address that intent”). 3. Ask the LLM to understand intent and choose the right UI or action. It used to be really, really hard for systems to determine user intent from natural language or other cues. Now LLMs just get it. They grasp underlying semantics, they get slang, they can infer from context. LLMs may hallucinate facts, but they’re brilliant at interpreting intent and the shape of the expected response. This makes them a powerful and reliable partner for interpreting user meaning and delivering an appropriate interface. Check out the demo and give it a try yourself. Start writing; the interface is listening. Link in the comments (because you know, LinkedIn).

49

1 Kommentar

It's still too early to predict how the AI-assisted dust will settle, but a very safe bet is closing the traditional gap between static, annotated mockups and functional code. And perhaps eliminating the concept of a "handoff" entirely. Which, by the way, handoffs honestly aren't a far cry from days gone by of waterfall-esque, tossing work over the fence which we've oddly embraced as an acceptable way of working today. Vibe coding, context engineering, or [your preferred term] is poised to close the design-dev gap like we've never seen before. Josh King's comment on one of my recent posts sums this up nicely: "Recently my handoff wasn't a big annotated Figma file, but a github repo and URL." And while this still nods at the need for a handoff, as AI tooling improves we'll probably see designers and developers working in the same repo much earlier in the process than today. Rather than a handoff, a continual handshake instead. Trust me, I've tried to poke holes in this new way of working—not in the (flawed) tools themselves but the overall concept—and I can't see a path forward that's anything other than the industry galvanizing around killing the handoff. And I think that's largely a really good thing. — Josh's full comment: https://lnkd.in/eRq4UGDp #ai #design #vibecoding #contextengineering

87

51 Kommentare