Jan Bačík

@jmeistrich @brotzky I’d love that! I’m doing this today with custom logic, and it’s the most complex part of the chat scroll view in Good Assistant. Having it in LegendList itself would be fantastic ❤️

@Honason @brotzky I'd love to have it! A custom scroll engine would also allow us to pre-render items at the precomputed scroll offsets to reduce blanking during a long scroll. Tracking an issue for that: github.com/LegendApp/lege…

Chat might be one of hardest UIs to build. - Rich text editor - Stream/pause/refresh data - DOM weight - Scroll position - Markdown/Custom UI Easy to build but hard to get right. Just getting this scroll restoration perfect was brutal 👇

@jmeistrich @brotzky Is it possible to control the scroll animation in LegendList? I’d like to use the automatic scroll to bottom feature, but native animated scroll is not as smooth as custom easing. Like the easing in that video ☝️

@brotzky If you want to use LegendList for the chat, I have a solution for this coming tomorrow which should feel perfect.

@Indiveri_ @interaction You might like good-assistant.ai

Poke (@interaction) was pretty cool but for some reason now doesn’t work. Any other ai personal assistant app suggestions?

@ProjetoLabo @dhruvtwt_ @nvidia GLM 5.1 is not on their API. Minimax is, but that’s much smaller model.

@Honason @dhruvtwt_ @nvidia Está enganado, oferecem sim. GLM 5.1 e MiniMax-M2.7 já estão disponíveis.

Why is no one talking about this? @nvidia is offering around 80 AI models via hosted APIs absolutely for free. You get access to MiniMax M2.7, GLM 5.1, Kimi 2.5, DeepSeek 3.2, GPT-OSS-120B, Sarvam-M etc. This plugs straight into OpenClaude, OpenCode, Zed IDE, Hermes agent and even with Cursor IDE. Setup: – Grab API key: build.nvidia.com/models – base_url = "integrate.api.nvidia.com/v1" – api_key = "$NVIDIA_API_KEY" – select model (e.g. minimaxai/minimax-m2.7) If you’re building or experimenting, this is basically free inference. Lock in and start building today anon. Thank me later.

@gabsdv Thank you! I like native tabs, but I also wanted to have this morphing animation between the chat input and other tabs, and that just wasn’t possible…

@Honason This looks really great! Usually I try to not hide expo's navigation but you pulled it off really well here

Custom native tabs in Good Assistant 2 (work in progress). I realise most users won't see a big difference between the default experience, and these fluid tabs that morph into an input in the chat screen, but I love it. And I build the assistant just as much for me as it is for everyone else, so I'm OK with sinking hours into these animations until they feel great, but now they do! It's @expo router's native tabs, but they are hidden & not rendered, with this custom UI using standard navigation to go from tab to tab.

@benjaminakar Pretty!

⌘+K

Can't explain the benchmarks. I like Composer and I often use it for easy tasks, and it's really impressive to make Kimi K2.5 this much better with post training. I see the difference with bigger models like Opus or even Sonnet in situations where Composer "doesn't notice" something that needs to be noticed in order for it to do all necessary changes. For example, when it updates one parameter, it will (possibly) affect logic elsewhere in the file, or other files. That's one example where I see the bigger models succeed and Composer often fail. I think the price feels juuust about right, borderline. At 2/3 of Sonnet's cost, it feels like it's about there with capabilities. And of course, for not challenging tasks, it's great, reliable.

@Honason Would love to see specific examples! Composer 2 performs about as well as Opus and outperforms Sonnet on public evals. If you're not seeing the same results, we would love some examples to help improve the next version.

It’s an excellent model. My personal opinion is it’s near Opus quality (indistinguishable to me for most tasks), but way faster and way cheaper. Enjoy!

@baguvixon @chalaska Nope, but I knew I wanted silent switches. And Nuphy seems to be the only great low profile keyboard that can be ordered with them, so that made the decision. And happy reviewers…

@Honason @chalaska Did you compare to Lofree?

What keyboard purchase don't you regret? Currently on apple keyboard and feeling adventurous.

@FireworksAI_HQ any plans for function calling support for GLM 5.1?

@bensig “compression fits your entire life context into 120 tokens - 30x lossless compression” - how? 120 tokens is couple of sentences. How can you put anywhere near enough context in that?

My friend Milla Jovovich and I spent months creating an AI memory system with Claude. It just posted a perfect score on the standard benchmark - beating every product in the space, free or paid. It's called MemPalace, and it works nothing like anything else out there. Instead of sending your data to a background agent in the cloud, it mines your conversations locally and organizes them into a palace - a structured architecture with wings, halls, and rooms that mirrors how human memory actually works. Here is what that gets you: → Your AI knows who you are before you type a single word - family, projects, preferences, loaded in ~120 tokens → Palace architecture organizes memories by domain and type - not a flat list of facts, a navigable structure → Semantic search across months of conversations finds the answer in position 1 or 2 → AAAK compression fits your entire life context into 120 tokens - 30x lossless compression any LLM reads natively → Contradiction detection catches wrong names, wrong pronouns, wrong ages before you ever see them The benchmarks: 100% recall on LongMemEval — first perfect score ever recorded. 500/500 questions. Every question type at 100%. 92.9% on ConvoMem — more than 2x Mem0's score. 100% on LoCoMo — every multi-hop reasoning category, including temporal inference which stumps most systems. No API key. No cloud. No subscription. One dependency. Runs on your machine. Your memories never leave. MIT License. 100% Open Source. github.com/milla-jovovich…

@colmtuite @base_ui Date and time pickers 🤓

What are some obscure, uncommon, and/or challenging UI components you want @base_ui to ship?

@sayansenguptaa This level of UX would make me pay extra for the software ❤️

Made an animation for clearing context where things goo woooosh ✨

@vanschneider Thanks! Looking forward 🙂

@vanschneider Great! By the way how do you do it? Embedded WebGL on a webview, or do you reimplement it completely with native APIs?

@vanschneider Really nice. Will you show this effect in the app as well, or is this destined to be web-only?

@arunidesign …and focused design. Very nice! Is the page going to be real?

your landing page doesn't need: - bento boxes - feature - benefits - social proof - how it works - pricing - faq your landing page needs storytelling.