Evan

“HaaS” Hero’s journey as a Service (THE J IS SILENT) @withopine

One 6-phased plan away from glory

@kimmonismus I think we will look back and nostalgically miss the brief period when models needed the current level of handholding

MYTHOS BENCHMARKS, OFFICIAL. HOLY MOLY Anthropic cooked!!

Recently did something similar where I have a frontend that lets the user build “dumps” that can contain any modality, that goes to an ingestion pipeline that scrapes/formats/etc, then an agent goes in and extracts unique evidence and observations (ledger based truths) and upserts/relates those to existing entities and topics into a graph database. Finally a markdown based wiki is auto generated like and auto freshened overtop of the canonical data like a visor. Have a reflection and recall agent tool as well that does agentic search through the graph/evidence/observations

LLM Knowledge Bases Something I'm finding very useful recently: using LLMs to build personal knowledge bases for various topics of research interest. In this way, a large fraction of my recent token throughput is going less into manipulating code, and more into manipulating knowledge (stored as markdown and images). The latest LLMs are quite good at it. So: Data ingest: I index source documents (articles, papers, repos, datasets, images, etc.) into a raw/ directory, then I use an LLM to incrementally "compile" a wiki, which is just a collection of .md files in a directory structure. The wiki includes summaries of all the data in raw/, backlinks, and then it categorizes data into concepts, writes articles for them, and links them all. To convert web articles into .md files I like to use the Obsidian Web Clipper extension, and then I also use a hotkey to download all the related images to local so that my LLM can easily reference them. IDE: I use Obsidian as the IDE "frontend" where I can view the raw data, the the compiled wiki, and the derived visualizations. Important to note that the LLM writes and maintains all of the data of the wiki, I rarely touch it directly. I've played with a few Obsidian plugins to render and view data in other ways (e.g. Marp for slides). Q&A: Where things get interesting is that once your wiki is big enough (e.g. mine on some recent research is ~100 articles and ~400K words), you can ask your LLM agent all kinds of complex questions against the wiki, and it will go off, research the answers, etc. I thought I had to reach for fancy RAG, but the LLM has been pretty good about auto-maintaining index files and brief summaries of all the documents and it reads all the important related data fairly easily at this ~small scale. Output: Instead of getting answers in text/terminal, I like to have it render markdown files for me, or slide shows (Marp format), or matplotlib images, all of which I then view again in Obsidian. You can imagine many other visual output formats depending on the query. Often, I end up "filing" the outputs back into the wiki to enhance it for further queries. So my own explorations and queries always "add up" in the knowledge base. Linting: I've run some LLM "health checks" over the wiki to e.g. find inconsistent data, impute missing data (with web searchers), find interesting connections for new article candidates, etc., to incrementally clean up the wiki and enhance its overall data integrity. The LLMs are quite good at suggesting further questions to ask and look into. Extra tools: I find myself developing additional tools to process the data, e.g. I vibe coded a small and naive search engine over the wiki, which I both use directly (in a web ui), but more often I want to hand it off to an LLM via CLI as a tool for larger queries. Further explorations: As the repo grows, the natural desire is to also think about synthetic data generation + finetuning to have your LLM "know" the data in its weights instead of just context windows. TLDR: raw data from a given number of sources is collected, then compiled by an LLM into a .md wiki, then operated on by various CLIs by the LLM to do Q&A and to incrementally enhance the wiki, and all of it viewable in Obsidian. You rarely ever write or edit the wiki manually, it's the domain of the LLM. I think there is room here for an incredible new product instead of a hacky collection of scripts.

MOBILE DEVS you can now use Claude Dispatch to develop and test your app from your phone! No more setting up simulator infra, building your app, or managing Xcode. Fully remote

@naaiyy_ @0xSero Great choice to fork off Zed

I'm working on this, you should check it out ! github.com/Glass-HQ/Glass… Glass is a fork of Zed, it has a Browser (chromium with CEF under the hood), Full IDE and Terminal. ACP for AI coding, with BYOK with a ton of providers, ability to use your existing subscriptions with Claude Code, Codex, Gemini, Cursor etc. The next release will be big. Huge improvements on UI, fixes, new layouts etc.

Please list all the apps you know of this category that's not already on the list: 1. Cursor Glass 2. Factory Desktop 3. Codex App 4. OpenCode App 5. Claude Desktop 6. CMUX I am looking for this mythic ADE: 1. Has browser 2. Has filesystem 3. Has BYOK 4. Has good ui

@simoarcher Can’t wait to see this in person

Craziest thing about turboquant is that within 3 days of the paper release someone had told an AI how to reproduce it, coded it, distributed it for free and it’s now on my computer

Anyone else speaking to their agent like please audit the fundamental canonical elegance of this API call

@LandseerEnga @expo This is exactly what I’ve been waiting for wow!!!

React Native devs: Your background agent can now hot reload in the cloud. We built a CLI that tunnels your local @expo app to a cloud simulator so every code change shows up live. Claude Code can now: - Edit a component - Watch it render live - Confirm the change No simulator on your laptop. No screenshots. A live dev loop, in the cloud, at scale.

@swmansion @jmeistrich Beautiful work!

react-native-enriched-markdown just leveled up! 🚀 We've added EnrichedMarkdownInput, which means the library is now a fully rich-text suite for React Native. 💪 You get a Telegram-style selection-based toolbar, seamless Input-to-Output sync with EnrichedMarkdownText, bold/italic/underline/strikethrough and links – all with a flexible API for custom UIs.  ✨ Works on iOS, macOS, and Android.  Try the nightly release: npm i react-native-enriched-markdown@nightly P.S. Expect new styles soon – we're just getting started!

@josh_maldonado iOS and Android I sense @expo

my first app is finally live. I've been building with AI tools for a year now and it feels good to finally have a product I'm proud of out in the world. it's called MealScanner: an AI-powered nutrition tool that gives you a full macro breakdown from a photo of your meal. that's it. snap and eat. iOS + Android. links below. go try it and let me what you think!

@josh_maldonado YESSS CONGRATS JOSH!!!!!

It’s wild how important zig and rust are

It’s wild how important pre 2017 books are going to become

@ziusko @coffman_r @jmeistrich This is super cool and I cannot wait to use this wow

And with a huge help from @coffman_r we managed to add a support for AI-chats-like interfaces. Built on top of LegendList by @jmeistrich with a KeyboardChatScrollView component internally and it runs super smooth 🔥

Okay, last improvements before a public release 👀 I added a functionality to KeyboardChatScrollView that changes position of content if input is growing (typical messengers such as Telegram/WhatsApp functionality)

@falcon_ide Big win for nodes everywhere

announcing FALCON GX a new design tool for the curious designers that embrace the creative chaos, the beautiful accidents, wandering, tinkering their way to something great now in private beta falcon.so

GitHub workflow to run a security scan on a @reactnative / @expo codebase on commit, just place it in .github/workflows/rnsec.yml gist.github.com/e-simpson/b8bd…

Vol. 9 on Friday may have been one of the best yet! Toronto’s creative tech scene is growing Emily Switzer laid out philosophical frameworks for craft and taste. Justin chatted aggregators vs generators + open vs closed source in gen-media. We closed with J Lee's Invalid Symbols screenings that had the room fully locked in. Shoutout to @newsystems_ for venue + programming support. Vol. 10 soon. We’re planning something special.

@ziusko Amazing work here!

And a quick update on this: I merged PR and published 1.21.0-beta.3 version to npm that has this component😎 Documentation for new component can be found here: - kirillzyusko.github.io/react-native-k… - kirillzyusko.github.io/react-native-k… Need to fix some other bugs and I will publish stable 1.21.0 soon!

@jmeistrich Appreciate the itemLayourAnimation!

By popular request, Legend List 3.0.0-beta.36 has some fun new features. And it’s faster too! - web: useWindowScroll to virtualize the window - mobile: itemLayoutAnimation to animate item transitions - many fixes and optimizations 🔥 Getting closer to a stable release...