As a fun Saturday vibe code project and following up on this tweet earlier, I hacked up an **llm-council** web app. It looks exactly like ChatGPT except each user query is 1) dispatched to multiple models on your council using OpenRouter, e.g. currently: "openai/gpt-5.1", "goog…↗

I recorded a 37-min tutorial on how to prompt with Gemini 3 to get the best UIs https://t.co/LEhYqdXnNI↗

We put together a prompting guide for Claude Opus 4.5 based on extensive internal testing by our research and applied AI teams. Here's what we've learned so far about getting the best results:↗

Okay, as a designer, I'm over the moon with Gemini 3. You can create far better results with the right prompting and references. Let's get into the details. So I started underwhelmed. A non-descriptive "create a landing page" will give you generic purple results. Barely better…↗

Claude doesn't do image generation. Except it does—just add a skill that connects Nano Banana Pro to your workspace. https://t.co/s8eUsvAl1j You're welcome.↗

I built a @claudeai skill that it takes any input and converts it into different kinds of art for my site using Nanobanana 3.0. - Blog header art - Tech - Comics Available for free in our public Personal AI repo!↗

Take a peek if you’re a Cursor fan and want to do more with us. We’re intentionally keeping this program small, but doing our best to bring on new members! https://t.co/lGaz8KUK9v↗

Cursor just shipped the feature I've been harassing them about for weeks. Worktree paths can now be copied. I'm finally free of the evil built-in VS Code terminal 🫡 https://t.co/L08svmmRkQ↗

I built osgrep It’s a local code search tool that understands natural language. Works as a standalone CLI or a plugin for Claude Code. No API keys or subscription. I wanted the power of "semantic search" without the latency, price, or privacy trade-offs. Video in realtime. ht…↗

I don’t know wtf this is but if you want MIT-licensed code search + CLI that you can use for yourself and for your devtools, check out https://t.co/9mMKJ43KAt↗

we shipped find all today, which allows you to query the web like a database. give a natural language prompt, get a high-quality dataset back. bring structure to the world's largest, unstructured repository of info - the web. hedge funds use this for idea generation, gtm tea…↗

https://t.co/cIwA9BwVm5 files are better: just-in-time indexing. Broke my 800-line Convex rules into 400-line context-specific chunks. Agents load only what's relevant. https://t.co/uUJOhBtGkh↗

5. Vision is improved. Opus is better at image processing and data extraction, especially with multiple images. For dense images, give it a cropping tool to "zoom" in on regions. We've seen consistent uplift on image evals with this.↗

others: better run in a sandbox llms are not safe me: literally hooks up claude to whatsapp xD you can limit actions to specific phone numbers tho. Supports both twilio and reverse engineered whatsapp sync (using a second phone is way easier than Meta WA Business verification) …↗

2 months and 5.7k downloads later It's fun building open-source coding agent orchestration tooling with @arnestrickmann https://t.co/ZlQn8O87DT↗

Figured out a better way how to share multiple agent files with codex. Tell it to read files on startup. https://t.co/IFXc6wFCAA↗

LFG 🔥 @FactoryAI Thanks @bentossell for sending it over all the way to Germany 😅 Got some SF vibes rn https://t.co/eNlPtnX2n6↗

@darrenjr virtual filesystem backed by db will work just fine imo↗

3 weeks ago, I was tasked to revamp the website for the AI Engineer Code Summit in NYC. With the conference officially kicking off TODAY, I revisited an original design to highlight every speaker. (demo ft. @kevinhou22 & @barry_zyj) This is the most stacked speaker lineup I ha…↗

So excited to be back in the unpaid labour game (OSS), recorded a 5 minutes video of what contributors and I accomplished in my shitty coding agent pi today: https://t.co/8PWFiNcmT5 - message queuing - tool output expand/collapse - thinking level and model cyclcing shortcuts -…↗