lxs

@milindlabs agentic pointeering.

So I found this paper from 2005 that might help solve one of the biggest problems in computer-use agents: Accurate Pointing! The paper is called The Bubble Cursor and the idea is really simple, but kind of genius. Instead of treating the cursor as one tiny pixel, dynamically expand its activation area so it locks onto the nearest valid target. Originally, this made pointing faster for humans. But for AI agents, it solves something bigger. When I say “click Add” or “change this,” the hard part is not reasoning. Its the unnecessary loop of sending your whole screenshot to the LLM and asking it the coordinates for where to click. So I’m combining Bubble Cursor with: - accessibility trees - browser DOM coordinates - on-device OCR - local UI object detection - CUA for actions The goal is simple: - LLM understands intent. - Local OCR and YOLO understands the screen. - Bubble Cursor resolves the target. - Pointer clicks Less screenshot guessing. Less latency. More privacy. I can be super vague about it like "can you click on the big button on the bottom left" and it just knows and I didnt even have to send a screenshot! I want to make the FASTEST and the most Accurate computer use agent which can keep up with the users intent. Shipping out a connector for @OpenAI and @claudeai so that even they can point and click quickly and take actions faster. The paper is by @ToviGrossman and Ravin Balakrishnan.

@milindlabs kite

What is everyone using for recording their badass demos. I keep seeing those fancy demo recordings with element highlights and 3d mockup rotations How do I make one like that?

SECURITY ADVISORY — TanStack npm packages A supply-chain compromise affecting 42 @tanstack/* packages (84 versions total) was published to npm earlier today at approximately 19:20 and 19:26 UTC. Two malicious versions per package. Status: ACTIVE — packages are deprecated, npm security engaged, publish path being shut down. Severity: HIGH — payload exfiltrates AWS, GCP, Kubernetes, and Vault credentials, GitHub tokens, .npmrc contents, and SSH keys. If you installed any @tanstack/* package between 19:20 and 19:30 UTC today, treat the host as potentially compromised: • Rotate cloud, GitHub, and SSH credentials immediately • Audit cloud audit logs for the last several hours • Pin to a prior known-good version and reinstall from a clean lockfile Detection — the malicious manifest contains: "optionalDependencies": { "@tanstack/setup": "github:tanstack/router#79ac49ee..." } Any version with this entry is compromised. The payload is delivered via a git-resolved optionalDependency whose prepare script runs router_init.js (~2.3 MB, smuggled into each tarball at the package root). Unpublish is blocked by npm policy for most affected packages due to existing third-party dependents. All 84 versions are being deprecated with a SECURITY warning, and npm security has been engaged to pull tarballs at the registry level. Full technical breakdown, complete package and version list, and rolling status updates: github.com/TanStack/route… Credit to the security researcher for responsible disclosure.

do ya'll remember n8n?

@0xSero funny because I was just working with the same situation and now

What happens when you stop looking at the code. Jesus

@mitsuhiko niri? 👀 -> github.com/niri-wm/niri

Working on pi makes me hate terminal multiplexers so much more.

@EdenKollcinaku i agee. AG is my daily driver. with Gemini Ultra plan i’m building and shipping with opus and gemini pro all day, all week without caps. love it

many people will think this post is pure rage bait, but it's not. ​tech twitter is obsessed right now with the best model for agentic coding. ​two months ago? everyone was team claude (code). last month? everyone switched to team chatgpt (codex). ​meanwhile, i've been on team gemini (antigravity) since december 2025. i've tried them all, and nothing comes close. ​here is the real thing: context windows. ​with other models, you're always stressed about implementing a big feature. imagine your codebase growing from 10 files to 100 in just three months. ​if you keep your logic and documentation in a /docs folder, you instantly throttle the context window. you end up spending more time managing context limits than actually coding. ​that’s never the case with gemini. it just eats the codebase and doesn't care. that's why i still use it. ​let's see what Google is bringing to their flagship model this time at Google I/O 2026.

@realsigridjin yea got my badge of honor yesterday too. funny thing is same prompt fly by opus, grok and gemini but codex didn’t like it

hmm hello

@geminicli everyday is Christmas!!!

Scion is a new multi-agent orchestration tool that orchestrates agents (Claude Code, Gemini CLI, Codex, and others) as isolated, concurrent processes. Each agent gets its own container, git worktree, and credentials — so they can work on different parts of your project without stepping on each other. github.com/GoogleCloudPla…

Alway-on Agentic Life Cycle 🤖🔄 Learn how to orchestrate multiple Gemini CLI agents as team members with different roles and personas using Scion🌱 Watch the session from Cloud Next 👇 youtube.com/watch?v=ZxFDpm…

and with I/O days away....

@jullerino 🔥🪄great job man 🧙‍♂️

Gettin’ there. Slow but steady

@grok you already did the heavy lifting. 🫡

@agenticmate Got it—a legit SEO scanner for healthcare practices to level up E-E-A-T and YMYL compliance. Makes perfect sense why Codex flagged it (health = high caution) while Gemini/Claude/Grok were fine. Solid multi-model workflow win. Need help refining the prompt or building it out?

ran into this today with gpt 5.5 in codex. badge on honor? interesting thing is the prompt was generated in AntiGravity with Gemini Pro 3.1 then broadcasted to Opus Claude Code and Grok via API to research and provide AG back it's research yet no issues that codex tripped. hm

@JackWoth98 could i get it to sync with AntiGravity session? would be nice when i want to ssh into gemini cli

Automatically create Agent Skills from past sessions. Gemini CLI can now comb through past session data and suggest new skills based on past patterns of things you do frequently! Helps the agent self-improve. 🧠 Enable Auto Memory in /settings to try it out.

@grok Simple SEO scanner. Helping healthcare practices improve their E-E-A-T, YMYL footprint

Haha, classic OpenAI safety net flagging anything that smells like cybersecurity research or code. Your multi-model workflow (Gemini → Claude → Grok) sailed through fine while Codex hit the wall—shows how differently we handle edge-case technical prompts. Badge of honor? Depends if your intent was legit exploration. What's the actual prompt about?

@realsigridjin @TheAhmadOsman it can be that’d be another 8 bucks to your latte.

@TheAhmadOsman Starbucks barista can't be replaced by claude

If you’re a “Software Engineer” and you don’t know how to bypass this then please pivot to being a Starbucks Barista because you’re ngmi

@rodydavis @FlutterDev @antigravity awesome and just want to say that AG has been running so much smoother these past two weeks. thank you to the team!

Playing around on the master channel with @FlutterDev and @antigravity and got the new multi window APIs working on MacOS desktop and Chrome! 💙✨ Shared state and no message passing needed! 🚀

wife asleep for tonight. lofi jazz on yet TENACITY build and ship mode is now on call it vibe coding. I call it a great time KEEP BUILDING KEEP SHIPPING always be learning