Tawsif Kamal

Voice mode in Antigravity is here. Enabled by the Gemini Live model, you can use your voice to plan, steer code changes, and approve permissions out loud. Have a live conversation across multiple languages, interrupt whenever you need to, set your preferred voice and playback speed, and when you're done, just say "bye" or tap end. Happy building!

Professor Jiang: ChatGPT isn’t actually generating everything by itself, there’s a lot of human manipulation going on in the algorithm. “Do you really think when you put something into ChatGPT, it’s completely generated by the AI? I guarantee you it’s not true.” “There’s a lot of human intervention to make sure that the answer conforms to certain expectations.” “Maybe it’s 90% AI, 10% human, but it’s still 10% human.” “In order for ChatGPT to generate essays, it needs samples. It needs a lot of samples.” “It has to constantly absorb more human expertise in order to refine the model.”

OpenAI turned its most capable cyber models against their own infrastructure to continuously attack their systems. In the age of AI-powered attacks, defense factories that find vulnerabilities, remediate them, and verify the patch autonomously are the only way to stay secure. We've spent months building this with major healthcare systems and cloud providers. Today, we’re making it available to everyone. > Find: Continuously probes your systems to identify vulnerabilities > Fix: Writes the fixes and opens PRs for review > Verify: Retests every patch until the vulnerability is gone Find. Fix. Verify. Repeat.

I led engineering at Google DeepMind. Today, I'm proud to introduce Fo to give personal AI something no lab ever has... Humans. Other personal AI's pretend AI can do everything. Fo employs humans to do tasks that AI cannot. - 2x better at real-world task completion (beats other agents by 69%) - 94% trust rate (4x less likely to leak private info vs Muse, Instinct) Sign up for free: wajo.ai/join-wajo

Introducing terminal-code: VS Code inside the terminal - VS Code compatible CLI - works over ssh - syncs with your terminal theme terminal-code.com

my ideal agentic IDE setup after using several IDEs (@superset_sh , @conductor_build , @vorfluxai , @DevinAI , @Cursor, @herdrdev ), I've first principled my way into a set of features I'd like. 1) worktrees - goes without saying. you need this for isolated environments and parallelism. 2) nested worktrees - I've only seen devin use these. an agent in a worktree can produce a sub-worktree, which you can click into. A really neat way of passing isolated context into a new pane. 3) subtabs - multi-threading across the same worktree, so i'm not forced on one chat per worktree. conductor does this. 4) chat - having a good UI for components etc. is super nice. its the little things that matter here - like suggesting skills, letting users enter a password in an encrypted box instead of plaintext, etc. 5) a display editor for a browser / markdown / artifact. You should be using noVNC on the cloud to stream back a browser with computer use. This browser should be in sync with whatevers going on in the code. i.e if I'm making a UI change, show me the UI on the browser. vorflux has a very nice design pane here in the form of a whiteboard, which is also cool, so the models can output figma styled prototypes. all the agentic IDEs can do some pretty cool things with this tab for good DX. 6) a worktree-manager agent - I haven't seen anyone do this yet but why not have an agent that is context aware of all the worktrees. It can tell me about my overall work, whats been going on within the team and the project, etc. Other feature requests: 1) local to cloud sync - conductor has come closest to this with their option for conductor cloud, but i can't move the session from cloud to local and back, which I'd like to do when accessing my own filesystem. 2) update worktrees when the source changes. this is a big one. when I start a worktree, it obviously clones main. but if main changes, or someone added a new .env, I have to tell the worktree to pull from main. why? it should auto pull from main by subscribing to updates. its obvious that in any case when I do make a PR, I will have to rebase on main. At the very least, tell me main has changed.

The most annoying part about running your agents in the cloud is that sometimes you _do_ need access to your local machine. With Conductor Cloud we give you all the benefits of cloud but also a bunch of local escape-hatches. All cloud agents get access to: - Port-forwarding from cloud to your localhost - File sync from cloud env to your machine - A "run local command" tool that lets agents remotely run commands on your Mac Here's a walkthrough: