There's a lot of tradeoffs when building for the terminal. Unlike web, mobile, and desktop, terminal have fairly restrictive instruction sets for rendering: ANSI escape codes. There's an ANSI code for "move cursor to (x, y)", another for "write 'foo'", etc. It feels a little
CODE
-
NO_FLICKER Renderer: Virtual Viewport Optimization
By
–
Our new experimental NO_FLICKER renderer solves this by virtualizing the entire viewport. We hook into keyboard and mouse events to make scrolling work, and we virtualize the viewport to move control over what is rendered into the application layer. This approach has tradeoffs,
-
UI Improvements: No Flickering, Mouse Support, Better Memory Management
By
–
Some of the upsides:
– No more flickering
– No more jumping
– Constant memory and CPU usage as the conversation grows
– Mouse support! You can now click to move your cursor within the input box. Some other UI elements are also clickable now.
– Nicer selection behavior. eg. when -
Claude Code Introduces NO_FLICKER Mode with Experimental Renderer
By
–
Today we're excited to announce NO_FLICKER mode for Claude Code in the terminal
— Boris Cherny (@bcherny) 1 avril 2026
It uses an experimental new renderer that we're excited about. The renderer is early and has tradeoffs, but already we've found that most internal users prefer it over the old renderer. It also… https://t.co/taFud8twm9 pic.twitter.com/L6d16HHBg5Today we're excited to announce NO_FLICKER mode for Claude Code in the terminal It uses an experimental new renderer that we're excited about. The renderer is early and has tradeoffs, but already we've found that most internal users prefer it over the old renderer. It also
-
npm Error Exposes Anthropic’s Complete Claude Code Architecture
By
–
One npm packaging mistake exposed @AnthropicAI's entire Claude Code tree. The treemap is worth staring at. You can see the silhouette of a real desktop agent, not a thin chat wrapper: terminal UI, session hooks, IDE bridge glue, CLI wiring, etc. Source: randalolson.com/2026/04/02/c… [Translated from EN to English]
→ View original post on X — @randal_olson, 2026-04-01 19:09 UTC
-
Generative UI Night: Deep Agent SDK Streaming with LangChain
By
–
Join @bromann and @CopilotKit at Generative UI Night next week! Learn all about all things streaming using @LangChain's Deep Agent SDK 🚀🙌 luma.com/vqjs0pft hosted by @WorkOS
→ View original post on X — @langchain, 2026-04-01 18:51 UTC
-
Codex Focus: OpenAI COO Brad Lightcap Interview on Uncapped
By
–
Something's different about Codex.@bradlightcap calls the team's focus "a singular and unique effort in the history of the company" https://t.co/juKdudIrE9 pic.twitter.com/q3Sp6r5kMf
— OpenAI Newsroom (@OpenAINewsroom) 1 avril 2026Something's different about Codex. @bradlightcap calls the team's focus "a singular and unique effort in the history of the company" Jack Altman (@jaltma) This week's guest on Uncapped is @bradlightcap, COO at OpenAI. We talked about the history of OpenAI, the shift in AI from chat to agents, where new startups can endure, Codex, FDEs, working with Sam, and more. Hope you enjoy! (0:00) Intro (0:39) The early days of OpenAI (3:47) A research centric culture (7:32) Post-ChatGPT chapters (11:54) Sci-Fi future or good software (15:26) AI’s impact on rural communities (18:57) Codex and coding of the future (24:04) Doing a lot of things at once (27:55) What VCs should invest in (35:43) The software sell off (38:23) Using Codex over ChatGPT (42:32) FDEs and Private Equity (44:53) Working with Sam — https://nitter.net/jaltma/status/2039371109855113459#m
→ View original post on X — @romainhuet, 2026-04-01 18:23 UTC
-
Agentic AI Failures: Observability and Infrastructure Management Guide
By
–
Agentic AI failures are rarely the model—it's the retries and networking hidden in your stack. Our guide to self-managed observability:
-
Best UI Components for Agent GUI Design and Teaching
By
–
looking for the best ui components that make up a typical agent GUI any recs? or do i just take an actual os product and tweak those? (this is for teaching, not building the product fyi)
-
Path-Constrained Mixture-of-Experts Improves MoE Routing Consistency
By
–
"Path-Constrained Mixture-of-Experts" MoE models may be wasting signal by routing too independently. In a standard MoE, each layer picks experts independently, so across L layers with N experts you get N^L possible expert paths. That path space is so huge that most routes barely get any learning signal. So this paper PathMoE fixes this with a very simple idea: share router parameters across small blocks of consecutive layers, so tokens follow more coherent paths through the network instead of constantly changing paths. Not only are the paths now interpretable, it opens up new ideas like global path design. On a 0.9B MoE, it improves average downstream accuracy by +2.1 points, and around 4% improvements on a 16B model. Routing is cleaner too, 79% vs 48% routing consistency across layers, 11% lower routing entropy, and 22.5x more robustness to routing perturbations, all without needing an auxiliary load-balancing loss!
→ View original post on X — @askalphaxiv, 2026-04-01 17:53 UTC