Microduck 로봇 학습·Kimi Code 원격 제어·Wan 3.0 통합 생성·PRAXIST 연구 루프

So your telling me it's Open Source and I can print STL parts for it? Take my money. … https://gr illz-microduck-simulator.onrender.com

just had a quick look at microduck_rl the codebase is very elegant, i’d recommend reading the agent.md since it contains quite a few fun quirks for reward modeling like head tracking too tight impairs walking cause the head is 38% of the duck’s weight, so it naturally oscillates -> to solve this, smooth the head tracking error using ema, essentially penalizing only the dc bias also a few other quirks as you dig deeper into the codebase like how they model the backlash of the motor by adding an unactuated hinge (with very small range) in series with the motor gg @antoinepirrone

Kimi Developers

@KimiDevs

1일 전

Kimi Code New Feature: Remote Control Kimi Code 0.39.0 add Remote Control as an experimental feature for accessing a local web session remotely. Enable it with KIMI_CODE_EXPERIMENTAL_REMOTE_CONTROL=1, then run kimi rc, kimi web --remote-control , or /remote-control to start it.

Wan 3.0 is now on @DeepInfra ! Video + audio in one model, with 30-second generation and multimodal references. Now available on DeepInfra.

Wan 3.0 is now live on @PixelDojoAI ! Check out the guide and start creating.

Jerry Liu

@jerryjliu0

1일 전

PDF parsing is fun because there's an infinite variety of enteprise documents . For each document category, there's a long tail of work to build more precise bounding boxes, confidence scores, and domain-specific annotations so that you provide any downstream agent rich metadata without it having to reinvent this from scratch. Take forms for example. Besides simply outputting it into markdown, we put in the work to detect every annotation, field, checkbox, and section. That way you immediately get structured information as to whether a form is filled without a separate LLM extraction step. You also get source citations for free! Doing this well is hard. There's a rabbit-hole of optimizations you can do, including extracting annotations for every document type. It's hard to properly render visual formats like charts, handwriting into digitalized information. The more you skip this step, the more work you're creating for any downstream agent. Precise bounding boxes are a necessity for precise citations on any type of document. You can aggressively tune the model + harness so that the accuracy/cost on any document subtype is much more competitive than the frontier models. Whether you're parsing forms (see the enriched forms option in "processing options") or any other type of doc, check out LlamaParse ! https:// cloud.llamaindex.ai

This is not just another coding agent. In a comparison with Claude Code + Opus 4.8, PRAXIST achieved: • 49 gold medals across 75 tasks, vs. 35 for Claude Code • ~$3K in token cost vs. ~$38K for Claude Code That's better benchmark performance at roughly one-twelfth the cost! With PRAXIST, you define a measurable technical goal + an evaluator for what success looks like, and PRAXIST figures out the rest: • Multiple agents explore different approaches in parallel • They run experiments and validate the results • They learn from failures and share findings • Each new round builds on what previous experiments learned • You get the final solution + the evidence showing how it got there

Builders are shipping with Qwen3.8. Next Tuesday, they're going live to show you how. Qwen3.8 Show & Tell | 1 Sept | 17:00 GMT+8 · A local AI copilot that never touches the cloud · A full business running on WhatsApp (Qwen3.8-27B) · An autonomous agent that ships while its creator sleeps Register now: https:// luma.com/859h4n12 #AlibabaCloud #BuildWithQwen