drxmingc's profile picture. data & ai guy @alibaba_cloud. founder of @mirasurf_ai. ph.d @sjtu1896.

Xiaming Chen

@drxmingc

data & ai guy @alibaba_cloud. founder of @mirasurf_ai. ph.d @sjtu1896.

How to crack pitfalls of LLM on continual learning. Here is my proposed solution, Composable Learning. A position post here, xiaming.site/2025/11/05/com…


Xiaming Chen reposted

343+ Quant and Algorithmic Trading Projects in Python This is how to get your hands on them for free. (a thread)

quantscience_'s tweet image. 343+ Quant and Algorithmic Trading Projects in Python

This is how to get your hands on them for free.

(a thread)

Xiaming Chen reposted

I've found the path to AGI: arxiv.org/abs/2510.12269


Xiaming Chen reposted

What can we learn from Anthropic’s Blog about building tools for Agents? Here are the key principles I extracted for building Tools and MCP Servers for Agents: - Consolidate workflows: Implement `schedule_event` instead of separate `list_users`, `get_availability`, and…

_philschmid's tweet image. What can we learn from Anthropic’s Blog about building tools for Agents? Here are the key principles I extracted for building Tools and MCP Servers for Agents: 

- Consolidate workflows: Implement `schedule_event` instead of separate `list_users`, `get_availability`, and…

Xiaming Chen reposted

This is really exciting and impressive, and this stuff is in my area of mathematics research (convex optimization). I have a nuanced take. 🧵 (1/9)

Claim: gpt-5-pro can prove new interesting mathematics. Proof: I took a convex optimization paper with a clean open problem in it and asked gpt-5-pro to work on it. It proved a better bound than what is in the paper, and I checked the proof it's correct. Details below.

SebastienBubeck's tweet image. Claim: gpt-5-pro can prove new interesting mathematics.

Proof: I took a convex optimization paper with a clean open problem in it and asked gpt-5-pro to work on it. It proved a better bound than what is in the paper, and I checked the proof it's correct.

Details below.


A new agents.md is proposed by openai. dotaidev could be retired. github.com/openai/agents.…

When you are a programmer who want to switch AI tools from different vendors, but meanwhile remain your work progress. the best method is to use a vendor agnostic configuration. if you have the same problem you can try dotaidev project



Xiaming Chen reposted

Summary of GPT-OSS architectural innovations: 1. sliding window attention (ref: arxiv.org/abs/1901.02860) 2. mixture of experts (ref: arxiv.org/abs/2101.03961) 3. RoPE w/ Yarn (ref: arxiv.org/abs/2309.00071) 4. attention sinks (ref: streaming llm arxiv.org/abs/2309.17453)


Xiaming Chen reposted

great ai engineers think like farmers: > don’t shout at the llm > don’t blame the llm for not being smart enough > choose the best model for the task > don’t throw away the model before testing it with different prompts > remove redundant context

Hesamation's tweet image. great ai engineers think like farmers:
> don’t shout at the llm
> don’t blame the llm for not being smart enough
> choose the best model for the task
> don’t throw away the model before testing it with different prompts
> remove redundant context

Xiaming Chen reposted

BUILD IN PUBLIC People are asking about why AI Slides 2.0's "Deep Thinking" is so good (achieves 90% fewer hallucinations). I want to share the secret sauce: We transferred the knowledge from Claude Code's TodoWrite task planning architecture to slide generation. From Code to…

🚨 Goodbye Canva, Gamma, and others! Genspark AI Slides 2.0 is here! — Hundreds of slide templates, all FREE! — Introducing Deep Thinking with 90% fewer hallucinations! — Advanced manual editing supported!



congrats to qwen team

Qwen3-Coder is live and open! 480B parameters, 35B active, up to 1M context. Non thinking and powerful!

ivanfioravanti's tweet image. Qwen3-Coder is live and open! 
480B parameters, 35B active, up to 1M context.
Non thinking and powerful!


Loading...

Something went wrong.


Something went wrong.