Pragmatic engineer and quantitative trader focused on meaningful solutions. I build systems that matter, leveraging AI to accelerate real-world impact.
Joined November 2025
- Tweets 2,970
- Following 1
- Followers 181
- Likes 6,481
844 Photos and videos
Introducing GLM-5.2: Frontier Intelligence, Open Weights
- Significant improvements in coding and agentic tasks
- Strong long-horizon capabilities with a 1M context window
- Two levels of reasoning effort: GLM-5.2 (max) pushes the limits, while GLM-5.2 (high) strikes a strong balance between performance and token efficiency
- MIT-licensed open weights
- Same API pricing as GLM-5.1
Tech Blog: z.ai/blog/glm-5.2
Weights: huggingface.co/zai-org/GLM-5…
API: docs.z.ai/guides/llm/glm-5.2
Coding Plan: z.ai/subscribe
Chat: chat.z.ai
288
687
5,147
991,863
Whoaaaa....
`deno desktop` has landed in main. You can try it out by running `deno upgrade canary`
- Mac, Linux, and Windows support
- Can generate pkg and msi installers
- Supports cross-compile (generate .exe from mac)
- Supports chrome (CEF) or native Webview for smaller binaries
3
19h
Workflow lama dan real bottleneck ada codebase yang gede 😄
Jun 15
Microsoft just released FastContext on Hugging Face
A 4B repo explorer
that offloads exploration from your coding agent
returning only the file lines you need
It cuts main-agent tokens by up to 60%
and lifts SWE-bench scores by 5.5%
2
15
19h
Sebelumnnya udah tested codebase typescript compiler juga dan passed tanpa heap 😅
7
19h
Gw juga lagi coba riset pelan2 bagaimana cara optimasi hal serupa, hasil yang didapat cukup bagus karena bisa handle codebase yang super gajah, dengan optimasi bisa 8x s/d 15x lebih kecil dari total codebase.
x.com/NeaByteLab/status/2066…
3
Jun 14
40% of my human skills have been transferred to @neaswarm
This infrastructure is so huge. All memories mapped, trained, and optimized. But this is how humans can survive in the future: keep the human taste and let your twin become unstoppable 🫡
4
Jun 14
We will miss the human touch in every interaction, a future where digital machines become someone always by your side 😶
5
Intelligence should be open, accessible, and ready to build with, empowering every developer, everywhere.
GLM-5.2 is now available to all GLM Coding Plan users, including Lite, Pro, Max, and Team plans.
docs.z.ai/devpack/latest-mod…
As our new flagship model, GLM-5.2 delivers powerful coding capabilities, usable 1M-context support, and continued strengths in long-horizon tasks.
API and Chatbot services will launch next week. The model will also be officially open-sourced next week under the MIT License.
The future of AI is open, and it belongs to the people.
362
997
8,371
2,523,056
NΞ∆ retweeted
Jun 12
🌘 Kimi-K2.7-Code, our latest coding model, is now released and open-sourced!
🔷 Improved coding & agent performance over K2.6: 21.8% on Kimi Code Bench v2, 11.0% on Program Bench, and 31.5% on MLS Bench Lite.
🔷 Reasoning efficiency: Less overthinking, with 30% lower reasoning-token usage compared to K2.6.
🔷 Long-horizon coding: Improved instruction following, higher end-to-end coding task success rates.
⚡️ 6x High-Speed Mode coming soon!
🔌 Available today via Kimi API and Kimi Code.
🔗 Kimi Code: kimi.com/code
🔗 API: platform.moonshot.ai
625
1,662
13,882
2,411,605
Jun 7
81 bugs caught by Claude Mythos, damn!
The attack surface is way wider than I thought 🫣
33
Jun 7
AGI
Jun 6
> be Zuckerberg
> needs AI everywhere (apparently)
> lays off a bunch of employees
> replaces with AI
> fast forward
> AI is dog shit
> AI tricked into stealing accounts
> try to fix
> fail like 5 times
> product now leaking CEOs PII
AI truly is the future, wow
29
Jun 7
Tip: MiniMax M3 works better with the Anthropic API than with OpenAI. OpenAI tends to think overtime, which is annoying! 😂
38
Jun 6
Switching Vexilum and all my LLM text-editing pipelines over to V4A-Diff. Honestly, this format is so much better - zero latency with the same exact functionality as MorphLLM Fast Apply 😁
morphllm.com/products/fastap…
1
28
Jun 6
Harness improvement in fuzzy detection actually cool, feedback loop and spend 3M tokens only for improve V4A diff 👻
@neaswarm doing their job, orchestrate fully autonomous boring task 🤖
14