Sabitlenmiş Tweet
Three years ago, I started a microservices demo project. Last month, I noticed it had grown to 80+ stars and multiple forks. That was my sign to keep building.
Just shipped some major updates:
👇🧵
English
Adnane Miliari #nanodev
135 posts
Adnane Miliari #nanodev
@nanodev_ma
Software Engineer || Backend Developer || Moving slowly but surely ! 🐢
Morocco Katılım Şubat 2025
393 Takip Edilen54 Takipçiler
Adnane Miliari #nanodev retweetledi
Adnane Miliari #nanodev retweetledi
You can contribute easily to Mrrakc.com by using build.Mrrakc.com
Access code: MrrakcIsLive!
English
Adnane Miliari #nanodev retweetledi
1. Custom maps:
Like in mrrakc.com/maps/ you can build custom maps based on the data and schema.
There is a ready to use schema for maps here github.com/Iduoad/Mrrakc/…
And example data is here
github.com/Iduoad/Mrrakc/…
English
Adnane Miliari #nanodev retweetledi
Excellent.
OpenAI and Anthropic are competing in a way that soon there will be nothing left for us to do.
But to use these amazing tools, I need money to buy subscriptions. For money, I need a job.
If you take away all the jobs(if not all, majority of them) then we won't have jobs and no money.
So, leave few things for us, please.
OpenAI@OpenAI
Codex for (almost) everything. It can now use apps on your Mac, connect to more of your tools, create images, learn from previous actions, remember how you like to work, and take on ongoing and repeatable tasks.
English
Adnane Miliari #nanodev retweetledi
Adnane Miliari #nanodev retweetledi
I think people are understating Cursor’s technical achievements for Composer 2.
Small team of 40 researchers, doing large scale mid-training and RL at a scale that no startup has done before, beat Opus 4.6 in Terminal-Bench.
Strong work needs the strong foundations of the Kimi base model, but it’s the post training that makes this model frontier.
Kimi.ai@Kimi_Moonshot
Congrats to the @cursor_ai team on the launch of Composer 2! We are proud to see Kimi-k2.5 provide the foundation. Seeing our model integrated effectively through Cursor's continued pretraining & high-compute RL training is the open model ecosystem we love to support. Note: Cursor accesses Kimi-k2.5 via @FireworksAI_HQ ' hosted RL and inference platform as part of an authorized commercial partnership.
English
@bernoussama claude opus4.6 slightly better on claude code, but not sure if the diff comes from claude/opencode rather than the models themselves
English
@enlamp Level 1 — JWT authentication (On the SecurityConfig class)
Level 2 — Role-based access (PreAuthorize -> @PreAuthorize("hasRole('app_user') or hasRole('app_admin')"))
)
Level 3 — User identity in AI context (screenshot)
English
@nanodev_ma exciting! would be interesting to see how you solve authn/authz for the AI layer. keeping an eye an this. ✍️
English
📢continuous learning alert: "agents as yaml" from docker
📦package and distribute custom agents with instructions, LLMs and tools as OCI artefacts, like plain old container images.
🤩docker unmatched developer experience strikes again
👇link to the tutorial in comment
djelal e5z@enlamp
a wish came true, in pure docker compose concise style, 🐳💕🤖
English
Adnane Miliari #nanodev retweetledi
Looks like it’s confirmed Cursor’s new model is based on Kimi! It reinforces a couple of things:
- open-source keeps being the greatest competition enabler
- another validation for chinese open-source that is now the biggest force shaping the global AI stack
- the frontier is no longer just about who trains from scratch, but who adapts, fine-tunes, and productizes fastest (seeing the same thing with OpenClaw for example).
Lee Robinson@leerob
Yep, Composer 2 started from an open-source base! We will do full pretraining in the future. Only ~1/4 of the compute spent on the final model came from the base, the rest is from our training. This is why evals are very different. And yes, we are following the license through our inference partner terms.
English
👌
Piotr Mińkowski@piotr_minkowski
I created my Claude Code template repo for the Spring Boot app with instructions, skills, and subagents.💡 It is desired to create an app that connects to the DB, is deployed on Kubernetes, and runs tests with Testcontainers, and some others 👉github.com/piomin/claude-…
ART
Adnane Miliari #nanodev retweetledi
Adnane Miliari #nanodev retweetledi
أندريه كارباثي "واحد من أهم الأسماء في مجال الAI، واشتغل على تطوير OpenAI"
نشر مشروع بيقيّم كل وظيفة على مقياس من 0 ل10 حسب قد إيه الذكاء الاصطناعي ممكن يستبدلها.
- جمع كل الـ 342 مهنة من بيانات مكتب إحصاءات العمل
- دخل كل واحدة على نموذج لغوي باستخدام معيار تقييم تفصيلي
- عمل خريطة تفاعلية، حجم المربع فيها = عدد الوظائف، واللون = قد إيه الوظيفة دي متعرّضة للذكاء الاصطناعي
أهم نقطة في التقييم بتاعه كان لو ناتج الشغل كله رقمي بطبيعته، والوظيفة ممكن تتعمل بالكامل من البيت، يبقى مستوى التعرض للذكاء الاصطناعي عالي تلقائيًا
المقياس كان كالتالي:
0–1: عمال الأسطح، عمال النظافة
4–5: ممرضين، بيع بالتجزئة، أطباء
8–9: مطوري برمجيات، مساعدين قانونيين، محللي بيانات
10: ناس بتفرّغ التقارير الطبية صوتيًا
وفي الآخر كان متوسط تقييم الذكاء الاصطناعي لكل ال342 مهنة كان 5.3 من 10
النظام بالكامل مفتوح المصدر، من سحب البيانات، لتقييم النموذج، للتصميم البصري، كله متاح
t.co/iWIh8qhvdu
العربية
Adnane Miliari #nanodev retweetledi
اللهم انا لا نسألك رد القضاء ولكن نسألك اللطف فيه 😭
Claude@claudeai
Introducing Code Review, a new feature for Claude Code. When a PR opens, Claude dispatches a team of agents to hunt for bugs.
العربية
even more interesting, more thinking does not increase BS detection. Sonnet 4.6 does even slightly better than Opus 4.6 "high". More tokens != less BS.
djelal e5z@enlamp
the BS benchmark: this may explain why Anthropic is winning the coding battle. nobody has time for BS during software dev., whatever the context window, price per M. token. honorable mentions for opensource Kimi and Qwen3
English
Adnane Miliari #nanodev retweetledi
