Kurniawan

🚀 Building AI & Python automation in public
🛠️ AI • Python • Web Automation • Prompt Engineering
📖 Learning every day.
🌐 Tutorial, experiments, etc

[Kurniawan @kurniswan]

AI doesn't just need better models.

It needs better ways to measure progress.

If the benchmark is wrong, every comparison built on top of it becomes less meaningful.

Better evals benefit the entire ecosystem.

[OpenAI @OpenAI]

We audited SWE-Bench Pro, one of the most widely used AI coding benchmarks, and found it no longer reliably measures frontier coding capability.

We find 30% of SWE-Bench Pro tasks to be broken, and are retracting our previous recommendation that the research community use it as a leading coding eval.
https://t.co/wDdSEjBe4F

[Kurniawan @kurniswan]

@OpenAI Better benchmarks don't just measure progress.

They shape what the entire industry optimizes for.

[Kurniawan @kurniswan]

Every major AI release creates a short window of opportunity.

While most people wait for reviews, builders are already experimenting.

The next few days won't be about who has GPT-5.6.

They'll be about who ships something useful with it.

[Sam Altman @sama]

GPT-5.6 sol launches thursday!

happy building

[Kurniawan @kurniswan]

@sama The biggest winners won't be the people asking "what's new?"

They'll be the ones shipping before everyone else catches up. 🚀

[Kurniawan @kurniswan]

@StarPlatinum_ What goes down will come up...eventually

[Kurniawan @kurniswan]

@arieagung @KavezuAzila @AuliaBismar berarti pakai gmail yang bisa dikorbankan ya?

[Kurniawan @kurniswan]

@k2sbhai this is interesting, I should try this
Are there any country restriction?

[Kurniawan @kurniswan]

Hari ini aplikasi kantor dipasangi anti-bot untuk mencegah para pegawai untuk menggunakan bot atau tool otomasi dalam membantu pekerjaannya.

[Kurniawan @kurniswan]

4/8
DeFi yield tinggi? Iya, tapi risikonya juga tinggi. 2024 aja rugi $2.9M+ gara-gara hack, smart contract bug, rug pull. Bukan "passive income" tanpa risiko.
📎 https://t.co/8ynuiV6kLd

[Kurniawan @kurniswan]

1/8
gw nemu postingan rame bahas Web3 & janji cuan. definisinya gak salah, tp claim "banyak cuan" perlu konteks lebih dalem. gw coba cek 🧵

[Kurniawan @kurniswan]

3/8
Play-to-Earn: bener bisa dapet NFT/kripto. Tapi 93%+ game blockchain gagal di tahun pertama. Ekonomi token sering gak sustainable. ⚠️
📎 https://t.co/xxCMOuT1G8

[Kurniawan @kurniswan]

7/7
Intinya, skill di Web3/DeFi itu emang beneran ada nilainya dan bisa menghasilkan.

Tapi prosesnya itu maraton, bukan kompetisi sprint yang viral semalam langsung kaya. Stay curious, stay safe, and do your own research! 🧠✨

[Kurniawan @kurniswan]

1/7 Beneran bisa dapet 25 juta dalam 2 minggu cuma modal HP sejuta + skill Web3? Pas kemarin nanya ke AI, jawabannya menarik banget.

Karena agak panjang dan padat, aku coba bikin thread-nya di bawah ya! 🧵👇

[Kurniawan @kurniswan]

6/7 Poin pentingnya: kalau mau coba masuk ke dunia ginian, pakai uang dingin yang kamu siap kalau hilang 100%.

Fokus belajar mekanismenya dulu (misal cara kerja LP di platform kayak Meteora, dll), dan buang jauh-jauh mentalitas pengen kaya instan.

[Qwen @Alibaba_Qwen]

🚀 Qwen Code v0.14.0 – v0.14.2 are now available

Channels：Control Qwen Code remotely from Telegram, DingTalk, or WeChat — send a message from your phone, get results on your server
Cron Jobs ：Schedule recurring AI tasks — auto-run tests every 30 min, pull & build every morning, monitor logs on a timer
Qwen3.6-Plus ：New flagship model with 1M token context, 1,000 free daily requests
Sub-agent Model Selection：Assign different models to sub-agents — use a powerful model for the main task, a fast one for subtasks, save tokens without sacrificing quality
/plan：Enter planning mode before execution — AI maps out all files and steps first, you confirm, then it executes
Follow-up Suggestions：AI suggests 2-3 next steps after completing a task — "Add unit tests?" "Check similar files?" — click to continue
Adaptive Output Tokens ：Default 8K output, auto-escalates to 64K when truncated — no more manually tuning max_tokens
Ctrl+O Verbosity Toggle ：Switch between verbose and compact output mid-conversation — debug mode when you need it, clean mode when you don't

📋 Full changelog: https://t.co/zPMyCVIskN