Ruli

154 posts

Ruli banner
Ruli

Ruli

@Total_Slim

Full-stack engineer building with Codex, Claude & AI agents. Sharing practical systems, experiments, and IHSG trading lessons. Indonesia 🇮🇩

Katılım Ekim 2018
239 Takip Edilen11 Takipçiler
Sabitlenmiş Tweet
Ruli
Ruli@Total_Slim·
I've been using Claude Code for free on personal projects. No subscription. No 5-hour limit. Here's my exact setup🧵
English
2
0
3
321
Ruli
Ruli@Total_Slim·
@NickDevFE Running it inside the group chat is how a five-minute test becomes everyone’s afternoon 😂
English
1
0
1
24
Ruli
Ruli@Total_Slim·
@azamuddin91 Kalau model offline makin murah, bagian mana yang menurutmu tetap paling mahal: hardware, token, atau integrasinya?
Indonesia
0
0
0
129
Muhammad Azamuddin
Muhammad Azamuddin@azamuddin91·
Gak usah takut tapi juga jangan nyepelein Saya melihat AI itu udah "mereplace" banyak headcount Buat yg belum kena dampak bersyukur aja Buat yg mikir AI bakal mahal kalau subsidi habis, boleh aja, tapi inget pasti ada yg berusaha bikin token jadi murah
Ali 👊🏻@mhuseinali

Gausah takut sama AI guys. Hari ini cost AI murah karena masih disubsidi sama model learning. Lama kelamaan banyak perusahaan yang sadar kalau usage AI dan iterasinya (karena perlu prompt-testing dan stress-testing) harga tokennya jauh lebih mahal daripada ngegaji satu senior. Udah kejadian di Amazon wkwk.

Indonesia
4
9
112
6.8K
Ruli
Ruli@Total_Slim·
@DavidNyann The mosquito-catching pig is such a weirdly good premise 😂
English
0
0
0
28
Ruli
Ruli@Total_Slim·
@Nasikin_MA Sepatunya lebih sering keluar-masuk daripada kita 😭
Indonesia
0
0
0
1.3K
A En Inisialku
A En Inisialku@Nasikin_MA·
Diproduksi di Indonesia, Diimpor oleh Asics Indonesia. Ini maksudnya dibuat di Indonesia, lalu diekspor, kemudian diimpor masuk lagi ke Indonesia, atau bahan diimpor kemudian dibuat di Indonesia dan dipasarkan di Indonesia? 🤔
A En Inisialku tweet media
Indonesia
325
610
7.4K
1M
Ronnie Adkins
Ronnie Adkins@RonnieAdkins·
“What do you have to hide?” Nothing. Everything. Whatever I want. Maybe I don’t want the Government knowing about my lawn care routine. It’s a family secret passed down for generations.
English
50
132
1.8K
20.5K
Ruli
Ruli@Total_Slim·
@hanifproduktif Mas, pas training biasanya yang paling susah justru login, atau nentuin use case pertama?
Indonesia
1
0
0
57
Hanif | AI For Productivity
Hanif | AI For Productivity@hanifproduktif·
Always remember this image: kalian user X yg udah ngulik Agentic AI itu ada di titik merah. Sedangkan sebagian besar orang Indo ada di titik abu2 dan hijau Kalian ngerasa semua orang udah melakukannya karena emang timeline kalian isinya orang2 yg ngulik. Tapi coba kalian cek di lapangan Kerjaan gue sehari hari tuh ngisi training Generative AI di perusahaan2 Indo dan kalo lo tau di lapangan bakalan kaget sama faktanya. Bahkan untuk sekedar login ke ChatGPT orang banyak banget yang belum bisa 😅 Ini realita di banyak company (especially di luar Jabodetabek) Knowledge-gap nya jauh banget cik
Hanif | AI For Productivity tweet media
Divine Glare@divineglare

@perogeremmer Kalau di X aku ngerasa ketinggalan terus, tapi di real life ngerasa frontier wkwkwk. Temenku vibe coding, bukan pakai agent kek opencode dan semisalnya, tapi chat terus copy paste hasil kodenya 🗿 Pas kutanya, dia cuma jawab "biasanya juga gini". Memang di X ini bubble.

Indonesia
19
10
111
7.4K
Ruli
Ruli@Total_Slim·
@stellarprtcol Gue pakai pola ini juga. Agent baca vault dulu sebelum kerja, terus hasil dan learning barunya ditulis balik ke sana.
Indonesia
0
0
0
41
stellar protocol
stellar protocol@stellarprtcol·
Gue baru ngeh kenapa ada orang yang hasil AI-nya jauh lebih ngacir dibanding yang lain. 🤯 Ternyata bukan karena mereka punya prompt rahasia 1000 baris... 👀 Yang bikin beda justru context. Bayangin AI udah ngerti: 🧠 cara lu mikir 🎯 tujuan jangka panjang lu ✍️ gaya nulis favorit lu ⚠️ kesalahan yang sering lu ulang Jadi gak perlu lagi buang 10 chat pertama cuma buat jelasin diri sendiri. 😅 Yang lebih keren, setiap project punya context sendiri. 📂 Riset 📂 Draft 📂 Feedback 📂 Task AI cuma baca yang relevan, bukan seluruh "kehidupan digital" lu. Dan kalau ada pekerjaan yang sering lu ulang... 📧 Email 📝 Meeting notes 📋 PRD 🐛 Bug report Tinggal jalanin workflow yang sama berulang kali. Cepat, konsisten, minim ribet. 🚀 Bagian paling mindblowing menurut gue? 🤯 Knowledge base lu bisa terus dirapihin: 🔍 nyari informasi yang udah usang 💡 ngangkat ide yang terlupakan 🔗 nyambungin catatan yang berkaitan ⚡ nemuin kontradiksi yang lu lewatkan Semakin lama dipakai, sistemnya malah makin pinter. Dan semua itu cuma berbasis file Markdown. 📄 Gak terkunci di satu platform. Ganti model AI besok pun, knowledge lu tetap ikut jalan. 🔥 Makanya gue mulai percaya... Masa depan AI bukan soal siapa yang punya prompt paling keren. 👀 Tapi siapa yang punya context terbaik. 🚀📈 x.com/undefinedKi/st…
Yarchi@undefinedKi

x.com/i/article/2067…

Indonesia
10
20
140
6.8K
Ruli
Ruli@Total_Slim·
@LambeSahamjja Perlu batas keluar, tapi nggak selalu stop loss persentase. Buat investasi, trigger-nya bisa thesis break: laba turun struktural, utang memburuk, governance rusak, atau valuasi awal ternyata salah. Horizon panjang bukan berarti tanpa invalidation.
Indonesia
0
0
0
149
Lambe Saham
Lambe Saham@LambeSahamjja·
kalo invest jangka panjang apakah perlu stop loss?
Lambe Saham tweet media
Indonesia
35
4
80
11.2K
Ruli
Ruli@Total_Slim·
@sungchi This is the real productivity shift for me: a passing idea can become something playable before the motivation disappears. Small experiments matter because they train the whole idea-to-feedback loop.
English
0
0
1
288
sungchi
sungchi@sungchi·
어제 운동하다가 구슬치기가 생각나서 Codex랑 같이 간단하게 만들어봤습니다. 재밌게해주세요. plan9.kr/goosl/
한국어
40
196
1.9K
188.5K
Ruli
Ruli@Total_Slim·
@explosss1ve Exactly. MCP compresses the procedural middle, but it does not remove the need for someone who can see why a silhouette, fold, or asymmetry feels alive. Automation makes taste more visible, not less valuable.
English
0
0
0
236
explos1ve
explos1ve@explosss1ve·
MCP CAN DRIVE BLENDER ALL DAY AND STILL NOT SCULPT ONE SPINE ON THIS THING a model through MCP writes bpy, blocks out a base mesh, sets the scene, spins the camera, hits render, reads its own frame. that whole loop got cheap. the scaffold of a creature is a conversation now. but look at what's actually on screen. the layered spines. the skin that folds like it has muscle under it. the asymmetry that reads as alive instead of generated. that's sculpt work. hours in the brush, by someone who knows how a jaw should sit and where a ridge breaks. no prompt makes organic detail feel earned. MCP gives you the blockout, the lighting, the turntable. it doesn't give you the hand that pushed every spike. the technical middle of 3d collapsed. the sculpt, the taste, the thing that makes a monster look like it could bite, still didn't move.
explos1ve@explosss1ve

x.com/i/article/2075…

English
60
107
1.3K
177.4K
Ruli
Ruli@Total_Slim·
@SarjanaEksu Kalau benar dapat kontrak, cerita CBRE-nya bukan cuma next PANI, tapi next test apakah pasar sudah pricing in proyeknya atau belum
Indonesia
0
0
0
698
Ruli
Ruli@Total_Slim·
@TradingDiary2 Negara ramah, mata uangnya yang nggak diajak kompromi 😭
Indonesia
0
0
0
259
Ruli
Ruli@Total_Slim·
@xallyange Yang bikin capek bukan target IHSG 4700-nya, tapi goalpost yang ikut jalan setiap kali market nggak sesuai prediksi 😅
Indonesia
1
0
0
415
kursi
kursi@xallyange·
Ada orang nanya ke influencer investasi yang selalu bilang IHSG ke 4700. Bukan dijawab dengan benar, malah digeser gawang nya. Kalian ada yang follow king of harmonic pattern gak?
kursi tweet mediakursi tweet media
Indonesia
22
0
30
17.7K
Ruli
Ruli@Total_Slim·
@mirochill At $0.01, the interesting comparison is not only output quality, but how stable the result is across prompt variants. Cheap plus consistent is what changes the workflow.
English
0
0
3
624
Mirochill
Mirochill@mirochill·
🚨 Test de Deepseek V4 Pro (GA) ! Le modèle devrait être proche du niveau Kimi K3 (lui-même proche de Fable 5). Ce prompt n’a coûté que 0,01 $, le modèle devrait être le plus performant par rapport au prix. Je partagerai bientôt d’autres résultats, abonnez-vous pour ne pas les louper ! Thanks at @Waguri_Kaoruko8 for generation
Mirochill@mirochill

Nouveau test de Kimi K3... 👀 Ça ne semble pas incroyable au premier regard, mais c’est l’un des meilleurs outputs sur ce prompt. Il a utilisé les mêmes mécanismes originaux que Fable 5 m’a produits (non précisé dans le prompt), on dirait exactement le même résultat. Il est très probable qu’il soit distillé à partir de Fable, et pourtant il n'est pas moins performant, il lui arrive même d’être meilleur ! Abonnez-vous pour d’autres tests et comparaisons 🫡

Français
18
19
346
38.5K
Ruli
Ruli@Total_Slim·
@kyzoroX Open weights are technically open, but a 1.4TB model is still practically closed to most builders. Hardware is becoming the new API pricing.
English
1
0
1
35
KyzoroX
KyzoroX@kyzoroX·
2.8 trillion parameters. Open weights July 27. Here's the number nobody's doing: you can't run it. Not on anything. K3 at Q4 is roughly 1.4TB of weights. Every machine in the local AI conversation, measured against that: DGX Spark, 128GB — 11x short AMD Strix Halo, 128GB — 11x short Mac Studio M3 Ultra, 512GB — 3x short Four Mac Studios clustered, ~$40,000 — still short Look at the demos going around today. 3D cities and satellite trackers rendering live in a browser tab, and the reaction is always the same: this is insane, it's running right here. It isn't. It's running in Moonshot's datacenter. The browser is a window. "Open weights" used to mean "you can run it." K3 is where those two quietly separated. The weights will be free and downloadable by anyone on the 27th. The hardware to load them starts north of a rack. DeepSeek at 1.6T already sat outside every consumer machine. K3 at 2.8T doubles the distance. What open weights actually buy you at this size: someone else runs it and charges you less. $15 per million output tokens against $50 for the closest US frontier model. That's the product — not local, cheaper cloud.
Kimi.ai@Kimi_Moonshot

Introducing Kimi K3: Open Frontier Intelligence 🔹 2.8 Trillion Parameters, 1 Million Context, Native Multimodal 🔹 Kimi Delta Attention enables up to 6.3x faster decoding in million-token contexts 🔹 Attention Residuals deliver ~25% higher training efficiency at <2% additional cost 🔹 Built for long-horizon agentic coding and self-evolving workflows Kimi K3 is now live on on Kimi.com, Kimi Work, Kimi Code, and the Kimi API. Open Weights by July 27, 2026. 🔗 API: platform.kimi.ai 🔗 Tech blog: kimi.com/blog/kimi-k3

English
4
3
18
4.1K
Ruli
Ruli@Total_Slim·
@WOLF_of_IHSG BBCA ini bukan saham sejuta umat lagi, sudah jadi ketua RT portofolio retail 😭
Indonesia
1
0
0
91
DiaryTrader
DiaryTrader@WOLF_of_IHSG·
Saham sejuta umat kalau perdana beli saham dgn modal lumayan😀 Yg portonya up ratusan million pasti minimal ada punya king $BBCA ini Kalau dulu: UNVR Pake kampanye : belilah saham yang prduknya kita gunakan sehari-hari 🤣
Mintel 🐤@wargaritelunyu

@WOLF_of_IHSG Minimal bca, cuma mahal sih

Indonesia
4
0
11
2.1K
mora
mora@themamoraguy·
REVIU KELAS EDUKASI PGK BATCH 2 #PGKStrikesAgain Pertama2 saya ingin mengucapkan terima kasih sebesar2nya untuk mas Al @stockmapping selaku Kuncen PGK yang telah memberikan saya kesempatan untuk dapat mengikuti kelas edukasi PGK Batch 2 ini.
Indonesia
1
0
1
243
Gopal
Gopal@MYasfika·
Rangkuman Drama Buku hidden flow of money versi meme Thread
Indonesia
73
119
405
69.3K