kata siapa?

Develop good taste and level up everyday

[kata siapa? @kataridhan]

Sometimes, it's okay if you don't know what you're doing in life. Because everyone is totally just winging it, all the time. So, relax. Everything is gonna be fine.

[Rohit Ghumare @ghumare64]

If you want to go from AI beginner → AI builder & engineer, don’t just watch tutorials.

Build from great open-source repos.

Here are 17 free GitHub repos to learn AI/ML, LLMs, agents, RAG, math, and RL 👇🧵

[Harman @itsharmanjot]

A TEAM OF AI RESEARCHERS JUST OPEN-SOURCED THE BLOOMBERG TERMINAL FOR QUANT FINANCE.

A Bloomberg Terminal costs $25,000 per year per seat. Banks pay for thousands of them.

This thing reads every quant paper, every financial blog, every SEC filing, every arXiv preprint, and turns it into a searchable knowledge base. For free.

It's called QuantMind.

It just got accepted to the NeurIPS 2025 GenAI in Finance Workshop.

Here's what it actually does:

→ Ingests arXiv quant papers, financial news, blogs, and reports automatically
→ Parses PDFs, HTML, tables, and figures into structured knowledge
→ Tags every paper by research area and topic
→ Builds a semantic knowledge graph you can query in plain English
→ Plugs into DeepResearch, RAG, and MCP for multi-hop reasoning
→ Two-stage architecture: extract once, retrieve forever

Here's the wildest part:

The financial research industry publishes around 500 new papers and reports every single day.

Hedge funds pay six-figure salaries to junior analysts whose entire job is reading them.

QuantMind reads all of it. Tags it. Embeds it. Lets you ask it questions.

154 stars. 22 forks. 173 commits. MIT license. Python.

One honest note: this is a framework, not a magic alpha machine. You still need to know what to ask. But the "I haven't read that paper yet" excuse is officially dead.

The thing Wall Street charges $25,000 a year for is sitting on GitHub. Free.

Link in the comments.

[Harman @itsharmanjot]

10 GitHub repos so good they shouldn't be free.

1. AutoHedge

An autonomous hedge fund built in Python with four AI agents: a director generates investment theses, a quant validates them, a risk manager decides position size, and an execution agent places orders. Operates live on Solana. With 'pip install -U autohedge', you can start trading immediately.
repo → https://t.co/q22EzesLoD

2. Vibe-Trading

A trading system using a Directed Acyclic Graph (DAG) model, featuring 64 finance skills and 29 preset specialist agent swarms. Includes analysis methods like Ichimoku, Elliott Wave, SMC, Black-Scholes, full Greeks, and risk parity. Its crypto desk provides liquidation heatmaps and token unlock tracking. You can observe agents debating strategies in real time.
repo → https://t.co/LZ5CYGMC1W

3. Fincept Terminal

A Bloomberg Terminal replacement that runs on your laptop. CFA levels 1, 2, and 3 analytics. 20+ investor AI agents (Buffett, Dalio, Soros). 100+ data connectors, including Polygon, World Bank, and IMF. Bloomberg charges $24,000 a year. This is free.
repo → https://t.co/dMM1WZxrw9

4. LibreChat

Every model ChatGPT runs, plus Claude, Gemini, DeepSeek, and 20 more. Self-hosted. Native MCP support. You own the data, the history, the infrastructure. OpenAI charges $20/month to use their wrapper. This costs nothing to use your own.
repo → https://t.co/457utdZUIF

5. Open Higgsfield AI

A self-hosted cinema studio with 200+ AI models. Flux, Midjourney, Sora, Kling, Veo, GPT-4o, SDXL all in one interface. Text to image. Image to video. Cinema mode with pro camera controls. No subscription. Your data stays local.
repo → https://t.co/WHCzBSFBW4

6. Open-LLM-VTuber

A Live2D AI companion that runs offline, sees your screen, hears your voice, and never forgets. Inner thoughts are shown as a separate text layer, so you watch the reasoning happen before words come out. Pet mode floats it on your desktop. Swap the LLM in one config line.
repo → https://t.co/5XVKUPr35X

7. Claude Ads

A free Claude Code skill that runs 190 audit checks across Google, Meta, YouTube, LinkedIn, TikTok, and Microsoft Ads. 6 parallel subagents firing at once. Consolidates into a single Ads Health Score ranked by revenue impact. Agencies charge $4,000 a month for this.
repo → https://t.co/AJRfpSB7B6

8. Agentic Inbox

Cloudflare just open-sourced an email client where an AI agent reads your inbox and drafts your replies. Runs entirely on Cloudflare Workers. Each mailbox lives in its own Durable Object. Your email never leaves your Cloudflare account. One click deploys it.
repo → https://t.co/QEEMtzoliV

9. Camofox Browser

An open source headless browser that makes AI agents invisible to bot detection. Spoofs navigator properties, WebGL, AudioContext, and WebRTC at the C++ level. The browser does not look modified because it genuinely is not. Accessibility tree output drops token cost by 90%.
repo → https://t.co/95d0V3o7vO

10. Hyperframes

HeyGen open-sourced a video framework that does everything Remotion does without React, without JSX, without teaching your AI agent a new format. The agent writes HTML. The framework renders MP4. GSAP, Lottie, and Three.js all work. Same HTML always produces the same file.
repo → https://t.co/ekquvYvTNC

These are not toys. Each one replaces a paid product you're still being charged for.

Pick one. Install it. Plug it into your workflow.

100% free. 100% open source.

[Vivo @vivoplt]

Research papers you must read for AI Engineer interviews:

1. Attention is all you need (Transformers)
2. LoRA (Low rank adaption)
3. PEFT ( Parameter Efficient Fine Tuning)
4. VIT (Vision Transformers) 5. VAE (Variational Auto Encoder)
6. GANs ( Generative Adversarial Networks)
7. BERT ( Bidirectional Encoder Representation from Transformers)
8. Diffusion Models (Stable Diffusion)
9. RAG (Retrieval Augment Generation)
10. GPT (Generative Pre-trained Transformers)
11. MoE (Mixture of Experts)
12. RLHF (Reinforcement Learning from Human Feedback)
13. LLaMA (Large Language Model Meta AI)

[Priya @LoopandPixels]

If I had 6 months to become a GenAI Engineer.

I'd do this.

Stage 1: Python + Async Architecture FastAPI, asyncio, typing, Pydantic v2, event-driven design, API integration patterns.

Stage 2: Multimodal LLM Fundamentals Transformer architecture, SLMs vs LLMs, context window management, vision/audio inputs, token economics.

Stage 3: Structured Outputs + Prompt Ops JSON schema enforcement, function calling, prompt versioning, template management, few-shot optimization.

Stage 4: Advanced RAG + Knowledge Graphs Hybrid search, graph RAG, semantic reranking, metadata filtering, citation grounding, incremental indexing.

Stage 5: Agentic Workflows + Orchestration LangGraph/LlamaIndex, tool use, planning loops, multi-agent collaboration, human-in-the-loop handoffs.

Stage 6: Production GenAI Applications Streaming responses, optimistic UI, fallback models, rate limiting, cost-aware routing, session management.

Stage 7: Evaluation + Quality Assurance LLM-as-a-judge, automated eval harnesses, regression testing, hallucination metrics, RAGAS/DeepEval.

Stage 8: Inference + Infrastructure Optimization vLLM/SGLang, quantization (FP8/INT4), KV caching, speculative decoding, edge deployment, model distillation.

Stage 9: MLOps + Observability Distributed tracing, latency monitoring, cost dashboards, drift detection, CI/CD for prompts and models.

Stage 10: AI Safety + Compliance Guardrails, prompt injection defense, PII redaction, copyright checks, EU AI Act compliance, content filtering.

Stage 11: Open Source + Portfolio Ship multimodal agents publicly, write technical deep dives, record demo videos, contribute to orchestration libs.

Stage 12: Apply GenAI Engineer, AI Application Developer, LLM Infrastructure Engineer, Autonomous Systems Engineer roles.

Most people stay stuck watching tutorials.

Builders get hired.

(Bookmark it)

[Rohit Ghumare @ghumare64]

Learn Coding by playing games

1. Kubernetes
https://t.co/e2lskAhkSb

2. DevOps
https://t.co/JqMwECBqVL

3. Linux
https://t.co/8AZMxUVUTN

4. Git
https://t.co/7gn24CyQik

5. Python
https://t.co/K6KnVWxqj1

6. CSS & HTML
https://t.co/sgooWVCQ4u

7. Cybersecurity
https://t.co/mwso7Evxtj

8. Mobile Coding (like Duolingo)
https://t.co/DumVCTo4F9

9. For Complete Beginners
https://t.co/kfSqv3Daac

10. 25+ Programming Languages
https://t.co/3SeeGxFWNl

Follow @ghumare64 for more such tips!

[Suraj Sharma @suraj_sharma14]

If I had 6 months to become a GenAI Engineer.

I'd do this.

Stage 1: Python + Async Architecture FastAPI, asyncio, typing, Pydantic v2, event-driven design, API integration patterns.

Stage 2: Multimodal LLM Fundamentals Transformer architecture, SLMs vs LLMs, context window management, vision/audio inputs, token economics.

Stage 3: Structured Outputs + Prompt Ops JSON schema enforcement, function calling, prompt versioning, template management, few-shot optimization.

Stage 4: Advanced RAG + Knowledge Graphs Hybrid search, graph RAG, semantic reranking, metadata filtering, citation grounding, incremental indexing.

Stage 5: Agentic Workflows + Orchestration LangGraph/LlamaIndex, tool use, planning loops, multi-agent collaboration, human-in-the-loop handoffs.

Stage 6: Production GenAI Applications Streaming responses, optimistic UI, fallback models, rate limiting, cost-aware routing, session management.

Stage 7: Evaluation + Quality Assurance LLM-as-a-judge, automated eval harnesses, regression testing, hallucination metrics, RAGAS/DeepEval.

Stage 8: Inference + Infrastructure Optimization vLLM/SGLang, quantization (FP8/INT4), KV caching, speculative decoding, edge deployment, model distillation.

Stage 9: MLOps + Observability Distributed tracing, latency monitoring, cost dashboards, drift detection, CI/CD for prompts and models.

Stage 10: AI Safety + Compliance Guardrails, prompt injection defense, PII redaction, copyright checks, EU AI Act compliance, content filtering.

Stage 11: Open Source + Portfolio Ship multimodal agents publicly, write technical deep dives, record demo videos, contribute to orchestration libs.

Stage 12: Apply GenAI Engineer, AI Application Developer, LLM Infrastructure Engineer, Autonomous Systems Engineer roles.

Most people stay stuck watching tutorials.

Builders get hired.

(Bookmark it)

[Vlad Mihalcea @vlad_mihalcea]

As a software developer, it's very important to know Galton’s Law of Mediocrity (Regression Towards the Mean) in the context of AI.

According to Galton, as long as you are better than the average or have skills that AI cannot replicate (e.g., creativity, high agency, determination, communication, and teamwork), you are going to be fine in your career.

[priyanshu.sol @priyanshudotsol]

someone wrote a 680 page interactive book on cs algorithms

[CyrilXBT @cyrilXBT]

NVIDIA JUST MADE PAID AI APIS LOOK EMBARRASSING.

80 models.

Free inference.

No rate limits.

No data leaving your machine.

No monthly bill growing every time your usage scales.

The open source AI stack that most developers still do not know exists just got very hard to ignore.

Here is what you can run right now for free.

DeepSeek V3. MiniMax M2.7. Kimi 2.5. GLM 5.1. GPT OSS 120B.

Models that compete with the best paid APIs on the market.

Running on infrastructure you control completely.

The builders who figured this out 6 months ago are not paying $300 a month in API bills.

They are running agents at scale, experimenting without cost anxiety, and shipping products with margins that closed competitors cannot match.

The gap between people who know this stack exists and people still paying per token is not a preference gap.

It is a business model gap.

Three lines of setup.

Grab your free API key at https://t.co/EYXopCfCOV

Set your base URL to https://t.co/pP5pscJM91

Pick a model and start building tonight.

The infrastructure excuse just ran out.

Follow @cyrilXBT for the exact local AI setups, open source stacks, and free inference workflows I use to build without limits.

[blue @bluewmist]

No porn + avoid sugar + drink water + hit the gym + party less + Avoid drama + Dress well + Make money + Smell nice + Talk less

= Magic pill to life.

[Bardan - Digital Marketer @bardanslm]

to the point aja, di pernikahan:

yg suami butuhkan:
- segs
- waktu sendiri kadang2
- diberi pengertian dan disupport
- enggak dibanding2in
- didukung tanpa ditekan terlalu keras
- kepercayaan

yg istri butuhkan:
- uang
- suami inisiatif bantuin kerjaan rumah dan anak tanpa diminta
- suami dengerin curhat istri
- partner diskusi 2 arah dan terbuka, bukan 1 arah
- kesetiaan

ya tentu bisa sama/gantian kebutuhannya tergantung kondisi dan orangnya, tapi kira2 itu bukan ya gaes???

[The Vital Path @PathOfVitality]

You are bored because you are not doing side quests.

Life is not just work and lying in bed doing nothing.

Here are 50 side quests every man should complete:

[Sim @simpreetkaur_19]

Research papers you must read for AI Engineer interviews:

1. Attention is all you need (Transformers)
2. LoRA (Low rank adaption)
3. PEFT ( Parameter Efficient Fine Tuning)
4. VIT (Vision Transformers)
5. VAE (Variational Auto Encoder)
6. GANs ( Generative Adversarial Networks)
7. BERT ( Bidirectional Encoder Representation from Transformers)
8. Diffusion Models (Stable Diffusion)
9. RAG (Retrieval Augment Generation)
10. GPT (Generative Pre-trained Transformers)

[ISM ☀️ @ism_sol]

lu pada kadang kena rate limit kan kalo butuh data dan indicator realtime suatu asset

gua nemu nih Github repo tradingview yg bisa lu colok ke ai agent lu 🫵🏻
gua taro ya di komen

enaknya tuh :
> tanpa api key
> bisa lu bikin scan tiap 1 menit sekali
> dapet data saham indo, saham us, crypto ya intinya semua ticker yg ada di trading view
> gaperlu takut rate limit
> open source 100%

intinya kalo lu bikin mau bikin ai agent enak pake ini.

[nikhil @jhanikhil]

there are only like 5-6 tricks in systems

- divide and conquer
- caching
- indirection
- batching
- redundancy
- lazy eval

and then you apply arbitrarily complicated variations of these in clever ways to build whatever you want

[Bagas Naufal Insani @BagasN7]

Riil, kalo liat dari hasil survei nya skillsoft juga sertifikasi profesional yg bisa ngasih income paling gede itu ga jauh dari Data sama seputar Cloud.

Data bisa data engineer atau data analyst.

Cloud bisa DevOps, SysOps, Security Engineer, dsb.

Bahkan di lowongan Asia pasifik aja bisa nyentuh 3 digit perbulan atau 5 figures annual. *TnC berlaku ya

Apalagi kalo bisa nembus lowongan Worldwide khususnya US, Europe, sama Middle East.

Jadi buat yg masih meraba mau karir kemana, dari sekarang bisa coba mapping mau role apa, bisa dapat pengalaman real world nya gimana kalo di lowongan Indo, terus sisihin uang, tenaga dan waktu buat incar sertifikasi professional kayak di list ini.

Sisanya nanti tinggal serahkan aja sama yg maha pemberi rezeki deh. 😇

[Lambe Saham @LambeSahamjja]

Guys, ini gilaa sihh
Suster Natalia.
Perempuan yang tidak menikah.
Tidak punya harta pribadi.
Mengabdikan seluruh hidupnya untuk gereja dan umatnya di Labuhanbatu, Sumatera Utara.

Dan sekarang dia harus menanggung beban Rp28 miliar yang raib bukan uangnya sendiri
tapi uang 1.900 jiwa umat yang dia jaga amanahnya.

bahkan dia bilang
ke teman dia yang suster juga
dia akan masuk penjara.

dia cerita
Setiap kali ketemu umat yang sederhana itu,
saya selalu katakan:
mari, masa depan anak-anakmu melalui menabung.
Tapi sekarang masa depan mereka itu hancur di tangan saya.

Kronologi yang perlu semua orang pahami:
Credit Union Paroki Aek Nabara koperasi simpan pinjam di bawah naungan gereja sudah berjalan 45 tahun tanpa masalah.
Umat menabung perak demi perak.
Untuk sekolah anak.
Untuk biaya sakit.
Untuk masa depan.

Total yang terkumpul dan ditempatkan di deposito: Rp28 miliar lebih dari 1.900 anggota.

Di 2019 Andi Hakim Febriansyah
Kepala Kas BNI Unit Aek Nabara
mendatangi pengurus CU.
Menawarkan produk bernama BNI Deposito Investment
dengan bunga 8% per tahun.
Lebih tinggi dari deposito biasa.

Pengurus percaya.
Karena siapa yang tidak percaya kepada kepala kas bank negara yang datang dengan seragam resmi, ID card BNI, dan pick-up service resmi yang sudah berjalan sejak 2015?

Tujuh tahun berjalan.
Bunga masuk rutin setiap bulan.
Tidak ada masalah.

Sampai Desember 2025 dan semuanya mulai runtuh:

CU mengajukan pencairan Rp10 miliar untuk pinjaman ke anggota.
Bertahap minta Rp2 miliar dulu.

Januari 2026 tidak cair.
Februari 2026 tidak cair.

5 Februari Suster Natalia panggil Andi.
Andi bilang besok.
Besok tidak cair.
Andi minta semua bilyet deposito untuk pembaruan. Suster menyerahkan semuanya karena percaya.

Sore hari Andi sudah di jalan ke Medan katanya cuti.

Lalu 23 Februari bukan Andi yang datang.
Tapi kepala kas baru.
Dengan kalimat yang mengubah segalanya:
Per hari ini saudara Andi Hakim Febriansyah bukan pegawai BNI lagi.
Dan produk yang ditawarkan itu bukan produk BNI.

Suster Natalia pingsan lima menit.
Yang lebih ngeri dari hilangnya uang itu:
Bilyet deposito yang dipegang Andi dibakar. Sengaja. Supaya tidak ada barang bukti.

Tapi Andi salah hitung.
Satu bilyet tersimpan di tangan pastor lain yang kebetulan tidak ada di tempat saat pengambilan.
Satu bilyet itu yang menjadi bukti bahwa semua ini nyata.

Andi sudah menyiapkan skenario dari jauh hari. Tanggal 23 Februari itu hari yang sama dia ambil semua bilyet dia sudah mengajukan pengunduran diri. Dan dua hari kemudian dia terbang ke luar negeri bersama istrinya lewat Bali ke Australia, lalu ke New Zealand.

Sambil cuti dia masih angkat telepon Suster Natalia. Masih bilang "aman, Suster."
Masih janjikan pencairan.

Setelah red notice diterbitkan oleh Interpol dan Australian Federal Police Andi kembali ke Indonesia 30 Maret 2026 dan ditangkap di Kualanamu.

Di dalam pemeriksaan dia mengakui semua perbuatannya.
Uangnya?
Dipakai untuk sport center, kafe, mini zoo, tanah, dan berbagai aset yang kini sedang dilacak dalam proses TPPU.

Dan sekarang masuk ke bagian yang paling mengkhawatirkan:

BNI melakukan verifikasi internal sendiri.
Tanpa transparansi.
Tanpa melibatkan korban dalam proses.
Hasilnya: BNI bersedia mengganti Rp7 miliar.

Dari Rp28 miliar lebih.
Dan pada 26 Maret 2026 tanpa persetujuan CU-PAN BNI mentransfer Rp7 miliar itu ke rekening korban secara sepihak.
Seolah dengan mentransfer itu kasus selesai.

Kuasa hukum CU-PAN dari Gani Djemat & Partners menolak keras.
Karena:
Berdasarkan prinsip Vicarious Liability perusahaan bertanggung jawab atas tindakan pegawai yang dilakukan dalam kapasitas jabatannya.

Andi beroperasi dengan ID card BNI, jabatan BNI, fasilitas pick-up service BNI, dan atas nama BNI selama tujuh tahun.
Ini bukan tindakan pribadi yang kebetulan dilakukan oleh orang yang bekerja di BNI. Ini tindakan yang bisa terjadi karena dia adalah BNI di mata korban.

POJK Nomor 22 Tahun 2023 Pasal 10 ayat 1 juga menegaskan: pelaku usaha jasa keuangan wajib bertanggung jawab atas kerugian konsumen akibat kesalahan pegawainya.
Tidak ada klausul kecuali kalau pegawainya nakal.

Dan respons BNI yang paling menyakitkan menurut korban:

Enam kali mediasi.
Satu kali aksi damai.
Sepanjang itu tidak satu pun pejabat BNI dari kantor cabang atau wilayah yang mengucapkan kata "maaf" atau kami prihatin kepada korban.

Yang datang dari pihak BNI hanya satu permintaan berulang: Berikan kami bukti pendukung.
Padahal semua data transaksi ada di sistem BNI sendiri.
Semua perpindahan uang dari kas lancar ke rekening Andi tercatat di rekening koran BNI.
Bukan di tangan korban.

Baru Wakil Menteri BUMN yang mengundang korban dan itulah pertama kalinya ada pejabat yang mengucapkan kata permohonan maaf dan rasa prihatin.

Satu hal yang tidak bisa diabaikan:
Suster Natalia sekarang punya utang pribadi ke beberapa orang.
Karena ada anggota CU yang butuh uang untuk berobat yang tidak bisa dia biarkan meninggal di rumah sakit sementara dana CU tidak bisa diakses.

Dia yang tidak punya harta pribadi meminjam uang untuk membayar tagihan rumah sakit umatnya.

"Saya tidak bisa biarkan umat meninggal di rumah sakit, Pak."

BNI adalah bank BUMN.
Bank milik negara.
Diawasi oleh OJK.
Dijamin kepercayaannya oleh nama negara Indonesia.

Dan di bawah namanya selama tujuh tahun seorang kepala kas menjalankan skema penipuan yang menyedot uang 1.900 jiwa umat gereja yang menabung perak demi perak untuk masa depan anak-anak mereka.

BNI tidak bisa menyebut ini hanya masalah oknum lalu cuci tangan dengan transfer Rp7 miliar yang tidak transparan prosesnya.

Karena korban bukan menyimpan uang kepada Andi Hakim.
Korban menyimpan uang kepada BNI.
Dan BNI harus mengembalikannya penuh tanpa pengecualian.

Kalau tidak ini bukan hanya kasus kriminal biasa.
Ini adalah konfirmasi bahwa di negeri ini orang miskin yang menabung untuk masa depan anaknya bisa kehilangan segalanya karena sistem yang seharusnya melindungi mereka justru membiarkan hal ini terjadi selama tujuh tahun.

[Prompter @PromptLLM]

Absolute banging advice from Claude

[Mustafa @oprydai]

one-man robotics empires are a bad idea.

hardware isn’t software. you can’t just solo your way through it.

where it breaks:
• hardware complexity → mechanics, electronics, control, manufacturing. too many domains for one person to handle deeply.
• iteration speed → software ships in minutes. hardware takes days, weeks, sometimes months per loop.
• capital + supply chain → parts, fabrication, logistics. you’re dealing with the real world, not just code.
• testing in reality → things fail unpredictably. debugging isn’t just logs, it’s physics, noise, wear.
• scaling → one prototype is easy. producing 1000 units is a completely different game.

software lets a single person build leverage fast.
hardware demands teams, processes, and infrastructure.

romanticizing the “solo robotics empire” is misunderstanding the medium.

build alone to learn.
build teams to win.