Jupe.base.eth | 劉

[tempo.co @tempodotco]

Prabowo: Yang Merasa Indonesia Suram, Cari Negara Lain

[433 @433]

𝑻𝒉𝒆 𝒂𝒖𝒓𝒂 𝒌𝒊𝒏𝒈.

[diva @divaagurlxw]

As an AI Engineer. Please learn

>Harness engineering, not just prompt engineering

>Context engineering, not just long prompts

>Prompt caching vs. semantic caching tradeoffs

>KV cache management, eviction, reuse, and memory pressure at scale

>Prefill vs. decode latency and why they optimize differently

>Continuous batching, paged attention, and throughput optimization

>Speculative decoding vs. quantization vs. distillation tradeoffs

>INT8, INT4, FP8, AWQ, GPTQ, and when quantization hurts quality

>Structured output failures, schema validation, repair loops, and fallback chains

>Function calling reliability, tool contracts, argument validation, and idempotency

>Agent guardrails, loop budgets, tool budgets, and termination conditions

>Model routing, graceful fallback logic, and degraded-mode UX

>RAG architecture: chunking, embeddings, hybrid search, reranking, and freshness

>Retrieval evals: recall, precision, grounding, attribution, and citation quality

>Evals: golden sets, regression tests, adversarial tests, LLM-as-judge, and human evals

>LLM observability as a first-class discipline: traces, spans, tokens, latency, errors, and drift

>Cost attribution per feature, workflow, tenant, and user journey not just per model

>Safety engineering: prompt injection defense, data leakage prevention, and permission boundaries

>Multi-tenant isolation, cache safety, and cross-user context contamination prevention

>Fine-tuning vs. in-context learning vs. RAG vs. distillation and when each is the wrong tool

>Latency, quality, cost, and reliability tradeoffs across the full inference stack

>Production failure modes: hallucinated tool calls, malformed JSON, stale retrieval, runaway agents, and silent eval regressions

[Jupe.base.eth | 劉 @0xjupe]

@MrBeast 2356315677

[David Ch @chhddavid]

ITS SO OVER

Claude just killed Lovable forever

this is terrifying……

[Ruben Hassid @rubenhassid]

Claude is offering 13 AI courses & certificates.

All free. Here are all 13 links (+ my own guides):

1. Go to each link below. Enroll. It's free.
2. But honestly? My newsletter covers it better.
3. I'll explain at the end. Start with the official ones:

---

1 - Claude 101. Learn Claude for everyday work.
↳ https://t.co/OvBmlvnVqL

2 - AI Fluency: Frameworks & Foundations.
↳ https://t.co/cObZdwCmXP

3 - Introduction to Agent Skills.
↳ https://t.co/wZsD0PxJwi

4 - Building with the Claude API.
↳ https://t.co/RcCbfNjlzz

5 - Claude Code in Action.
↳ https://t.co/y29CC0GBaN

6 - Intro to Model Context Protocol.
↳ https://t.co/Qnrn0NHxyI

7 - MCP: Advanced Topics.
↳ https://t.co/0S5f4kESzG

8 - AI Fluency for Students.
↳ https://t.co/YIOopqo7WB

9 - AI Fluency for Educators.
↳ https://t.co/54oLlYjLGD

10 - Teaching AI Fluency.
↳ https://t.co/fHdgs6uNDM

11 - AI Fluency for Nonprofits.
↳ https://t.co/xFKngNz09m

12 - Claude with Amazon Bedrock.
↳ https://t.co/dV6xi7UiaL

13 - Claude with Google Cloud's Vertex AI.
↳ https://t.co/MEzAxODPW4

---

Official courses are good. But they're theoretical.
I wrote how-to guides that show you what to do.

Here's how to master Claude (for free):

1. Start here: https://t.co/jw2qdIcjnh
☑ The basics of Claude.
☑ How to prompt it the right way.
☑ The different types of Claude to master.

2. Move to Cowork: https://t.co/uWTpOI3Woc
☑ The more advanced Claude is Claude Cowork.
☑ How to prompt it and set it up properly.
☑ It's a long process. But worth every minute.

3. Set up Claude for teams: https://t.co/qxlcqhf8bM
☑ Setting up Claude for teams is different.
☑ This is the easiest 5-day plan I could find.
☑ 5 steps so your team runs on Claude in a week.

4. Use Claude Skills: https://t.co/jT4uB5Bdjw
☑ Stop prompting, build your first skill.
☑ 7 favourite hacks of Claude Skills.
☑ Access Claude's team skills.

5. Claude Computer: https://t.co/TxYuHPjgbV
☑ Access Claude Computer.
☑ Use cases of Claude Computer.
☑ Schedule tasks with Claude.

6. Claude Code: https://t.co/WYZd5ltVMW
☑ English is the new code.
☑ Code 100x faster.
☑ Prompt Claude Code the right way.

7. Bonus (to go even deeper).
☑ Claude for Excel.
☑ Claude interactive charts.
☑ How to move from ChatGPT to Claude.

---

All of this is free. Here's how to get it:
1. Go to https://t.co/psB7XxB2Y4. Add your email.
2. A pop-up will ask you to pay. Do not pay.
3. Open my welcome email & enjoy the free guides.

431,000+ people read it weekly. Join them.
♻️ Repost this so others get free AI education.

[CyrilXBT @cyrilXBT]

INSTEAD OF WATCHING NETFLIX TONIGHT.
Spend 1 hour with this.

Claude AI FULL COURSE that teaches you how to BUILD and AUTOMATE anything.

The people who watch this tonight will wake up tomorrow with a skill that most people will not have in 2 years.

The people who skip it will still be watching Netflix next year wondering why nothing in their life has changed.

Your call.

[Mario Nawfal @MarioNawfal]

🚨🇮🇩BREAKING:

Indonesia has issued a tsunami advisory warning after a powerful 7.6 magnitude earthquake struck the Maluku Sea.

Malut and Sulut provinces, along with parts of the Maluku islands and North Sulawesi, are under the advisory.

Authorities are urging residents in coastal areas to stay alert and prepare for possible waves.

Source: TravelGov, @WorldEQLocator

[MickΞy M @mickeymrdt]

Gila sih, ini bener-bener plot twist paling insane di industri tech tahun ini.

✨Bayangin, perusahaan sekelas Anthropic yang bangun Claude yang notabene saingan berat OpenAI malah kena blunder yang sifatnya human error banget cuy ckckck

Nih, gue ceritain kronologinya biar lo tetep update tapi nggak pusing sama istilah teknisnya:

1. The 4 AM Blunder

Jadi jam 4 pagi, tim Anthropic nge-push update buat tool mereka namanya Claude Code. Eh, ada satu file buat debugging (buat nyari error) yang lupa dihapus dan ikut ke-upload. Masalahnya, file itu isinya 512.000 baris kode rahasia mereka. Itu ibarat lo nggak sengaja nge-share resep rahasia Krabby Patty ke grup WhatsApp komplek.

2. Viral dalam Itungan Menit
Ada researcher namanya Chaofan Shou nemuin file ini, terus langsung di-post di X (Twitter). Dan Boom! 21 juta orang ngelihat. Sebelum tim Anthropic bangun tidur dan sadar kalau mereka "telanjang" di internet, kode itu udah di-downloaddan di-copy ribuan orang ke mana-mana.

3. The "Genius" Move dari Korea
Pas Anthropic mulai panik dan kirim surat ancaman (DMCA takedown) ke mana-mana buat hapus kodenya, muncul seorang developer dari Korea namanya Sigrid Jin. Dia ini user Claude paling loyal sedunia (katanya pake 25 miliar token setahun!).
Karena takut kena masalah hukum tapi pengen kodenya tetep ada, dia ngelakuin hal paling mad lad:

✅Dia nulis ulang seluruh kode itu pakai bahasa pemrograman Python dari awal sebelum matahari terbit.

✅Logikanya gini: Kalau lo cuma copy-paste, itu melanggar hak cipta. Tapi kalau lo tulis ulang pakai bahasa lain dengan gaya lo sendiri, itu dihitung karya kreatif baru. Jadi tim legal Anthropic nggak bisa nyentuh dia!

4. Bintang Baru di GitHub
Proyek "re-write" dia yang dinamain claw-code langsung meledak. Cuma dalam waktu singkat, dapet puluhan ribu starsdi GitHub (ini kayak dapet jutaan likes tapi buat anak koding). Dia bahkan nulis ulang lagi pake bahasa Rust biar makin kenceng performanya.

5. Permanen dan "Undercover" Fail
Sekarang kodenya udah di-upload ke platform decentralized yang nggak bisa dihapus sama siapa pun. Basically, rahasia dapur Anthropic udah jadi milik publik selamanya.

Stay Ciakakakaka😁😁

[MickΞy M @mickeymrdt]

✨Google baru aja ngelakuin sesuatu yang gila ciakakakak😁😁

✨Mereka ngeluarin Sertifikat AI Profesional dan itu bener2 GRATIS cuy

💎Yang lo dapet nih:
→ 7 kursus terstruktur
→ 20+ proyek langsung
→ Prompting, analisis data, coding, konten
→ Bonus 3 bulan Google AI Pro
Selesai dalam seminggu.

Ini tuh Level training yang dipakai Walmart & Deloitte.

Dulu harganya Rp30–80 juta di bootcamp. Sekarang nol rupiah gilee

✅Lo bisa cobain disini
https://t.co/lfpRtv9ixl

[Watcher.Guru @WatcherGuru]

JUST IN: Super Micro Computer $SMCI crashes 27% after its co-founder was charged with smuggling $2.5 billion worth of Nvidia AI chips to China.

[Rohit Singh | Mr. Chartist @Mr_Chartist]

📂 Fundamental Analysis
┃
┣ 📂 Business Understanding
┃ ┣ 📂 What Company Does
┃ ┣ 📂 How It Makes Money
┃ ┣ 📂 Product Mix
┃ ┣ 📂 Industry Position
┃ ┗ 📂 "Simple business, better clarity"
┃
┣ 📂 Revenue Analysis
┃ ┣ 📂 Sales Growth
┃ ┣ 📂 Segment Growth
┃ ┣ 📂 Geographic Mix
┃ ┣ 📂 Pricing Power
┃ ┗ 📂 "Growth should be real"
┃
┣ 📂 Profitability Analysis
┃ ┣ 📂 Operating Profit
┃ ┣ 📂 Net Profit
┃ ┣ 📂 EBITDA Margin
┃ ┣ 📂 PAT Margin
┃ ┗ 📂 "Margin tells quality"
┃
┣ 📂 Return Ratios
┃ ┣ 📂 ROE
┃ ┣ 📂 ROCE
┃ ┣ 📂 ROA
┃ ┣ 📂 Asset Turnover
┃ ┗ 📂 "Capital efficiency matters"
┃
┣ 📂 Balance Sheet Check
┃ ┣ 📂 Debt to Equity
┃ ┣ 📂 Borrowings Trend
┃ ┣ 📂 Reserves Growth
┃ ┣ 📂 Working Capital
┃ ┗ 📂 "Weak balance sheet, weak sleep"
┃
┣ 📂 Cash Flow Analysis
┃ ┣ 📂 Operating Cash Flow
┃ ┣ 📂 Free Cash Flow
┃ ┣ 📂 CFO vs PAT
┃ ┣ 📂 Capex Trend
┃ ┗ 📂 "Cash is harder to fake"
┃
┣ 📂 Shareholding Pattern
┃ ┣ 📂 Promoter Holding
┃ ┣ 📂 Promoter Pledge
┃ ┣ 📂 FII Holding
┃ ┣ 📂 DII Holding
┃ ┗ 📂 "Watch who is increasing"
┃
┣ 📂 Management Quality
┃ ┣ 📂 Promoter Background
┃ ┣ 📂 Capital Allocation
┃ ┣ 📂 Guidance Quality
┃ ┣ 📂 Governance Record
┃ ┗ 📂 "Numbers are good, intent should also be good"
┃
┣ 📂 Competitive Advantage
┃ ┣ 📂 Brand Strength
┃ ┣ 📂 Cost Advantage
┃ ┣ 📂 Distribution Reach
┃ ┣ 📂 Entry Barriers
┃ ┗ 📂 "Moat or just marketing?"
┃
┣ 📂 Industry Analysis
┃ ┣ 📂 Sector Tailwinds
┃ ┣ 📂 Industry Cycle
┃ ┣ 📂 Demand Drivers
┃ ┣ 📂 Competition Intensity
┃ ┗ 📂 "Good company, bad industry is trouble"
┃
┣ 📂 Valuation Check
┃ ┣ 📂 PE Ratio
┃ ┣ 📂 PB Ratio
┃ ┣ 📂 EV/EBITDA
┃ ┣ 📂 Peer Comparison
┃ ┗ 📂 "Great business, wrong price"
┃
┣ 📂 Red Flags
┃ ┣ 📂 Rising Debtors
┃ ┣ 📂 Falling Cash Flow
┃ ┣ 📂 Frequent Equity Dilution
┃ ┣ 📂 Related Party Issues
┃ ┗ 📂 "One red flag is a hint, many are a parade"
┃
┣ 📂 Annual Report Review
┃ ┣ 📂 Chairman Message
┃ ┣ 📂 Management Discussion
┃ ┣ 📂 Notes to Accounts
┃ ┣ 📂 Auditor Remarks
┃ ┗ 📂 "Footnotes hide the gossip"
┃
┣ 📂 Concall Review
┃ ┣ 📂 Growth Guidance
┃ ┣ 📂 Margin Commentary
┃ ┣ 📂 Capex Plans
┃ ┣ 📂 Demand Outlook
┃ ┗ 📂 "Tone also matters"
┃
┣ 📂 Investment Thesis
┃ ┣ 📂 Why Buy
┃ ┣ 📂 Key Triggers
┃ ┣ 📂 Major Risks
┃ ┣ 📂 Time Horizon
┃ ┗ 📂 "Conviction needs reasons"
┃
┗ 📂 Final Result
┣ 📂 Strong Business
┣ 📂 Clean Financials
┣ 📂 Fair Valuation
┣ 📂 Watchlist Ready
┗ 📂 Or Maybe Just a Fancy Trap

[Bayriku 🇯🇵 @buywle]

Jadi cowok di 2026 itu gila.

Kamu harus jadi milyarder hanya untuk kencan dengan wanita miskin

[The Kobeissi Letter @KobeissiLetter]

BREAKING: US oil prices surge above $90/barrel, up +14% on the day, now at the highest level since October 2023.

We expect to see US gas prices above $3.50/gallon by Monday.

[Ray Dalio @RayDalio]

https://t.co/ZQbEmI4txz

[Jupe.base.eth | 劉 @0xjupe]

@RayDalio @grok give me summarize and what to do for the impact ?

[Jupe.base.eth | 劉 @0xjupe]

@RayDalio @grok as indonesian , what do i need to do right now?

[Reads with Ravi @readswithravi]

I’m in love with this sentence:

“Don't let someone ruin your peace just because they can't find theirs.”

[Nav Toor @heynavtoor]

BREAKING: AI can now build financial models like Goldman Sachs analysts (for free).

Here are 12 Claude prompts that replace $150K/year investment banking work (Save for later)

[P @PikamonID]

Setuju nggak: orang yang sudah lama terbiasa sendiri biasanya tidak takut ditinggal?