Kingslayer

Learner - active villain

[纳特🇵🇬 @ayehuasca]

cron update silahkan dipake ke agent kalian

https://t.co/Jhjt60omio

[Ram Maheshwari @rammcodes]

WTF... Someone literally built a massive collection of open-source, in-browser tools that require no sign-up to use. 🤯

It includes tools across multiple categories:
→ Design & Graphics
→ Development
→ Productivity
→ Privacy & Security
→ AI
→ Education
...and many more.

No accounts. No sign-ups. Just open your browser and start using them :)

Follow me for more amazing AI, Coding & Web Dev insights 💎

[爱丽丝呀！ @BTCqzy1]

Codex 5.6 真的要让剪辑师失业了！

它现在的视频制作能力提升了不止一个档次，直接一句话生成视频！

重度实测一周，筛选出这 4 个Skill，直接实现全流程闭环：

1️⃣ HyperFrames （内置插件）

负责整片策划、动效生成、画面合成与最终渲染。它能把文字构思直接转化为高水准的动态视频框架，实现全流程闭环交付。

2️⃣ Remotion （内置插件）

基于 React 模板的视频生成工具。支持通过代码批量生产高质感视频，尤其适合数字人+动效结合的自媒体内容，实现真正的工业化生产。

3️⃣ video-use

专门处理真人素材的剪辑工具。能自动识别镜头、精准卡点并完成后期剪辑，让 AI 具备了职业级的剪辑逻辑。

地址：https://t.co/QqX1MJaemQ

4️⃣ Seedance

针对顶级视频大模型的提示词工具。专门生成���影级运镜和实拍感指令，补齐了 AI 视频在审美和镜头质感上的短板。

地址：https://t.co/r4RL7z744e

现在用 Codex 5.6 + 这4个Skill，从脚本→动效→剪辑→渲染，全程AI闭环。

[oxbfnohome @OxbfnoHome]

kasih skill ini ke agent untuk generate vidio secara otomatis

gue nemu repo open source dari heygen yang menurut gue keren banget, namanya HyperFrames
Bedanya sama video editor biasa?
Lu gak perlu buka After Effects atau Premiere lagi
Cukup kasih prompt ke AI Agent, nanti agent bakal:
✅ Nulis HTML
✅ Menyusun layout video
✅ Menambahkan animasi
✅ Menambahkan transisi
✅ Render langsung ke MP4
Yang menarik, HyperFrames menggunakan HTML sebagai format video, bukan timeline editor tradisional

jadi AI Agent jauh lebih mudah memahami cara membuat video karena hampir semua model AI sudah sangat jago menulis HTML

gw udah nyoba bikin vidionya menurut kalian ini udah bagus atau belum, soalnya gw hanya kasih screensoot lalu suruh agentnya bikinin vidionya, bagi yang ngerti mengenai vidio tolong koreksinya dong, gw butuh sarannya 🙏

https://t.co/HXW7F8LY03

[Risang Pradipta @rsngprad]

Mau reshare lg aah

Beberapa bulan lalu, di twitter juga, gw nemu 1 google drive isinya ilmu semuaaa.

Nih langsung aja https://t.co/qEvuMTjGAP

Silakan belajar sepuasnya yaaa. Semoga bermanfaat

Silakan bookmark dan bantu repost yaa!

[dwi budi @nayantaka]

Saya ingin belajar dari repo Meridian di GitHub dari Pak @0xyunss bot trading Meteora DLMM yang sudah jalan otomatis. Saya juga sering lihat dari postingan Pak @magersih dan pak @bengsharksol

Tapi saya belum berani terjun langsung. Bukan karena takut teknologi, tapi karena saya belum paham alur keputusannya: dari mana datanya, siapa yang menilai, dan kapan sebenarnya boleh eksekusi.

Jadi saya coba develop sendiri Multi-Agent Theater — versi edukasi, saya suruh codex dan grok build untuk belajar dari repo github.

Cara kerjanya (SOL di pool SOL-USDC):

1. Data dulu, tanpa wallet — pool diverifikasi ke API resmi Meteora DLMM (bukan asal tebak pair).
2. Skor & risiko — metrik DLMM (TVL, volume, fee/TVL, bin step) + gate token Jupiter (organic, holders, dll.).
3. Paper simulasi — rencana posisi off-chain (range harga, estimasi fee harian) disimpan di journal lokal, bukan transaksi.
4. Lima agent AI (CrewAI) — Scout (data cukup?), Analyst (metrik), Narrative (konteks token), Dip Catcher (timing), Manager (sintesis keputusan).
5. Guardrail Python — meski LLM optimis, aturan keras bisa menurunkan status; tidak ada auto-sign / auto-send.

Hasil analisis SOL–SOL-USDC kali ini: READY_TO_ANALYZE (bukan READY_TO_TRADE, bukan perintah beli).

Rekomendasi AI:

• Lanjutkan advanced off-chain analysis dengan pool address terverifikasi: 5rCf1DM8LjKTw4YqhnoLcngyZYeNnQqztScTogYHAS6
• Jangan eksekusi on-chain dari keputusan ini; READY_TO_ANALYZE = layak dianalisis lebih lanjut
• Gunakan paper position & journal sebagai simulasi, bukan instruksi transaksi
• Modelkan: active bin, range width, out-of-range probability, volatilitas SOL, impermanent loss, rebalance, slippage, bin depth, biaya tx
• Uji keberlanjutan fees/APY beberapa periode, bukan snapshot 24 jam saja
• Pantau apakah harga SOL tetap dalam range simulasi 72.23 – 78.25; bandingkan paper fees vs estimasi harian ~$5.14
• Dip-catching: tunggu drawdown nyata & stabilisasi; jangan chase saat 24h masih positif
• Refresh TVL, volume, fees, active bin, Jupiter risk sebelum naikkan status

Tentunya ini belum membuat saya “trader DLMM”.

Yang saya dapat: cara berpikir sebelum sentuh uang verify pool, baca risiko, simulasi dulu, baru diskusi dengan AI yang dibatasi guardrail.

Meridian tetap referensi arah; Multi-Agent Theater jadi kelas saya sendiri.

Saran dan arahan dari beliau2 para ahli sangat saya butuhkan.

Video dan gambar arsitektur saya drop di bawah

[Ily 🔻 @kliseanganku]

Indonesia’s biggest loss:

A thriving semiconductor industry.

[Haryo Aswicahyono @Aswicahyono]

Th 1985, Sudomo, Menaker waktu itu, melarang Firchild otomatisasi produksi chip. Akibatnya Fairchild kabur ke Malaysia

sering kali kebijakan yg kelihatan mulia, eg. perlindungan tenaga kerja saat ini punya efek jangka panjang yg buruk, hilangnya kesempatan kerja buruh masa depan

[Electra @Electrarythm]

There is a certain type of person everywhere now, especially online.

He consumes endless information every day: philosophy, psychology, productivity, spirituality, neuroscience, business, self-improvement, history.

He knows a little about everything and deeply experiences almost nothing.

His entire identity becomes built around understanding instead of living.

He watches videos about confidence instead of speaking confidently. Reads about discipline instead of becoming disciplined. Studies relationships instead of learning how to love. Consumes motivational content instead of taking action.

He feels intelligent because he is constantly mentally stimulated. But stimulation is not transformation.

Most of the time, knowledge becomes emotional protection. Reality is unpredictable. Reality humiliates. Reality exposes weakness. Books and ideas do not.

Inside information, he can continue imagining himself as intelligent, deep, insightful, different from ordinary people. So he remains trapped in preparation.

He constantly feels as if he is "becoming" someone, while his real life remains strangely untouched. He develops sophisticated language for problems he never confronts directly. He can explain human behavior beautifully while being unable to handle ordinary discomfort, rejection, uncertainty, loneliness, or risk.

He slowly turns life into observation instead of participation.

The internet rewards this personality heavily. He receives validation for sounding aware rather than becoming capable.

Eventually, he begins confusing self-analysis with growth and information with wisdom.

But beneath the intelligence usually exists the same thing: fear. Fear of failure. Fear of embarrassment. Fear of reality answering back.

Because action destroys fantasy. The moment he truly acts, he can no longer hide inside potential.

[Claude @claudeai]

Claude Fable 5 has been out for a couple of days. Some projects people have already built with it:

[Viktor Oddy @viktoroddy]

Claude Mythos is Insane!

❤️‍🔥Just recorded a 12-min tutorial on how to build animated, award-Winning websites with Claude Fable 5!

[yunus @0xyunss]

masih inget agent DLMM di @MeteoraAG gua yang bulan kemaren ngehasilin gua $1.5k, ini adalah tutorial lengkap 32 menit full dari awal banget setup hingga oprek.

kalian yang gabisa apa-apa udah dijelasin semua di sini.

kalo ada yang kurang tolong banget buat komen biar gua bikinin video lanjutan.

semua link untuk setup kebutuhan ada di komen

00:00 Intro
02:47 Persiapan Tools
04:32 Fundamental
09:08 Installing
12:13 ENV dan User Config
18:57 First Start
20:27 Telegram Community
24:12 Basic Troubleshoot
27:02 Screen Session
28:05 Hardcode
30:10 Penutup

[Luca @lucaxyzz]

Irawati, rahmat wibowo, skrng prihantini.

Energi netizen akan habis ngecancel satu satu. Gue no comment, lebih ke ‘dont blame the player, blame the game’ 😂

[Afshine Emrani MD FACC @afshineemrani]

1/5
I'm a cardiologist. I have spent twenty years watching cholesterol destroy arteries, trigger heart attacks, and kill people I care about.
Today, Eli Lilly presented data that may begin to end that era.
VERVE-102. A single infusion. One dose. It uses base editing to permanently turn off the PCSK9 gene in your liver.
Presented today at the European Atherosclerosis Society Congress:
88% reduction in PCSK9.
62% reduction in LDL cholesterol.
Sustained up to 18 months.
No treatment-related serious adverse events.
One infusion. Not daily pills you forget to take. Not monthly injections. One dose — and your cholesterol may stay low for the rest of your life.

[The Boring Marketer @boringmarketer]

how I plan to multiply my wealth:

1. generate cashflow as an ai practitioner
2. invest cashflow in ai picks and shovels

not financial advice!

[Nav Toor @heynavtoor]

Opus 4.8 added sound design and animations without being asked. it just decided the project needed them.

that's the detail from this thread that stopped me.

a Mars rover simulator. a 3D partnership visualizer. an ecosystem simulator. a full landing page where Opus acted as creative director and orchestrated other models.

7 builds. every one interactive. every one linked.

fewer coding errors. better spatial reasoning. more natural copy. and it self-corrects over long tasks instead of drifting.

this is the best showcase of what Opus 4.8 actually feels like in practice.

[Akshay 🚀 @akshay_pachaar]

Microsoft just open-sourced SkillOpt!

A framework for training agent skills like neural networks:

SkillOpt treats a plain markdown file as the trainable parameter of a frozen LLM agent, applying the same optimization discipline used in weight training: learning rates, validation gates, batch sizes, and epoch schedules.

The analogy maps precisely. The skill document is the parameter. Trajectory-derived edits are the gradient direction. An edit budget is the learning rate. A held-out split is the validation check.

Here's how it works.

A frozen model runs tasks with the current skill and produces scored trajectories. A separate optimizer model analyzes failures in minibatches, proposes structured add/delete/replace edits, and ranks them under a budget cap.

If the candidate skill improves performance on a held-out split, the edit is accepted. If not, it's rejected and stored so the optimizer avoids repeating failed changes.

The deployed output is a single best_skill. md file, typically 300 to 2,000 tokens. No weight changes, no extra inference-time calls.

The learned rules are compact and readable. These read like rules a thoughtful engineer would write after a day with the benchmark, except they were discovered automatically.

Learn more:

Paper: https://t.co/sdj5DW7t9h
GitHub: https://t.co/W3DcpBCni0

SkillOpt isn't the first system to treat skills as something you can optimize.

Hermes Agent independently built the same idea through a combination of skill_manage, Curator, and an optimization loop called GEPA that scores, mutates, and promotes skill documents across runs.

Two teams, different architectures, same conclusion: the skill file is the highest-leverage thing to optimize in a frozen-model agent.

I wrote a deep dive on how the Hermes agent works and covered all of these topics briefly.

The article is quoted below.

[Kingslayer @Kingslayergd]

@viktoroddy ninggal jejak

[Hanif | AI For Productivity @hanifproduktif]

Kalo kita all-in duit MBG 1 tahun (Rp 335 triliun) dan wowo beneran lengser payoutnya Rp 7,9 kuadriliun

Bisa untuk MBG selama 23 tahun tanpa membebani APBN

Ayok wok gas all-in buktikan kalo lo beneran patriot sejati