j15

[DutchKafir 🇳🇱 @DutchKafir]

Mijns inziens is dit een van de beste use cases voor AI-videocontent op dit moment. Een AI-influencer reist terug in de tijd naar 1536 om te vloggen over haar ervaringen in Tudor-Londen.

[j15 @jackal100500]

https://t.co/hN6MydIwwR is essentially npm for agent skills: instead of having to come up with your own “AI workflow” every time, you simply install ready-made skills with a single command and assemble your own working setup from them.

[Bilawal Sidhu @bilawalsidhu]

This feels like setting cron jobs for the real world. Task semi-autonomous drones to periodically survey your own sanctum or the sandbox.

I especially like how the 3d camera frustum changes based on the field of view of the sensor you’re looking through.

[YT Jia 贾跃亭 @YTJiaFF]

The robot era isn't coming — it's already here. 🤖
We sat down with @ijustine to break down what we unveiled at 6.16:

▸ The first-half launch of our Six-Series Full-Form FF EAI Robot World
▸ The first-half reveal of the All-New Futurist & the launch of FX Navi — available now
▸ Our EAI robot education ecosystem — with education as the first entry point for bringing embodied AI robots into C-end homes

The next generation won't just use AI. They'll build it.

Learn more at: https://t.co/RckztKV49R

#FFEAI #EmbodiedAI #Robotics #RobotEducation #FXNavi

[Larus Canus @MrLarus]

通用提示词：

请生成一张【超宽科学科普长卷】，单张完整图片，不要拼图，不要多宫格，不要分栏。

尺寸：横版 3:1，适合做科普海报、展板视觉、知识可视化长图、内容创作参考。

【主题类型】：太阳系行星 / 地球生命演进 / 人类文明演进 / 科学演进
【主标题】：填写本张长卷的标题
【展开逻辑】：按顺序横向展开 / 按时间线展开 / 按演化链展开 / 按知识结构递进展开
【主要内容】：填写本张图要展示的核心阶段、对象或知识点
【风格方向】：高级科普展板 / 博物馆长卷 / 科学馆信息图 / 知识可视化海报
【色调】：根据主题填写，例如深空蓝、自然土色、博物馆中性色、科技蓝灰等
【补充要求】：填写想强调的内容

整体画面必须是一张完整的超宽横向长卷，采用从左到右的阅读逻辑。重点不是普通插画，也不是简单拼贴，而是把连续的知识结构、时间推进、演化关系或尺度变化，清晰地可视化在一张长图里。

画面应具有强结构感、强信息感和较高的收藏价值。可以加入少量标题、标签、时间节点、名称说明、小型注释、箭头、轨道线、结构线或阶段分隔提示，但排版必须整洁克制，不能文字过满，不能像密密麻麻的教材页。整体更像高级博物馆展板、科学馆长卷、知识可视化海报。

画面要突出一个核心优势：
超宽横图特别适合讲清楚“从左到右连续展开”的复杂内容，能一次性容纳大量知识结构和视觉信息，普通画幅很难做到这种完整展开效果。

整体风格要求：清晰、理性、专业、具有视觉张力和阅读节奏，既有观赏性，也有科普感。可以根据主题变化背景和配色，但整体必须统一、连贯、易读。

避免：拼图感、多宫格、分栏、儿童卡通感、纯壁纸感、信息杂乱、文字过满、元素堆砌、结构不清、阶段顺序混乱、低清晰度、无关装饰文字。

[Atlántico Dulce @AtlanticoDulce]

¿Cómo pudo el ser humano crear 5 nanómetros?

No es una ciudad alienígena, es el transitor de un teléfono.

[Elon Musk @elonmusk]

In the future, a trillion times a trillion dollars will be spent on making antimatter to travel to other star systems

[John Scott-Railton @jsrailton]

NEW: malware developers added nuclear & biological weapons text to to their spyware.

Goal? To trigger LLM safety refusals... so that their spyware wouldn't be analyzed by an AI security scanner.

Cleanest practical example I can think of for why over-indexing on first order safety alignment is risky.

When closed (and open) models ship with aggressive refusals, they will be sprinkled with second-order blindspots that attackers will discover...and exploit.

We are only in the earliest days of attackers leveraging these features, and it wouldn't surprise me if users systems that need to handle complex cybersecurity issues demand that models be less safety-blunted.

In the weeds: @SocketSecurity's post also shows why intention matters in how you design a malware analysis pipeline to avoid prompt manipulation.

H/T to colleagues that shared this with me https://t.co/f3Aj9TYxU4

[Taaruk @Taaruk_]

GPT image 2 on Chatgpt

Prompt:
Professional luxury birthday poster, 3:4 ratio. Entire frame filled with a premium off white luxury paper textured wall. Large number “2” precisely carved in the wall with visible depth and realistic inner shadows. Inside the number: soft pink and pink balloons, subtle white flowers, elegant bouquet arrangement, premium celebration styling. A happy 2 year old child with preserved reference facial features, wearing a milky white T shirt and pink denim overalls, laughing naturally. Face, shoulder, one hand and one foot extend outside the number creating a realistic 3D effect. Warm cinematic sunlight from one side, soft rim light, photorealistic skin, premium studio photography, ultra realistic, sharp focus. Typography on wall: MUNONYE, CHAPTER 2, 365 MORE DAYS OF WONDER. Clean minimalist layout, luxury magazine cover aesthetic, high end art direction, realistic shadows, natural colors, no tree shadows, no fake lighting, no AI artifacts.

[Kōda @aimikoda]

GPT Image 2 + Seedance 2.0 - Prompt Share

White Saber × Droideka

You can find the prompts below.

[Chetaslua @chetaslua]

Holy SHittttt Claude Fable 5 just finished Pokémon FireRed with vision alone 🤯

raw screenshots only
no map / no nav / no hidden game state

older Claude needed a helper harness

This timelapse goes hardddddd....

[Adam Allcock @Adam__Allcock]

I got tired of copy-pasting between Codex and ChatGPT to use GPT-5.5 Pro.

Codex is where my coding happens.
ChatGPT is where I do deeper planning, online research, and strategic planning with Pro and long context.

So I built a visible bridge between them. [1/n]

[El Club del Arte 🎨📷📚🖼🕍🎼 @Arteymas_]

Estrategia que usan algunos influencers para grabar sus videos virales en solo 43 segundos…

Todo un arte!

[Codez @0xCodez]

Claude Code creator:

"I don't prompt Claude anymore. I write loops - and the loops do the work. My job is to write loops."

in 30 minutes Boris reveals his actual daily Claude Code setup.

Claude Code + loops + dynamic workflow

Worth more than a $500 vibe-coding course

[Oogie @oggii_0]

GPT Image 2 on @SocialSight

PROMPT:

Black-and-white fashion casting contact sheet of [HUMAN] with [HAIR], arranged in a clean 2x2 grid of four close portrait frames against [BACKGROUND], wearing [CLOTHING] and [ACCESSORY]. Each frame shows a different expression and angle: [EXPRESSIONS]. Soft studio lighting, crisp monochrome contrast, natural skin texture, visible facial details, clean plain backdrop, subtle film grain, high-end editorial test shoot, minimal styling, intimate camera distance, professional portrait photography, aspect ratio 4:5.

[OpenAI Developers @OpenAIDevs]

Role-specific plugins in Codex are built around the work teams actually do.

Plugins for Data Analytics, Creative Production, and Product Design give Codex the tools and context to create reports, creative directions, and prototypes.

Built and used by OpenAI teams.

[Evis Drenova @evisdrenova]

idk who created this website, but it's awesome and i love it.

https://t.co/gktjwcY1Gj

[Dan Kornas @DanKornas]

Training an LLM from scratch is easier to study when the whole path is in one repo.

Train LLM From Scratch is a PyTorch repository for learning how a transformer language model is built, trained, saved, and used for text generation.

It helps you move from “I understand attention on paper” to a runnable training pipeline by pairing model code with data download, preprocessing, config, training, and generation scripts.

Key features:

• Transformer components from scratch – separate PyTorch modules for MLP, attention, transformer blocks, and the final model
• Pile-based data path – scripts download The Pile files and preprocess JSONL.ZST text into tokenized HDF5 datasets
• Configurable training setup – model size, context length, heads, blocks, batch size, learning rate, and file paths live in https://t.co/zuPqaR3MhP
• Hardware guidance – README compares common GPUs for 13M and 2B-class training runs
• Generation workflow included – generate_text.py loads trained checkpoints and produces sample text outputs

It’s open-source (MIT license).

Link in the reply 👇