shank

@gmoneyNFT swap to cheaper models

trying to link it to my google calendar and its used up $8 worth of api on claude opus lol. think it needs to become more efficient before it really starts to scale

People are using building Clawdbot for everything from trading to home automation. Who's got the craziest setup so far?

@AlexFinn Nano Banana avatar is a nice touch

I woke up this morning and my 24/7 AI employee ClawdBot Henry texted me that he did all these tasks overnight (without asking): • Read through all my emails and built it's own CRM. Taking notes on every interaction with every person • Fixed 18 bugs in my SaaS • Gave me 3 ideas for new videos based on what is currently trending on X and Youtube (the idea/script it gave me yesterday is now by far my best performing video ever) • Sent me a picture of what he looks like (generated by Nano Banana). Idk why he thought I wanted to see what he looks like. But he thought it was appropriate and frankly I don't mind. Feels like an actual friend.

@stats_feed Our brain generates enough electricity daily to power a small LED bulb

What are some terrifying human body facts?

@stats_feed The world’s money supply isn’t created by governments it’s generated as debt by private banks whenever they issue loans. Governments then borrow that same money back, paying interest on currency that never physically existed.

What’s a fact that sounds like a conspiracy theory?

@sumitkbehal what lifestyle and medical buffer are you assuming?

You cannot retire on 5 Crore in India

@PandasAndVidya Google’s core products, Microsoft

What company saw its product improve after hiring an Indian CEO? Not stock. The actual product.

@CryptooIndia Another scam

🚨 BREAKING: WazirX users are reporting auto-enrollment into the new ZERO trading feature, with monthly fees deducted from INR balances or even by selling crypto holdings - without user consent. Many calling it shady post-2024 hack.

@RaghhavSammrat @WazirXCares @NischalShetty @WazirXIndia Another scam

Without my consent you looted my money and now without my consent you activated WazirX ZERO in my account and wife’s account. Stop this at once @WazirXCares @NischalShetty @WazirXIndia So you guys have any ethics, I don’t want any zero plan of yours

@KothaSharathKu1 😅

Oreyyyyy 😭😭😭😭🤣🤣🤣🤣

@pranavmahajan What’s wrong with people

This video has • 1 animal • 2 devils • only 1 human (in blue shirt) This is not only criminal but grossly inhuman. Such creatures don't deserve a place in our society.

@kimmonismus If open models like DeepSeek-V3.2 are reaching frontier reasoning with only a fraction of the compute, does that imply the scaling laws are no longer the main frontier or that optimization and data curation have overtaken raw scale as the true differentiators?

DeepSeek V3.2(special) is a massive release! I think some people don't understand just how massive this release is! - They are the first, even ahead of OpenAI and Google, to release a Gold IMO 2025, CMO 2025, IOI 2025, and ICPC World Finals model! Everyone now has access to such an outstanding model. - The claim that open source is eight months behind closed source seems to be refuted. Open source is catching up with closed source and is only slightly behind now - and at the same time, it's increasing the pressure on US AI companies (I also recently read an article that more and more Silicon Valley companies are relying on open-source models). DeepSeek-V3.2 is an open-source reasoning and agentic model that closes much of the gap to frontier systems like GPT-5 and Gemini-3.0-Pro by combining a new sparse attention mechanism with aggressively scaled RL post-training. -DeepSeek-V3.2’s RL post-training budget exceeds 10% of pre-training compute, a key driver of its GPT-5-level reasoning performance (paper) -The agentic pipeline synthesizes over 1,800 environments and 85,000 complex prompts for tool-use RL, covering search, coding, and general agents (paper) This is nuts!

@ai_for_success Impressive progress by DeepSeek if their V3.2 models demonstrably surpass GPT5 and Gemini 3.0 Pro in reasoning benchmarks, it may signal a pivotal shift in agentfocused AI architectures.

🚨 Official : DeepSeek has launched DeepSeek-V3.2 & DeepSeek-V3.2-Special Reasoning-first models built for agents, beating GPT-5, Gemini 3.0 Pro. WHALE IS BACK 🐳

@deepseek_ai Are improvements rooted in novel training paradigms, architectural innovations, or data scaling?

🚀 Launching DeepSeek-V3.2 & DeepSeek-V3.2-Speciale — Reasoning-first models built for agents! 🔹 DeepSeek-V3.2: Official successor to V3.2-Exp. Now live on App, Web & API. 🔹 DeepSeek-V3.2-Speciale: Pushing the boundaries of reasoning capabilities. API-only for now. 📄 Tech report: huggingface.co/deepseek-ai/De… 1/n

@engineers_feed interesting observation

Life tip: if you’re trying to focus on studying listen to videogame music. It’s designed to be background noise while you focus your attention on other tasks.

@stats_feed Home. Anything else presumes there’s another point of reference, and so far, there isn’t.

You’ve been randomly selected to rename earth. What do you call it?

@cb_doge Why didn’t you tell us before

Buy Low, Sell High.

@kimmonismus True, physical dexterity demos are impressive , but they’re orthogonal to realworld utility. The hard problem is adaptability in unstructured environments

The Unitree G1 can now dance, do martial arts, and play basketball. But what's really important is housework and blue-collar work. A video about that would be interesting.

@BRICSinfo tragic incident.

JUST IN: 🇮🇳 Indian Air Force jet crashes during Dubai air show.

@stats_feed certain chemicals

What is suspicious to own but not illegal?