RT @mahersharafeddi: بس يفوتوا على بلدك ويرموا الناس من البنايات، ويقتلوا الشخص اللي قال: أنا سوري، ويسرقوا ويحرقوا بيوت ٣٦ قرية، ويخطفوا ٨….

Say goodbye to asking AI what to do. 👋 Say hello to true automation. Today, @get_WRITER is releasing Action Agent – an enterprise-ready, autonomous AI agent – available in open beta. Here’s how it works:. Every time a user starts a chat with Action Agent, we spin up a virtual.

The most fun image & video creation tool in the world is here. Try it for free in the Grok App.

RT @rohanpaul_ai: Training an LLM to talk itself (self‑reflection) through mistakes can lift accuracy by up to 34.7%, turning a 7B‑paramete….

RT @samurai_611: 🔆 To my Druze brothers and sisters — in Syria, Lebanon, Israel, and everywhere around the world:. Our culture is a mosaic….

RT @RepAbeHamadeh: NEW. Congressman Hamadeh's Statement On the Barbaric Violence Being Committed Against the Druze in Syria .

RT @BillAckman: Why is this not front page news?.

Everyone’s chasing AGI like it’s the moon. Meanwhile, 90% of enterprise AI problems are just:.- Bad data.- Worse prompts.- And someone yelling “just use LangChain”.

LLMs can summarize a PhD thesis, write Python, translate Urdu, and generate Drake lyrics….…but still hallucinate the URL I literally gave it. AGI is coming. But slowly. And drunk.

China's going for quantity, America's betting on quality. Meanwhile, Europe, Canada & Israel are building their own AI sovereignty.

The new Cold War is being fought with algorithms, not armies .- China: 12 major LLM companies (52%).- USA: 6 major LLM companies (26%).- Rest of world: 5 companies (22%)

Prompt engineering is just gaslighting the model until it gives you what you want. “You meant to answer correctly. You’re helpful, honest, harmless… remember?”.

You want unlimited context? Think beyond RAM and RoPE. → Hierarchical Attention Routing: Embed attention selectors that recursively gate tokens across a temporal-spatial memory tree. Like Mixture-of-Experts meets the brain’s prefrontal cortex. → Learned Virtual Memory (LVM):.

RT @Get_Writer: WRITER Research: Better self-reflection leads to immediate performance gains 📈. Shelly Bensal, AI Research Engineer at WRIT….

AI: 10% magic, 90% cleaning up your data mess.

Reminder: Hype isn’t innovation.

AI innovation dies when CEOs start micromanaging research. Meta rushed Llama 4. OpenAI stalled post-GPT-4. Overhiring won’t fix it.

VCs: “A LOT of GPUs = AGI.”.@Get_Writer : “How about LLMs that don’t hallucinate insurance policies?”.Enterprise AI isn't sexy, but it's where the real impact lives.