Ask yourself the que... • 2m
The Next AI Battleground? Open-Source LLMs Are Gaining Fast GPT-4 may still lead the pack — but the real action is now in open-source LLMs, and the gap is closing *faster than anyone expected In just 3 months: - Mistral’s Mixtral matched GPT-3.5 on many tasks with just 12.9B active params. - Meta’s LLaMA 3 is outperforming most commercial models on multilingual benchmarks. - Qwen from Alibaba quietly became the most downloaded open model on Hugging Face. - **Groq’s blazing fast inference** is redefining how LLMs are served — 500+ tokens/sec on Mixtral. Why it matters: - Startups now have near-GPT quality models they can fine-tune and fully own - Cost of experimentation has plummeted. - The moat of proprietary LLMs is eroding. We’re entering a new phase where LLM innovation may become decentralized, developer-led, and globally democratized — especially valuable for emerging markets like India. Will OpenAI go closed-loop, while the world goes open-source? What’s your bet: Foundation models or open ecosystems?
Hey I am on Medial • 1y
Hi any body using foundational models (llms) in development if you doing so are you using closed like gpt or Gemini for opensource models if you are using closed source why only that because you can save money by using opensource with low parameter m
See MoreThatmoonemojiguy 🌝 • 18d
wormgpt found to be a wrapper for grok and mixtral ☠️ WormGPT, an uncensored Al tool used by cybercriminals, was found to be just a wrapper for Grok and Mixtral, two legitimate Al services. The two Al tools were jailbroken using manipulated system p
See More| Technologist | ML ... • 4m
In the ever-evolving AI landscape, a new player is making waves — Deepseek. While OpenAI, Google DeepMind, and Meta AI have been dominant forces, Deepseek is emerging as a formidable contender in the AI race.The recent buzz around Deepseek stems from
See Morestartups, technology... • 11m
Meta has introduced the Llama 3.1 series of large language models (LLMs), featuring a top-tier model with 405 billion parameters, as well as smaller variants with 70 billion and 8 billion parameters. Meta claims that Llama 3.1 matches the performance
See MorePassionate about Pos... • 5m
Bhavish Aggarwal’s AI startup, Krutrim AI, has begun hosting Chinese GenAI company DeepSeek’s open-source models on its cloud platform. Five models, ranging from 8 billion to 70 billion tokens, are now live on Indian servers at the world’s lowest p
See More19yo ✨ #developer le... • 1y
Meta, formerly Facebook, has unveiled two new open-source AI models called Llama 3 8B and Llama 3 70B, with 8 billion and 70 billion parameters respectively. 🚀 These models outperform some rivals and spark debate over open versus closed source AI de
See MoreDownload the medial app to read full posts, comements and news.