开源+赞助BTC基金 唯一的交易所赞助人OKX一直在这个行业做出自己的贡献OKX中文 (@okxchinese):🦾欢迎开发者们免费使用 OKX 开源技术，和我们一起让 Web3 变得更好：https://t.co/OuOKalFxNH ...

RT Mario Nawfal (@MarioNawfal)GROKIPEDIA IS ALREADY GRABBING WIKIPEDIA’S TRAFFICClaude blinked first and started pulling answers from Grokipedia instead of the old, dusty hall of "fact" citations.The Shift:• Claude now cites Grokipedia by choice, not by force• Grokipedia runs on open source data with real time updates and bias cleanup• Elon shrugs it off since the whole thing stays open and publicTraffic follows utility, and Grokipedia is already pulling the crowd toward the real truth!Source: @elonmusk, Be ...

Meta just dropped SAM 3 that is segment anything model and it allows you to use simple text prompting to segment anything in a video easily. Let me take a step back. There's this thing called rotoscoping.It is the extremely manual process that takes a team of dozens of people by manually segmenting different elements in a video. And now with SAM 3, it takes seconds. I'm partnering with Meta on this video to tell you about this incredible open-source open weights model that allows you to do some pretty incre ...

Zuckerberg Directs Pivot Away From Open Source AI At Meta. Listen for more on Bloomberg Intelligence. https://t.co/v0EzRmtIRr ...

The story was Metaypically is they hire ex-AI experts from Apple they've hired up so much talent you'd think that they have an insane bench and that things are going smoothly. Does this suggest that things are not going so smoothly with matter and their development of their offering. Well, in the first instance when the story had there was a kind of strange algo spike and then drop in the shares, but we're now down one and a half percent, so it's difficult to interpret the markets.Read a matter shifting to ...

[music] Today we're talking about RL environments and how to scale them. But the title is a little bit of a red herring. We'll talk a bit about the engineering pieces and like running these with thousands of parallel rollouts and sandboxes on hundreds of GPUs, but I'm mostly going to focus on a different notion of scale. Uh, and what I mean by scaling here is we there's a number of different ways we talk about scaling in the context of AI and research. We know about scaling laws and we talk about how much d ...

Open Model Impact - Open models democratize intelligence access, shifting focus to customer care and product excellence [1] - Open models transform users into makers, enabling specialization for specific use cases [2] Business Strategy - Success hinges on understanding and serving customers effectively, rather than resource dominance [1] - Customization of open models is key to creating unique value and applications [2]

AI Trends and Predictions - The AI industry is shifting towards specialization, with enterprises focusing on fine-tuning and specializing models for specific domains [6][8][82] - Open source technologies are driving transparency and adoption of AI agents, giving more power to enterprises and consumers [8][10] - The next wave of innovation is expected in world models, which are extremely data-intensive and will be the base layer for robotic opportunities [69][72] Challenges in AI Adoption - Agent memory is an unsolved problem, requiring agents to have persistent memory of both the user and itself [13][14][15] - Seamless communication between AI agents requires open source communication protocols [23][56] - AI security is crucial, with the potential need for a high ratio of security agents to cognitive intelligence agents [24][26] - Evaluating AI performance requires moving from academic benchmarks to real-world evaluations and reinforcement learning environments [34][38][39] Investment and Innovation - Capital investments are shifting from the model space to the agent space, driven by the focus on people and applications [58][59] - Enterprises seek AI solutions with high accuracy, small footprint, and data privacy [49][50] - Distillation, which involves making large models more efficient and smaller, is becoming important for cost-effectiveness [51][52] Enterprise Adoption Strategies - Enterprises should view model development as a software development platform, focusing on MVP and optimization over time [53][54][55] - Enterprises are adopting generative AI slower due to legacy systems and data locked in those systems [80][81] - Enterprises should focus on systems of smaller, specialized models rather than one model to solve all problems [83] Stochastic Mindset and Evaluation - AI compute is becoming more stochastic, requiring a shift in how we interface with and evaluate computers [30][32] - Verification of AI in specialized domains is challenging due to the difficulty and expense of expert verification [41] The Role of Open Source - Open source is critical for base models and communication protocols, enabling enterprises to build and compete with their own workflows [11][57] - A \$2 billion investment was raised with Nvidia's participation to support the US open source development ecosystem [11] Iteration and Mindset - Companies should iterate quickly, inspired by gradient descent algorithms, to gather information and find new opportunities [75][77] - Founders should pick a starting place that is exciting, big, and challenging enough to be worth the effort [79]

It is often two steps forward, one step back, but we are making rapid progress in showing people compelling content.It’s too much in flux right now, but, hopefully by next month, we should be able to open source literally all of the @X codebase. Nothing left out at all.Robert Scoble (@Scobleizer):The vibes are shifting here on X.Been talking to some on the inside of xAI.The new fully-completed algorithm is still a few weeks off, I hear. It will be run by a new Grok, and the report below makes me excited abo ...

Core Viewpoint - Mistral AI has launched the Mistral 3 series of open models, which are positioned as high-performance, cost-effective alternatives in the AI model landscape, particularly in response to competition from DeepSeek [2][4][28]. Model Details - The Mistral 3 series includes multiple models: Mistral 3 (14B, 8B, 3B) with base, instruction-tuned, and reasoning versions [5][19]. - Mistral Large 3, a state-of-the-art open model, features a total parameter count of 675 billion and 41 billion active parameters, trained on 3000 NVIDIA H200 GPUs [7][5]. Performance and Benchmarking - Mistral Large 3 ranks second in the OSS non-inference model category on the LMArena leaderboard, indicating it is one of the best-performing open models available [14]. - The model demonstrates strong performance in general prompt tasks and excels in image understanding and multilingual dialogue [7][14]. Collaboration and Optimization - Mistral has partnered with vLLM and Red Hat to enhance accessibility and efficiency for developers using Mistral Large 3, utilizing optimized checkpoints for better performance [17][18]. - The collaboration with NVIDIA focuses on advanced optimization techniques, ensuring that Mistral models leverage high-bandwidth memory for demanding workloads [17][18]. Cost-Effectiveness - Mistral claims that its models offer the best cost-performance ratio among open-source models, with instruction models performing comparably or better than competitors while generating tokens at a significantly lower rate [22][28]. Availability and Customization - Mistral 3 models are available on various platforms including Mistral AI Studio, Amazon Bedrock, and Azure Foundry, among others [25]. - The company also offers custom model training services to organizations seeking tailored AI solutions for specific tasks or environments [27].