Poe on X: "DeepSeek-V3 is now available on Poe! DeepSeek-V3 stands out as one of the most powerful open-source models available and excels in math, coding, and reasoning tasks. Its long context handling and performance across key benchmarks rival leading closed-source models. (1/2) https://t.co/BfE6xe69vU" / X
Poe on X: "QwQ-32B-Preview is now available on Poe! This chain-of-thought based reasoning model from Alibaba is useful for advanced math, coding, and logical reasoning, achieving top-tier scores across various problem-solving benchmarks that rival those of other reasoning models. (1/2) https://t.co/7LGjRJI4eR" / X
Is QwQ the same as Qwen? Google: No, while "QwQ" is closely related to "Qwen," it is not exactly the same model; "QwQ" stands for "Qwen with Questions" and is essentially a specialized version of the Qwen AI model designed specifically for tasks that require deeper reasoning and questioning, making it more focused on logical thinking and problem-solving compared to the standard Qwen model. Key points to remember: Base model: Both QwQ and Qwen are developed by Alibaba. Functionality: Qwen is a general-purpose AI model, while QwQ is specifically designed for tasks that require extensive reasoning and self-questioning. Meaning of "QwQ": "QwQ" is a playful way of representing a questioning or curious attitude, which aligns with the model's focus on deep thinking and doubt-checking.
https://poe.com/QwQ-32B-Preview-T…
Poe on X: "Today we’re introducing new subscription options to expand access to Poe, including a low-cost plan and higher point options. These plans provide access to all premium bots from OpenAI, Anthropic, Google, xAI, Runway, Ideogram, and many other providers. (1/7) https://t.co/fSk6mZvhB6" / X
Poe on X: "Now available on Poe: Claude 3.5 Haiku! The new Haiku model outperforms its predecessor across all metrics and exceeds Claude 3 Opus on several intelligence benchmarks. It’s fast, has improved instruction following, and is particularly strong with coding tasks. (1/2) https://t.co/DxJYv3fKwK" / X
Are there any specific decision-making methods that utilize new AI resources, e.g. large language models, VR, or other branches?
(Past Eisenhower Priority Matrix, and JFK Groupthink. On POE, o1-mini uses 1800 points per message out of 1M monthly subscription compute quota, whereas o1-preview uses 10000. ChatGPT-4o-mini uses 15. Gemini-1.5-Flash is 5. New Llama 3.2 are 475 for 90B and 115 for 11B.)
Dream Analysis - Poe
(Dream Anxiety would argue for better methods of recording those. AF meant Space on The Outer Limits. The weather, for Rangers seeing the hurricane forecasts and ancient poetics. Also a reminder from back in the day of EG&G before AI Winter. Now you would likely see a display case featuring earbuds with built-in chatbots. Upon further review, where most are trained for attrition, some have no one else in their domain. What keeps them going? It could simply be an ad for CoPilot. Or Minority Report, though Farrell now plays Oz Cobb. This bookmark is on Raindrop. Upon further reflection, the whole scene is a mnemonic now that could be used as a memory palace. Not quite deconstruction, but perhaps a quest for new methods under AI, if not Inception. Gibson went from Neuromancer to Jackpot. Meta tosses in trendy wristbands. Ironically had to fix this link.)
Poe on X: "The just-released Llama 3.2 models are now available on Poe! Llama 3.2 90B and 11B are cutting-edge multimodal models that show high-performance in vision-related tasks, including analyzing charts and graphs, describing images, and answering questions about pictures. (1/2) https://t.co/LHf5pSJmqH" / X
OpenAI on X: "We're releasing a preview of OpenAI o1—a new series of AI models designed to spend more time thinking before they respond. These models can reason through complex tasks and solve harder problems than previous models in science, coding, and math. https://t.co/peKzzKX1bu" / X
(o1 is directly aimed at STEM. o1-mini is much faster, but each runs out at 10 messages per day. o1-preview is slower than usual but formatted and purports to go through reasoning, although the steps are not shown. When asked how to learn STEM or if it is equivalent to conventional, it goes into detailed comparison and offers the hybrid option. The model is trained as of September 2021. File uploads like PDFs are limited in size. o1-preview-128k could answer on a par with @Web-Search or Perplexity elsewhere: What is the critical review of the new OpenAI LLM chatbots for o1-preview and o1-mini?)
Poe on X: "Ideogram 2.0 is now available on Poe! This model is the most advanced image generator available and excels at producing high-quality, lifelike images, especially for human features. It also offers industry-leading text rendering, seamlessly blending text into your images. (1/2) https://t.co/IiuqqzYxr6" / X
Poe on X: "Llama 3.1 405B is now on Poe! This new model is the first frontier-level open source AI model, outperforming leading models across multiple benchmarks. It excels in long-context understanding, mathematical reasoning, multilingual tasks, and general knowledge. (1/2) https://t.co/O967DKoxeK" / X
Poe on X: "You can now upscale your images on Poe with the @playground_ai upscaler! This bot will sharpen and add ultra-fine detail to any uploaded or generated image, increasing the resolution by 4x. It can also improve the definition and clarity of low-res photos. (1/2) https://t.co/UCscHa7br9" / X