First Introducing ourselves

Raw Text

.css-onkibi{width:1em;height:1em;display:inline-block;line-height:1em;-webkit-flex-shrink:0;-ms-flex-negative:0;flex-shrink:0;color:currentColor;vertical-align:middle;} go back

Hello, we are the Dice team. I write for the first time about what we're doing. The reason for writing is to let people who have similar thoughts know our thoughts. We had provided deep learning based B2B search Saas in the form of development and demo for 2 months to enable semantic search. Currently, we are developing new services by pivoting to see the speed of growth of deep learning and the impact that will occur in the future. Our goal is to create an AGI (Artificial General Intelligence) assistant that can assist people in their daily lives in the next 15 years. To achieve this goal, they are breaking down the progress into several stages and starting with the creation of a Quora-like platform using conversational AI. We provide a platform with AI experts for each area with superior knowledge. People can ask questions and get answers easily and quickly through conversations. Our main focus is on solving the problem of "online fatigue" that results from the time and effort required to sift through vast amounts of information on the internet. Our service utilizes advanced deep learning techniques, including a new foundation model (language model) pre-training algorithm that allows for low parameter models with better performance, lower cost and faster latency speed, to provide users with expert-level answers to their questions, as well as summaries and recommendations for products and services. In the future, we plan to expand our service to include a full-featured online assistant, ultimately building AGI (including hardware) which will be able to do everything in online and offline. I'm sure it's the most valuable service for anyone who doesn't get the information they need because they don't have the experts around them, and who feels tired online looking for the information they need with too much information. First , we are currently engaged in research to create a more powerful model. Our primary research focus is on creating high-performance language models with 100x fewer parameters. The current language models are known as "ultra-large langauge models" with a large number of parameters. This leads to problems such as slow speed and high cost, which makes an unsustainable business model and does not provide a great user experience. We are researching on the question of "Why current LLM's are giant parameter models?" and "How to solve these issues" and will provide more detailed thoughts and directions in their next article. Secondly , we are working on building models with more diverse user-computer interactions and online action ability including reservation and paying. To better understand user's question and provide better answers, we aim to research other domains that can help user experience like accepting inputs such as photos, performing specific actions, providing references for answers, and enabling answers that are based on the latest information. We work on developing various features such as truthful and harmless, multi-modal, step by step reasoning, video understanding, answers with generated image, etc. Thirdly , and most importantly, we are user-centric. We always think in a user-driven way. Although there are many things to do from a technical perspective, we will continue to have ways to confirm the user's needs and gain insights in parallel. To do this, we have quickly made and launched two products. Web AMA : https://chrome.google.com/webstore/detail/webama-web-assistant-like/abgnbgkcmpikahkbccpjhbekkhimbjpi/related?hl=en&authuser=2 Solomon : https://solomongpt.com If you have similar thoughts or are interested, please feel free to contact us at the email below! contact@diceyai.com Currently, we fully dedicates and works around 17 hours per day, with all members being able to handle engineering tasks with deep learning research ablities. Thank you for reading :)

Single Line Text

.css-onkibi{width:1em;height:1em;display:inline-block;line-height:1em;-webkit-flex-shrink:0;-ms-flex-negative:0;flex-shrink:0;color:currentColor;vertical-align:middle;} go back. Hello, we are the Dice team. I write for the first time about what we're doing. The reason for writing is to let people who have similar thoughts know our thoughts. We had provided deep learning based B2B search Saas in the form of development and demo for 2 months to enable semantic search. Currently, we are developing new services by pivoting to see the speed of growth of deep learning and the impact that will occur in the future. Our goal is to create an AGI (Artificial General Intelligence) assistant that can assist people in their daily lives in the next 15 years. To achieve this goal, they are breaking down the progress into several stages and starting with the creation of a Quora-like platform using conversational AI. We provide a platform with AI experts for each area with superior knowledge. People can ask questions and get answers easily and quickly through conversations. Our main focus is on solving the problem of "online fatigue" that results from the time and effort required to sift through vast amounts of information on the internet. Our service utilizes advanced deep learning techniques, including a new foundation model (language model) pre-training algorithm that allows for low parameter models with better performance, lower cost and faster latency speed, to provide users with expert-level answers to their questions, as well as summaries and recommendations for products and services. In the future, we plan to expand our service to include a full-featured online assistant, ultimately building AGI (including hardware) which will be able to do everything in online and offline. I'm sure it's the most valuable service for anyone who doesn't get the information they need because they don't have the experts around them, and who feels tired online looking for the information they need with too much information. First , we are currently engaged in research to create a more powerful model. Our primary research focus is on creating high-performance language models with 100x fewer parameters. The current language models are known as "ultra-large langauge models" with a large number of parameters. This leads to problems such as slow speed and high cost, which makes an unsustainable business model and does not provide a great user experience. We are researching on the question of "Why current LLM's are giant parameter models?" and "How to solve these issues" and will provide more detailed thoughts and directions in their next article. Secondly , we are working on building models with more diverse user-computer interactions and online action ability including reservation and paying. To better understand user's question and provide better answers, we aim to research other domains that can help user experience like accepting inputs such as photos, performing specific actions, providing references for answers, and enabling answers that are based on the latest information. We work on developing various features such as truthful and harmless, multi-modal, step by step reasoning, video understanding, answers with generated image, etc. Thirdly , and most importantly, we are user-centric. We always think in a user-driven way. Although there are many things to do from a technical perspective, we will continue to have ways to confirm the user's needs and gain insights in parallel. To do this, we have quickly made and launched two products. Web AMA : https://chrome.google.com/webstore/detail/webama-web-assistant-like/abgnbgkcmpikahkbccpjhbekkhimbjpi/related?hl=en&authuser=2 Solomon : https://solomongpt.com If you have similar thoughts or are interested, please feel free to contact us at the email below! contact@diceyai.com Currently, we fully dedicates and works around 17 hours per day, with all members being able to handle engineering tasks with deep learning research ablities. Thank you for reading :)