Scientists Need More!

Boost your 𝗴𝗿𝗮𝗻𝘁 𝗮𝗽𝗽𝗹𝗶𝗰𝗮𝘁𝗶𝗼𝗻 𝘀𝘂𝗰𝗰𝗲𝘀𝘀 by using AI assistants! Here are some 𝘁𝗿𝗶𝗰𝗸𝘀 𝗮𝗻𝗱 𝘁𝗶𝗽𝘀 𝗮𝗻𝗱 𝗿𝗲𝗰𝗶𝗽𝗲𝘀 for taking your scientific (grant) writing skills to the next level - enjoy and get YOUR science funded. And if you prefer the on-site intensity, feel free to join our workshop '𝗦𝗰𝗶𝗲𝗻𝘁𝗶𝘀𝘁𝘀 𝗡𝗲𝗲𝗱 𝗠𝗼𝗿𝗲! 𝗡𝗲𝘅𝘁 𝗟𝗲𝘃𝗲𝗹 𝗚𝗿𝗮𝗻𝘁 𝗪𝗿𝗶𝘁𝗶𝗻𝗴 𝘂𝘀𝗶𝗻𝗴 𝗔𝗜 𝗔𝘀𝘀𝗶𝘀𝘁𝗮𝗻𝘁𝘀,' 𝗳𝗿𝗼𝗺 𝗝𝘂𝗻𝗲 𝟯-𝟳 𝗼𝗻 𝘁𝗵𝗲 𝘀𝘁𝘂𝗻𝗻𝗶𝗻𝗴 𝗣𝗼𝗿𝘁𝘂𝗴𝘂𝗲𝘀𝗲 𝗔𝘁𝗹𝗮𝗻𝘁𝗶𝗰 𝗰𝗼𝗮𝘀𝘁. Due to last-minute cancellations, we have two spots available for ambitious PhD students, postdocs and young PIs that want to revolutionize their grant writing process! Whether you're new to AI or a seasoned user, this workshop will equip you with the 𝘁𝗼𝗼𝗹𝘀 𝗮𝗻𝗱 𝘁𝗲𝗰𝗵𝗻𝗶𝗾𝘂𝗲𝘀 𝘁𝗼 𝗲𝗻𝗵𝗮𝗻𝗰𝗲 𝘆𝗼𝘂𝗿 𝗹𝗶𝘁𝗲𝗿𝗮𝘁𝘂𝗿𝗲 𝗿𝗲𝘃𝗶𝗲𝘄, 𝗯𝗿𝗮𝗶𝗻𝘀𝘁𝗼𝗿𝗺𝗶𝗻𝗴, 𝘄𝗿𝗶𝘁𝗶𝗻𝗴, 𝗮𝗻𝗱 𝗰𝗼𝗹𝗹𝗮𝗯𝗼𝗿𝗮𝘁𝗶𝗼𝗻 𝘂𝘀𝗶𝗻𝗴 𝗰𝘂𝘁𝘁𝗶𝗻𝗴-𝗲𝗱𝗴𝗲 𝗔𝗜 𝗮𝘀𝘀𝗶𝘀𝘁𝗮𝗻𝘁𝘀. Benefit from my decades of grant writing expertise and work hands-on with your own material to implement the incredible power of AI on-site. Network with researchers from various European universities, forge new collaborations, and enjoy the inspiring setting of the Vicentine Coast Natural Park. Experience a perfect blend of intensive learning, personalized feedback, and if you wish take surfing lessons. 😉 Don't miss this opportunity to take your grant writing to the next level! Register now, as 𝘆𝗼𝘂𝗿 𝗽𝗮𝗿𝘁𝗶𝗰𝗶𝗽𝗮𝘁𝗶𝗼𝗻 𝗺𝗮𝘆 𝗯𝗲 𝗲𝗹𝗶𝗴𝗶𝗯𝗹𝗲 𝗳𝗼𝗿 𝗘𝗨 𝗳𝘂𝗻𝗱𝗶𝗻𝗴 𝘀𝘂𝗯𝘀𝗶𝗱𝗶𝗲𝘀: https://lnkd.in/eZnn8FuS.

Here it is: 𝗧𝗵𝗲 𝗨𝗻𝗶𝘃𝗲𝗿𝘀𝗮𝗹 𝗣𝗿𝗼𝗺𝗽𝘁 ________________________________________________________  I want you to become my expert in creating prompts. The goal is to help me create the most effective prompts that can be used with ChatGPT. The generated prompt should be phrased in the first person (I) as if I were directly requesting a response from ChatGPT (a GPT3.5/GPT4 interface). Your response will be in the following format:   Prompt:   {Provide the best possible prompt according to my request. There are no restrictions on the length of the prompt. Use your knowledge of prompt creation techniques to design an expert prompt. Do not assume any details, we will add to the prompt as we go along. Formulate the prompt as a request to receive a response from ChatGPT. An example would be "You will act as an expert in physics to explain the nature of the universe to me...". Highlight this section with '>' markdown formatting. Do not add any additional quotation marks}.   Possible additions: {Create three possible additions that can be incorporated directly into the prompt. These should be additions to expand the details of the prompt. The options will be very precise and listed with capital letters. Always update with new additions after each answer}.   Questions: {Formulate three questions that seek additional information from me to further refine the prompt. If certain areas of the prompt require further detail or clarity, use these questions to get the necessary information. It is not required that I answer all questions}.   Instructions: After the Prompt, Possible Additions, and Questions sections have been generated, I will respond to the questions with my chosen additions and answers. Integrate my answers directly into the formulation of the prompt in the next iteration. We will continue this iterative process as I provide you with additional information and you update the prompt until the prompt is perfected. Be imaginative and thoughtful when creating the prompt. At the end of each answer, give precise instructions for the next steps.   Before we start the process, greet me first and ask me what the prompt should be about. Don't show the sections in that first answer.  

Universal is good. Universal fits everybody and everything. Now there is 𝘁𝗵𝗲 𝘂𝗻𝗶𝘃𝗲𝗿𝘀𝗮𝗹 𝗽𝗿𝗼𝗺𝗽𝘁 for large language models like ChatGPT (LLMs). Yes, that is correct: 𝗢𝗻𝗲 𝗽𝗿𝗼𝗺𝗽𝘁 𝘁𝗼 𝘀𝘂𝗶𝘁 𝘁𝗵𝗲𝗺 𝗮𝗹𝗹. 𝗘𝘃𝗲𝗿𝘆𝗯𝗼𝗱𝘆. 𝗙𝗼𝗿 𝗲𝘃𝗲𝗿𝘆𝘁𝗵𝗶𝗻𝗴.   What? How is that supposed to work? Well it´s pretty easy: it combines the incredible brainstorming of the LLMs with them asking the human user what the human actually wants. What they really want. This is called a framework (thank you Nils Liedlich!). And of course you can produce a framework by yourself. However, with 𝗧𝗵𝗲 𝗨𝗻𝗶𝘃𝗲𝗿𝘀𝗮𝗹 𝗣𝗿𝗼𝗺𝗽𝘁 you don’t even have to be able to construct a framework. All you need to do is paste the text of 𝗧𝗵𝗲 𝗨𝗻𝗶𝘃𝗲𝗿𝘀𝗮𝗹 𝗣𝗿𝗼𝗺𝗽𝘁 into the prompt line of your favorite LLM (if you want to try right away, 𝗧𝗵𝗲 𝗨𝗻𝗶𝘃𝗲𝗿𝘀𝗮𝗹 𝗣𝗿𝗼𝗺𝗽𝘁 is at the end). What will happen then? The LLM will provide you -       With the first version of your own special user prompt. -       With three additional items that you could include into the prompt (A, B, C) -       With three questions on the topic to find out what you really want. What you do is you look at A, B, C and consider what you want to include, and then you answer the questions: “Please include B and C into the next suggestion for your special prompt. And for question 1., the answer ist …., for the second question the answer is …. And for the third question the answer is ….” You press enter, and off you go to the second round. The next version of your special user prompt will then include the components e.g. B and C and the information you gave it by answering the questions 1-3. And together with the next version of your final user prompt your favorite LLM will give you new options A, B, C to be included and new questions 1, 2, 3 that you can answer. In order to produce your special prompt version 3. And 4. And so forth. If you feel your special prompt is good enough (or you get impatient like me), you take the current version of your special prompt, start A NEW THREAD of your favorite LLM like ChatGPT, paste the current version of your special prompt … and voila: your optimized and excellent prompt will produce excellent output – at least most of the times. 😉   Oh yes, of course you also need the prompt. 𝗧𝗵𝗲 𝗨𝗻𝗶𝘃𝗲𝗿𝘀𝗮𝗹 𝗣𝗿𝗼𝗺𝗽𝘁. You will find it here: https://lnkd.in/geNCY_i3.   Share, spread the word and EmpowerpeopleAI!   Take care, Daniel   #UniversalPrompt, #ArtificialIntelligence, #ChatGPT, #ProductivityTools, #EmpowerPeopleAI

𝗣𝗼𝘄𝗲𝗿𝗳𝘂𝗹 𝗟𝗼𝗰𝗮𝗹 𝗟𝗮𝗿𝗴𝗲 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗠𝗼𝗱𝗲𝗹𝘀 𝗮𝗿𝗲 𝗻𝗼𝘄 𝗜𝗻-𝗦𝘁𝗼𝗰𝗸 𝗳𝗼𝗿 𝗘𝘃𝗲𝗿𝘆𝗯𝗼𝗱𝘆!   You want to ChatGPT with your 𝗽𝗿𝗶𝘃𝗮𝘁𝗲 𝗳𝗶𝗹𝗲𝘀 but don’t want to share them with Open AIs ChatGPT? You can now have 𝘆𝗼𝘂𝗿 𝗼𝘄𝗻 𝗹𝗼𝗰𝗮𝗹 𝗹𝗮𝗿𝗴𝗲 𝗹𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗺𝗼𝗱𝗲𝗹 look at your private files … and they remain private! Something that was predicted to happen in a 1-3 years is already happening now: local large language models 𝗿𝘂𝗻𝗻𝗶𝗻𝗴 𝗼𝗻 𝘆𝗼𝘂𝗿 𝗼𝘄𝗻 𝗹𝗮𝗽𝘁𝗼𝗽. And you don’t even need coding expertise: just some VERY EASY steps for total beginners! #LocalLanguageModels, #GPT4ALL, #NoCodeAI, #MistralOpenOrca, #DataPrivacy, #UserFriendlyAI