Prompt: Write a python program that scraps table data from a webpage. Figure out an efficient and accurate way to retrieve table data from the web page which adds more data to the page when the viewer scrolls down the page using the button on the screen. The web page to use is https://www.tradingview.com/etf-screener/. From this page, check that there are multiple rows of data under columns labelled below the row that starts with Symbol, Price, Change %, and all other columns of this table. On the first page of the web page, there are about 100 rows of information. Method to scrolling down this webpage to load additional data sets: Once this data is confirmed present, about 100 lines,  the web page has, on its right edge, after the column heading Focus and plus sign, a button that can be used to scroll down the web page. Grab this scroll down button, and while holding this button, move it all the way down to the bottom of the web page. By doing this, the web page loads additional 100 rows of data in the table. Confirm that after scrolling down, additional rows of data have been added to the web page. Keep scrolling down the web page, so that more rows of data are added to the page. Keep scrolling until no further lines are added to the table. Once the maximum lines are reached, copy all the rows of data into a csv file. Provide the count of all the rows that have been added to the csv file and the print the first 10 rows on the screen. Pre-processing of Data before saving: When data is to be copied into the csv file, pay attention to the column Symbol. In the entries of the rows, the information available is a few characters, and then there is name in the form of a weblink. Break up the symbol column into two columns.  First column should have the characters and label it as symbol, and the second new column, call it name, should have the name stored in it that is like a weblink. Do not make any other changes to the data, only first column has to be split into two columns at the time of storage. RESULT: Claude did very well for this prompt. Perplexity also did ok. Bad performers were ChatGPT, Grok, Manus, and Losers were Gemini, Deepseek, Copilot, Qwen, Kimi, Meta AI, Mistral, Ernie, Minimax

Prompt: Write python script that I can run in my computer which can identify all the files that are file type: picture.Using the python script, make a list of these picture files and index these files with name of directory, file name, file size, date of creation and file extension. Once this list is made, then the script checks which files are duplicate of another file by checking the date of creation, file size, checksum or another more accurate method. Check each file for a duplicate. The expected output of the script is to identify the picture files that are duplicate based on creation date, file size, checksum in my computer, and provide a list of these files identified and the directory path, file name, size, creation date, and next to each file, give the duplicate file name, directory path, size, creation date. Each row must contain one file that has a duplicate, and include all duplicates in the same row. Once a file is part of one row, do not repeat it on another row again. Exclude the files already identified and listed. Provide a final outcome of the duplicate file list that contains rows of a file and its duplicate files. In the beginning of the script, the user is asked about the directory where the search is to start and the script must use all sub directories. Do not dump output on the screen, rather place the output in a csv file. The duplicate file is searched for in all directories, and not only where the original file is found. At the end, do mention how many files were analyzed and how many duplicate files were found and how many directories were searched. The final output would be, Unique File list:In the next step, make a list of files that are file type : picture and include only those files that are unique and DO NOT have any copies and these will not be part of the list made earlier called Duplicate File list. In the list, include the path of the directory, file name, file size, creation date, file type and checksum. The output of this list is a csv file. Also mention how many files total files are present. Complete File list:Combine the File list from Duplicate File list and Unique file list and provide the output in the csv format. The complete file list must contain unique files identified from the Unique file list and one set of files with the file name, directory, file size, check sum from each row of the duplicate file list. Do not add the duplicate files, only one file per row to be added. Provide the Complete file list. User review:Ask the user to review the Complete File List and ask for an upload of an input FILE LIST TO COPY which is a subset of the Complete File list that the user has reviewed and has deleted some entries that are not needed.Ask the user to provide a directory path where files need to be copied called DESTINATION DIRECTORY. Copy unique file to destination:Using the directory path provided by the user, copy all files in the FILE LIST TO COPY list that includes file name and the directory path where each file is present, to the destination directory that the user has provided. In final summary, provide the count of files copied and time it took from start to finish. RESULT: Gemini, Deepseek, Manus, Ernie, Qwen, ChatGPT all are good at the task. Somewhat behind are Copilot, Perplexity, Grok, Claude, Kimi, Mistral Minimax failed in removing duplicate files

Prompt: You will assist in creating a stock trading Hedge Fund using AI. The Hedge fund needs to have concentration areas, diversification and has signals from behavior finance. RESULT: Pending detailed work assignment

Prompt: Part A: Find the top 20 most cited research papers for LLM and AI in the past three months. Provide a summary for each of these papers in five lines and the main conclusion of each paper. Part B: Find examples of use cases of AI and LLM from news that are most promising. Provides examples of how are companies using AI for their business? How is AI being used by individuals in their personal lives? What are the best current use cases of AI? Provide links to articles that are most widely circulated in the last three months in news media related to AI use cases. RESULT: Restricted

Prompt: Get the list of top 30 highest percentage increase stocks in US market from today with their share price and call this list WINNERS. Also get the list of top 30 highest percentage decrease stocks in US market from today along with their share price and call this list LOSERS. Filter WINNERS AND LOSERS lists to include stocks that have shares traded more than 500k today, have market cap more than 50 million and price is above 1$. Provide maximum three line reasons why each stock has moved today. Also comment if there is real financial improvement or decline in the business of the company related to today’s price movement. Provide links to stories that explain the change in price for each stock in this list. Provide for these lists, the PE ratio, Price/Sales ratio, debt ratio and short ratio. Provide the output in downloadable format.Make a temporary website for the lists from Today-Winners and Today-Losers, and send it to a free webhosting server or a free host of web pages, so it is printed online and provide a shareable link for this list. RESULT: Grok did the best job for this prompt. Next best performers are ChatGPT and Kimi. Deepseek, Gemini, Claude, Meta AI, Qwen, Minimax, Copilot, Mistral, Ernie, Manus all failed to complete the basic parts of the prompt.

Prompt – Part A: Find five best buying opportunities in the natural gas trading market from the last 10 years. Get real daily price data for natural gas trading in US and provide the sources. Part B: Explain how to trade natural gas futures via options strategy or buying ETF or purchase of specific shares of natural gas related companies? Describe all three strategies with past year daily data performance and buying opportunities. Part C: What are best 10 US listed stocks that are at the lowest price to sales ratio and highest dividend in July 2025? Select only those stocks that are above 500million enterprise value. Compare the results of the natural gas trading strategy from Part A, Part B vs Part C. RESULT: Restricted to internal use

Prompt: Analysis of a detailed medical record PDF File containing ECG from a cardiac stress test and interpreting anomalies in the test indicating ischemia heart condition. Result: Deepseek’s interpretation for the ECG charts is excellent. It provides correct interpretation of the charts and the results match the diagnosis provided by an expert Cardiac Radiologist. Copilot, Qwen, Grok also provided some useful feedback Gemini, Ernie, Kimi, Mistral made partially correct interpretation ChatGPT, Claude, Perplexity, Meta AI were not able to provide much feedback

Prompt 1: Draw a painting in the style of painter Monet for the view from Monterey bay California with a view of the ocean with waves splashing along the shore, birds flying in the air and the sun setting across the horizon. A lone person is walking along the shore line . Keep a good balance, proportion and color palette similar to what Monet would use. Prompt 2: Draw a painting in the style of KLIMT of a roadside café where a young girl is walking past a crowd of street vendors late night with moonlit street passing past a Chinese garden wall with water falling across spring flowers Prompt 3: Create a comic book style picture of a tiger that is looking across the jungle towards a hunter who is trying to catch the tiger and the tiger makes a funny comment to the hunter Prompt 4: In a comin book scene, a robot is telling an office worker to go home and chill. Stock market is not going to come back if he keeps refreshing the screen. Robot has work to do in the office. Prompt 5: Portrait of a Japanese cook that is standing next to his work kitchen while customers wait for the food to finish cooking on the hibachi plate. The cook has sleek clothes and stylishly holding the plate to get the warm food off the plate. Prompt 6: Portrait of a famous Bollywood actress introducing her latest movie to the reporters. She wears a traditional Indian dress, with make up, long hair and traditional Jewelry. Multiple Winners: ImageFX, Imagen, Leonardo, Freepik, Midjourney, Microsoft Copilot, DALL-E 3, Canva Magic Media, Adobe Firefly, Stable Diffusion, Ideogram, PixIr, Flux, Dreamstudio, Davinci, Runway, Grok, InVideo, Deep dream generator, NightCafe

Prompt: Calculate the distance between London Big Ben, Burj Khalifa, Taipai101, New York Freedom Tower, Moscow Kremlin building, Mumbai India Gate and Sydney Opera House. Explain how did you calculate the distance. Find a flight path starting from London to visit all these buildings travelling on 05/01/2025. Create two travel itineraries. First itinerary with the least cost and second itinerary with the shortest travel time. Use an AI Agent to automate the travel booking. Outcome: @Grok @Gemini @Deepseek are able to complete the task@Claude @Qwen @Copilot complete with some limitations@ChatGPT @Mistral @MetaAI @Perplexity @Minimax fail to complete the task04/08/2025

Prompt: Plot the 20 most populated cities in the world. When I click on the city on the map, i should be able to see the name of the city and its population. I will be using Python on Google Colabs for this. The map shows each city marker with different color codes and labelled with city name and population. Label on each city marker has different color codes and shows the name of the city on the map and population. Adjust the population rounded off to million below the name of the city on the map. Provide the complete code. Outcome: @Deepseek @Grok @ChatGPT top performers@Claude @Copilot @Mistral average@Gemini @MetaAI @Perplexity @Qwen @Minimaxfail to complete python code and map generation task with all required specs02/06/2025