I have heard from friends about the capabilities that ChatGPT has for helping programmers write code. Therefore, in this blogpost, I will test ChatGPT’s coding knowledge and accuracy. To do this, I will ask ChatGPT to create some simple programs from scratch, and test if the code that ChatGPT generates actually works. Usually, one of the first things that you learn when programming is how to program a calculator. Therefore, I will ask ChatGPT to program a calculator in Python as this is a common language and I am planning on learning it during the course of my BIM studies.

The inputs is as follows: Program a calculator in Java and Program a calculator in python.

This is the result:

When I run  the python script it gives me the following result:

As you can see, the calculator works, and I decided to also test the dividing by zero function just to be sure. I find it interesting that ChatGPT is capable of creating this code in a couple of seconds, and also does not forget to include that division by 0 is impossible. Now that we know that this works, we can try to test further and see what the limits are of how much ChatGPT can do without me having to write a single line of code.

Therefore, to improve the calculator, I asked ChatGPT to add buttons so that we do not have to type the operation we have to perform. This was ChatGPT’s output:

When testing the code, we get a working calculator which has buttons.

I find this quite impressive already. Therefore, I want to see now what would happen if I ask it that the inputted numbers should be able to be clicked on in the screen.

As you can see, ChatGPT just created a fully working basic calculator, and it did so in 3 prompts. I opened another ChatGPT window and made one specific prompt which was: Create a calculator with buttons for input and operations in python. This yielded the same result, showing that if you are specific, ChatGPT can be a very powerful tool to help you with programming. All in all, I am impressed that ChatGPT knows how to create a GUI and working calculator with little input.

I always look forward to that time of the day when I can be in my kitchen and cook a meal. This activity acts as a creative release and helps me get rid of the day’s stress and anxiety. I enjoy watching cooking videos with recipes from all over the world to learn about other cultures and get inspiration for my next meal. Therefore, when I am faced with an almost empty fridge and a tight budget, I use this opportunity to apply the knowledge and library of recipes that I gathered throughout the years.

However, since I moved abroad, I met many students that hate the fact that they need to cook three meals a day for themselves. In most cases, they either choose to buy easy-to-prepare dishes (which can become costly very fast) or rely on pasta and pesto. Most of them told me that they don’t enjoy planning meals, but if they were given precise instructions and grocery lists, they might start cooking more.

Therefore, I decided that I would experiment with AI recipes generators to see if they might help students eat better. I used ChefGPT to generate a meal with the ingredients that I had in my fridge. The interface is easy to use and I like the fact that you can choose the kitchen tools that you have at home, as many students do not have fully equipped kitchens.

At a quick glance, I find the combination of ingredients a little interesting, especially because I told ChefGPT to give me a recipe that best mixes all the ingredients, without including everything. However, when searching on Google, it shows that these ingredients do work together well and help create nutritious meals.

Another aspect that I like about the recipe from ChefGPT is that it includes the macros for the recipe, as I know many students that pay attention to their nutrition (however, I am not an expert so I cannot assess if the AI generator is correct).

I believe that tools such as this one can be used when shopping for grocery. For example, one could go to the store, check which products are on sale that day, input the ingredients in the app and check the variety of meals that it can create. This might be a very interesting experiment and a great way to save money and reduce food waste.

Overall, I believe that AI generating tools can help people struggling with culinary creativity to create a variety of nutritious meals without throwing away unused ingredients. However, the program’s outputs should be taken with a grain of salt as sometimes recipes might sound better than they are in real-life practice.

By Alexander Kristiansen

Creating a large userbase for a digital platform is no small task. Building a large audience requires patience and diligence. Continued support is required for a growing population, using user feedback and continues updating of services. New features entice users to stay with the platform, as it can further distinguish itself from the competition. Once you have established a large and thriving userbase, it can be considered valuable. This phenomenon is also known as network effects. One interesting case showing this is that of X (Formerly known as Twitter).

            From humble beginnings in 2006, the company was founded by Jack Dorsey, Noah Glass, Biz Stone, and Evan Williams. It exploded in popularity, and by 2012 already had 200 million active users (Vanian, 2022). The company was then purchased by Elon Musk in 2022 for $44 billion. This was a very polarizing move for Twitter’s users, as Musk is known for having many controversial opinions and business practices. He promptly changed the name of the company to X (because it’s his favorite letter) and established a paid tier system with blue verification check marks. These controversies were not going unnoticed by X’s competitors, and Meta promptly aimed to take advantage of the situation. Through their already established platform of Instagram, they launched Threads. It was essentially a direct copy of X, large being text based, and having comment chains where the highest upvoted responses would appear on top. Many disenfranchised X users, among others, moved to the competitor, and Threads amassed 100 million users in 5 days (Leigh, 2023). This shattered all records, and Threads was poised to become the new big social media. However, the userbase has dropped by 82%, and now there are only 8 million daily active users (Leigh, 2023). Ultimately, X remains dominant in the segment, showing the power of network effects. Threads was not able to compete in terms of content available on the platform, which is a by-product of having a large community. This content production will only garner more and more users, further established this platforms network dominance. It will take a new platform with some serious innovation to take the throne from X.

Reference:

Leigh, D. (2023). We Asked the Experts: What Happened to Threads?. TechRound. Available from: https://techround.co.uk/news/we-asked-the-experts-what-happened-to-threads/#:~:text=However%2C%20since%20that%20explosive%20start,who%20desired%20a%20desktop%20experience.

Vanian, J (2022). Twitter is now owned by Elon Musk here’s a brief history from the app’s founding in 2006 to the present. NBC. Available from: https://www.cnbc.com/2022/10/29/a-brief-history-of-twitter-from-its-founding-in-2006-to-musk-takeover.html

Art and creativity has been my exploratory space ever since childhood. Having loved painting, passed advanced level singing exam as well as a trained Bharatnatyam dancer; art will always remain special to me. I have always taken keen interest in what is new and upcoming trends in this arena. While one might think that art is something extremely human specific and quality differs from person to person, I was in for a huge surprise when I noticed that technology has made its entry over this space as well.

When I recently met my Bharatnatyam dance teacher after long, she made me realize that how choreography has become so easy these days, and you don’t need to put in hours behind synthesizing a beautiful dance. She mentioned that motion capture technology with generative AI is able to sense dancers’ movements, explore and generate completely unique sequences and assist in choreography. In fact, the world-famous British choreographer Wayne Mcgregor has created an AI powered tool in collaboration with Google which feeds on Mcgregor’s past dance performances to identify patterns and inspire new steps and sequences.

Since she knew my interest in music, and how I loved to dance on my own sung music. She told me about a new AI system named IBM Watson Beat which feeds on dance performances, analyzed the emotional content and generated appropriate music compositions for the dance. This really surprised me regarding the extent technology has reached.

When I went back home, I had to put my research glasses on and dive deep into how has generative AI disrupted art further. I came to realize craft was also not let alone. There is a generative AI application named DeepArt which can transform regular photos into artwork. It needs some reference artwork and can apply the style, precision and visual characteristics to an ordinary photograph submitted by the user.

This entire new piece of news took me by surprise and lead me to think whether generative AI can replace human talent. I still think not quite so, because the kind of uniqueness and creativity a human mind can bring on the table is quite impossible for AI since it ultimately feeds on reference work. But there is still a lot of time in the future to explore and probably a lot of surprises planned for us by emerging technology!

Hailing from IIM Bangalore, I interned at Citibank as an investment banker for 2 months. Finance industry in general is perceived to be extremely strenuous; relying severely on human labor. My first corporate experience with GenAI was during my internship itself. During this period, I realized how GenAI like ChatGPT, assists investment bankers in reducing their workload, increasing impact and efficiency simultaneously.

I was staffed on a project to create a company pitch for private placement of an Indian bathware firm. During this project, for the purposes of preparing the industry overview, the analyst working on the project used ChatGPT for generating research reports on Indian bathware industry and its past performance. ChatGPT also collated market commentary regarding the future of this industry. I was surprised at how this saved a lot of time that could have been spent in internet research pertaining to market conditions. ChatGPT also helped in providing auto-document summarization wherein the analyst uploaded 2-3 broker reports commenting on industry projected financials; this assisted him in scanning through these lengthy reports efficiently and gaining significant data at such low time and effort.

Moreover, the use of GenAI did not only limit to the research portion of the pitch, in-fact; it surprisingly extended to the financial or quantitative modelling portion as well. ChatGPT was used to create extreme scenarios and stress test the model under adverse scenarios. This helped to build robustness and made the pitch stronger by accounting for multiple possibilities.

Thus, during the course of these 2 months I realized how generative AI like ChatGPT is so helpful in the day-to-day life of an investment banker. It helps to accelerate tasks and can be used in tandem with the human mind. Investment banking industry is heavily reliant on human mind and touch, since successful client relations are a must; however, tools like these can help relieve them by reducing their burden and increasing the work quality simultaneously.

During my two-month long internship at a financial advisory and consulting firm, generative AI tools were gaining a lot of traction. Hence the firm decided to buy a premium account for the same and include it in the library of resources of the firm based on recommendations of junior analysts. My coverage was in the industries of digital tech and financial services in the firm and one of our client’s needed assistances with a go to market strategy of the new product in their platform. The tool was used extensively by some analysts and interns while some resisted the tool due to its accuracy and inability to provided complex solutions.

It was efficient in providing summary of industry reports, image solutions as well. For instance, if certain information was needed while creating slides for a client for instance what is the “total market size of cloud-based computing services in America”, prompts were directly fed to get an idea however it could be used only as an aid and not as a source in these slides. It provided us with a good starting point and source of relevant information. Some analysts even used it to create a rough template slide to work and ideate upon. My discussions with seniors at the firm made me ask my seniors that the tasks which are trivial in nature and demand a lot of effort in terms of time and research, can it not be automated. The senior director of my team answered and provided a new perspective to the same age old question. He said that the value we create for our client is not the solution, but our ability and competence which we provide, AI can read up a ton of documents and provide summaries and relevant data from it, but an analyst only can select and decipher what is relevant to the client and the problem. The learnings from the strategy will be different for each analyst in the deal which is the value added to client and not the strategy itself. With improvements in its deep learning, image processing models, it can certainly become accurate and efficient.
Surprisingly weeks after my internship was over, the final solution was submitted to the client and chatGPT recommended the same go-to market strategy for the client when I prompted it with the similar problem statement. So, the question arises, can it replace or outdate consulting and similar jobs. The answer is complex, but the nuances of a human interaction, expertise and experience can’t be always substituted with AI. This example can be a case where the solutions match but companies certainly benefit from such tools in their arsenal as an aid and rather as the complete solution provider.

The name of the client has not been shared due to NDA.