No ratings yet.

I’ve recently been diving into the world of AI art, and I have to say—it’s pretty exciting! There’s something fascinating about typing out a few words and watching AI turn them into art. As a fan of starry nights and stunning landscapes, I decided to see what Leonardo AI could create. I found myself playing around with different prompts and getting more and more amazed at how close AI can come to producing exactly what I have in mind.

Let me take you through my journey, sharing the images I generated at each stage.

The First Attempt: A Realistic Starry Night

I kicked things off with a simple idea—a starry night over some mountain backdrops. The prompt was straightforward enough, so I started with:

Prompt 1:

“Create a realistic starry night with mountains as backdrop.”

Result:
But, to be honest, the result wasn’t quite what I had in mind. Instead of a crisp, lifelike image, it ended up looking more like a stylized painting. The stars felt too dreamy, and the overall vibe wasn’t really close to photorealism. It was clear I needed to get a bit more specific to really capture what I was envisioning.

Getting It Right: Ladakh’s Starry Night

Determined to get the image just right, I decided to focus my prompt and get into the details. I thought Ladakh’s rugged mountains against a glowing night sky would be the perfect subject for a do-over:

Prompt 2:

“I want a photorealistic starry night image with the mountains of Ladakh in the backdrop. I don’t want constellations, rather a galaxy view. Make it like it’s a photo taken by a pro photographer.”

Result:
This time, it was exactly what I was after! The sky was rich with stars without being overwhelming, and the galaxy view made it all the more mesmerizing. The mountains of Ladakh were beautifully lit by moonlight, creating a surreal but totally realistic scene. It looked like something you’d find in a photography magazine, and I was absolutely thrilled with how it all came together.

Channeling Van Gogh in Prague

After recently visiting Prague, I wanted to try something totally different. Being a huge Van Gogh fan, I thought: why not see Prague through his eyes? So I gave the AI another shot, this time asking it to create something in a classic painterly style.

Prompt 3:

“Create a painting of Prague in the style of Van Gogh, with his swirling skies and vivid colors.”

Result:
And wow, the AI really delivered! It transformed the beautiful architecture of Prague into a vibrant, swirling masterpiece, as if Van Gogh himself had taken a trip there. The skies were animated with bold strokes, and the colors were just as vivid as I’d hoped. It was a perfect blend of Prague’s charm and Van Gogh’s iconic style.

AI Art: Enhancing Creativity, Not Replacing Artists

As I wrapped up my AI art adventures, I couldn’t help but think about the ongoing debate: will AI replace human artists? Personally, I don’t think so. Sure, the quality of AI-generated art is impressive, but it still lacks that emotional depth and personal touch that only a human can bring to their work. AI is great for quick results and can even spark new ideas, but it doesn’t quite match the creative spark or intent behind human-made art.

I see AI art as more of a creative assistant—a tool that makes art more accessible, even for those of us without traditional artistic skills (like me). It opens up new possibilities and approaches rather than limiting creativity.

For now, I’m excited to keep experimenting and see what this tech has to offer!

No ratings yet.

Here is a quick look at how a friend of mine uses generative AI to support her legal work at a big law firm. While AI doesn’t yet have the ability to provide reliable case law or specific judgments, it shines in simplifying and summarizing lengthy legal texts. Here is how her process usually works:

When she gets a request from a client, she inputs the questions into a GenAI such as Claude. It provides a general overview and suggests which laws might be relevant, which is usually a pretty good starting point. Then, she dives deeper into legal databases, looking for specific judgments and articles. After identifying useful sections, GenAI helps her craft a clear, concise summary to send back to the client. Legal workers should be aware of the potential risks when dealing with GenAI to support their workflow. Firstly, you should always make sure to omit sensitive client information from your prompt to avoid potentially breaking confidentiality and non-disclosure agreements. In addition, as with other GenAI outputs, you should check for hallucination to avoid damage to the firm’s reputation. Finally, consider that GenAI is usually trained on data that is often more than a year old, which can make it difficult to use in a fast-changing environment such as the legal field.

One of the major advantages is that AI can simplify complex legal language, making it more digestible for non-lawyers. This makes communication with clients much easier and helps them understand their legal situation more effectively. Furthermore, GenAI could have the potential to dramatically reduce working hours for lawyers in big law firms, with a lot of them currently working as much as 80 hours a week, leading to numerous burn-outs (Bottorff, 2024). In short, generative AI tools have become a great tool for efficiency, but it still needs a human touch to ensure accuracy and depth.

References:

Bottorff, S. (2024, February 7). How to survive big law Hours | LawMatics. Lawmatics | Engage, Convert, Nurture, Grow • Best in class Client Intake and Marketing Automation for the modern law firm. https://www.lawmatics.com/blog/how-to-survive-big-law-hours/

No ratings yet.

The way we purchase goods is forever altered by the explosion of digital platforms. Whether it is Marktplaats, E-bay, Amazon or Airbnb, the new models of connecting buyers and sellers is truly revolutionary. In this blogpost, we will explore the impact of one such platform called Marktplaats. If you’re not familiar with Marktplaats, in essence it is a platform where people can sell and purchase primarily second hand items through an auctioning system.

One of the biggest impacts, Marktplaats has had on our society is the promotion of sustainability through a circular economy. Whether it is furniture, toys, or even cars, the accessibility to purchase and sell has never been lower. By making the process of selling your obsolete items easier, it has encouraged people to not throw away their old stuff. On the side of the buyer, it has made sure that finding what you need and placing a bid on the item is very easy. Furthermore, the site also facilitates communication between a buyer and seller, reducing any possible frictions that may arise with other methods. If you have a question about the product, you can simply ask about it within the application.
The seller of an item also has the freedom of choosing the kind of online market Marktplaats utilizes, you may set a fixed price, an auction, or use direct contact to negotiate a price. This has the benefit of maximising the seller’s surplus, further promoting selling instead of throwing away. 

All of these features combine into a ubiquitous use of the application within the Netherlands, increasing sustainability one product at a time.

Besides individuals that make use of Marktplaats small businesses also have their place within the platform. By increasing the ease of access to selling and advertising, it removes the need for a physical store for many small businesses. This allows for people to easily sell items of their side projects, whether it is art woodworking or pottery, there is no need to set up your own website or have a store.

Unfortunately, there is one major drawback of the platform, and that is the amount of information asymmetry that occurs when making a purchase. Whether it is used cars, electronics or any kind of machine, while you may be able to test what you’re about to purchase, the items may have defects that do not show until later that the seller doesn’t disclose. Marktplaats does try to combat this kind of behaviour by implementing reviews for both buyers and sellers and having payment insurance when items are sent via the post office. Of course these solutions does not entirely solve the problem, yet they are a step in the right direction.

Overall, Marktplaats is a major contributor to reducing consumerism and improving sustainability within the Netherlands while also promoting small business growth.

References
https://pixabay.com/illustrations/dollar-consumerism-business-1644801/
https://duurzaamheid.nl/artikelen/op-marktplaats-komt-de-circulaire-economie-van-nature-tot-leven/

5/5 (1)

Artificial intelligence has made its footing in the stock market, driven by algorithmic trading at its very core. More and more investors and hedge funds now turn to the might of AI-driven algorithms for quicker, more incisive decisions on trading. These algorithms can process huge quantities of information, identify patterns, and execute trades in microseconds, something no human can do. The question that piqued my interest was whether these algorithmic models can outperform even today’s best traders.

In many of these cases, the answer to the question would be an easy yes, as seen in the case of the hedge fund Renaissance Technologies, popularly known for its Medallion Fund, which makes investment decisions based on AI-based quantitative models. The fund has been able to maintain better performance than its human counterparts, with an average annual return of approximately 66% over the last two decades (S., 2024). Another example is Bridgewater Associates, which uses AI to critically analyze global economic data to assist in trading (Jensen & Moriarty, 2024).

The real edge that AI has over humans is the fact that it can analyze both structured and unstructured data on the stock market, including in the news, social media, and financial reports, enabling more accurate predictions for future market movements (Gnibus, 2024). Furthermore, these insight-driven trades are executed much faster than any human trader. This saves a great amount of time, and as they say in the world of trading, time is money. With great speed and precision, AI-driven trading has now taken Wall Street by storm, with 60% of all trades being executed by algorithms (Schlaepfer, 2024).

However, AI isn’t foolproof per se. The most critical issue it faces is the dependence on historical data. Although this may sound normal, the over-reliance on past information does not help in the case of unprecedented events like the 2008 financial crisis or the COVID-19 pandemic, leading to poor decisions in such scenarios. Programs trained on old data can often fail to adapt to the abrupt market swings and eventually cause large losses. Additionally, AI-controlled strategies contribute to volatility, as seen in the case of a “flash crash”, when automated high-frequency trades create huge, unexpected market fluctuations, and can lead to major setbacks for many players (Needhi, 2024).

I personally feel that although AI has tremendous potential, it still needs to be used with and not instead of human traders. Rather than replacing human counterparts, AI should be used to enhance stock market decisions ultimately made by humans. What are your thoughts on this – will AI become the norm when making trading decisions and eventually dominate the stock market?

References

Gnibus, A. (2024, April 15). Why Unstructured Data is Having a Moment, Thanks to AI. Medium. https://medium.com/@alexgnibus/why-unstructured-data-is-having-a-moment-thanks-to-ai-ad546e5a7158

Jensen, G., & Moriarty, J. (2024, May 30). Are we on the brink of an AI investment arms race? Bridgewater. https://www.bridgewater.com/research-and-insights/are-we-on-the-brink-of-an-ai-investment-arms-race

Needhi, J. (2024, July 9). AI’s Role in the 2024 Stock Market Flash Crash: A Case Study. Medium. https://medium.com/@jeyadev_needhi/ais-role-in-the-2024-stock-market-flash-crash-a-case-study-55d70289ad50

S, R. (2024, September 27). How Jim Simons built the best AI that solved the market. Medium. https://suenrankine.medium.com/how-jim-simons-built-the-best-ai-that-solved-the-market-f3e23319db29

Schlaepfer, D. (2024, July 2). Humans vs. AI in The Stock Market: The Worst Trade Ever Made? Forbes. https://www.forbes.com/sites/forbesbooksauthors/2024/06/18/humans-vs-ai-in-the-stock-market-the-worst-trade-ever-made/

No ratings yet.

Over the past two months, I’ve been trying to learn React, a powerful JavaScript library for building user interfaces. As a newcomer to React, and coding in general (except for that one python course in every business bachelor), I quickly realized that learning React goes a lot deeper than I initially thought. With this comes the endless array of questions that arise with every new function you’re trying to write.

In the beginning, like many others, I found myself searching for solutions on forums, reading documentation, and piecing together tutorials. While these are invaluable resources, they often left me with information I found wasn’t exactly tailored to my specific questions or project setup. I would look on StackOverflow or Reddit and pray that some person from a random corner of the world had had my exact problem once before. But why search for this perfect scenario when you have a personal coding assistant that has the knowledge of all these forum users combined?

Instead of sifting through pages of StackOverflow posts or scrolling through Reddit threads, I started using ChatGPT to ask hyper personalized questions. For example, I could describe my project’s context in detail: what kind of components I’m working with, what the general goal is and just paste my entire codebase. The answers I received were tailored precisely to my situation, bypassing the need for trial and error from someone else’s scenario.

This however isn’t all rainbows and sunshine. Besides the occasional mistakes (which is to be expected), I started becoming very reliant on genAI to just create my code. Why would I try to type out and understand a function, if ChatGPT could do it perfectly and 10 times as fast. This however, does not really fit the description “Over the past two months, I’ve been trying to learn React”. I have thus been trying to limit my use of ChatGPT (in this context) to just asking questions about code I have actually written myself. This way I try to understand the code I am writing first, and if all else fails resort to my ‘coding assistant’. I feel like this is a much healthier way of “learning how to code”. I think it is really easy to make yourself believe you’re good at writing code, if you’re working with a chatbot on your side.

What do you think, what is your personal experience with ChatGPT and coding? Do you have any recommendations for LLMs to use for this besides ChatGPT? I was also thinking, will these forums such as StackOverflow disappear, why would I try to find a similar problem when I can just get the perfect answer through a chatbot? Let me know what you think!

No ratings yet.

Generative AI has already given us an abundance of opportunities which were not deemed possible only 3 years ago. From having a crappy version of ChatGPT (which was already groundbreaking at the time) to generated videos which can hardly be recognized as AI. The technologies are being improved regularly, leading to better versions of LLMs (such as ChatGPT and Google’s Gemini) which ultimately makes many tasks easier and faster to perform for humans.

If I relate it to my personal situation regarding using these text generating forms of AI, I can say it has made studying easier. From simply using it to improve my writing in formal language and grammar, to asking for solutions on specific assignment-related questions. Especially the latter has made tasks easier to do as the GenAI can provide with a starting point to use as a thought. If I for example ask Gemini to give me three examples of firms that use data monetization and how these firms do so, almost instantly I receive an answer which I can use to start in my answer. The time it saves me doing my own research on these simple questions might look marginal, but once added up can really make a difference. Therefore, one can argue that these forms of GenAI increase a student’s productivity quite significantly.

However, a potential downside I have noticed in this process, is the passive learning I experience. With using AI, I hand the thinking process to Gemini, after which I only have to read and reason which response best suits the question I need to answer. If an answer is not to my liking, I can simply ask the LLM to give a new answer, or slightly change my prompt in the hope it changes the answer accordingly. My own learning might become more passive than actively having to think on my own answers. This can be considered as becoming lazy in my opinion. During the guest lecture, Stefan van Duin also mentioned that one of his biggest concerns regarding GenAI in the future, is that it can make us lazy human beings. In the context of this thought, I feel the process of making us lazy has definitely already began.

In conclusion, I argue that the LLMs have made us more productive and gave us opportunities of enriching ourselves much easier and faster than we might have ever could have imagined. However, we should be careful in the usage of AI and our own development. We should not become lazy in our own thinking process and fully utilize LLMs for our tasks. Perhaps universities should focus on making assignments that are not possible to do with these LLMs. All in all, It is a double-edged sword for which we need to be careful in the increasingly better LLMs that we will see in the, foreseeable, future.

(I couldn’t come up with a saying for the last sentence. So, I intuitively asked Gemini, who quickly came up with ‘double-edged sword’. Just an example on how embedded it might already be in my system!)

No ratings yet.

In recent months, I have had the opportunity to integrate generative AI into our marketing and SEO strategies, and the results have been great. These tools, ranging from text-to-text generation to market intelligence, have transformed how we approach content creation and optimization.

Generative AI has been particularly effective when competing for various search queries. In practice, it has helped streamline the process of determining blog topics, ensuring we target the most relevant and competitive keywords. With generative AI, we do not just pick topics but we select them based on their potential to rank well. By analyzing search queries, we could create SEO-ready blog posts, increasing our chances of appearing in top search results.

Moreover, generative AI has proven invaluable in checking the competition. Before crafting a piece, we could analyze how competitors structured their content, what keywords they targeted, and how their strategies performed. This allowed us to optimize our own content to outperform theirs.

On the marketing front, AI tools allowed us to inspect the competition with a level of efficiency we had not experienced before. We were able to create SEO-ready advertisements by understanding competitor strategies and adjusting our approach accordingly. This has streamlined our advertising efforts, making them more targeted and impactful.

Overall, my experience with generative AI has highlighted how powerful these tools can be in marketing and SEO. They not only save time but also provide strategic insights that give us a competitive edge. Looking ahead, I am excited to see how these tools continue to evolve and improve our processes. Additionally, I believe there is room for further improvement in personalization. While AI is efficient, adding more nuanced, human-like creativity in content creation could enhance engagement. Integrating deeper data analytics for real-time adjustments would also be a valuable feature to keep pace with changing market trends.

No ratings yet.

Not so long ago creating your own website required a deep understanding of coding and it took a good amount of time. However, today we are able to create websites in no time and with no  knowledge of coding, thanks to AI-driven web tools. Having used these tools myself, I’ve experienced firsthand how easy they make website development to be. But are these tools the holy grail they seem to be?

What makes AI-driven web tools great?

Some characteristics of AI-driven web tools that make them so appealing are their ease of use, easy customization, scalability, and time saving efficiency (Haim, 2024). Platforms like Wix ADI (Artificial Design Intelligence) have revolutionized web development through these characteristics. From my own experience with Wix ADI, I found its ease of use to be remarkable. It offers a simplified setup where answering questions about your businesses or project will result in multiple layout options that meet your needs (Goldstein, 2022). The customization could not be easier through its intuitive editing where you can easily adjust colors, fonts, images without any coding knowledge. Further, the scalability the AI tool offers is amazing, because you can use the built-in features for marketing and social media promotion to grow your website with it (Goldstein, 2022). Lastly, how long does it take to achieve these results? Just a few minutes to get a fully functioning and customized website. The AI can handle everything, even color schemes and the layout of the design. 

(Wix ADI, n.d.)

Are AI-driven web tools the future? 

The appeal is clear: no coding required, faster development time, and professional results. However, the convenience of these AI tools comes with limitations. From my experience, complex projects cannot rely on these tools, at least not entirely. Even so, for small businesses and startups, these tools are life changing. Not having to invest large amounts of money and lots of time in building a website is revolutionary.

What AI-driven web tools should you use?

There are numerous AI-driven web tools available, but choosing the right one depends on several factors. Based on my experience, your budget is crucial. I’ve used tools that offer free versions, but the premium features are where the real customization and scaling power comes into play (Haim, 2024). Additionally, the type of website you’re building matters. For instance, if you’re building an e-commerce store or restaurant booking site, you’ll need a tool that specializes in those features (Haim, 2024). Skill level is also key. While some AI tools, like Wix ADI, handle most of the work for you, others require more input, so your comfort with web design should guide your choice.

Conclusion

Overall, my experience with AI-driven web tools has shown that they empower anyone, regardless of coding knowledge, to bring a website to life. While there’s still room for evolution in these tools, they have already revolutionized web development for small businesses and personal projects. As the technology evolves, more of these tools will evolve.

References

OpenAI. (2024). ChatGPT (Oct 3 version) [Large language model]. A futuristic panoramic illustration of AI-driven web development. https://chat.openai.com/chat

Goldstein, K. (2022, June 19). Wix ADI: How Design AI Elevates Website Creation for Everyone. Wix Blog. https://www.wix.com/blog/wix-artificial-design-intelligence

Haim, I. (2024, May 7). 6 Best AI Website Builders of 2024. Elementor. https://elementor.com/blog/best-ai-website-builders/?utm_source=google&utm_medium=cpc&utm_campaign=21398986558&utm_term=&lang=&gad_source=1&gclid=Cj0KCQjwu-63BhC9ARIsAMMTLXRSbna5db2d_T_x82zTHgGjcjcuu-zHnKmuHLZS5zt8mi83PInrhzEaAozTEALw_wcB

Wix ADI. (n.d.). Ai Website Builder. Wix.com. Retrieved October 2, 2024, from https://www.wix.com/ai-website-builder?utm_campaign=toplink-ai-website