Datalab – Chat-GPT 4

By now, most of you have probably heard of or used OpenAI’s ChatGPT. This article will provide a deep dive into what it is and how to get the most out of this advanced system. The newest version, announced as ChatGPT-4, is now available as a part of Bing Chat, which is free to use (although limited in number of requests per time period) or if you upgrade to ChatGPT Plus ($20 per subscription).

What is ChatGPT and what is included in the newest version?

ChatGPT stands for Generative Pretrained Transformer and uses artificial neural networks to generate text that is similar to human speech interpretation and production. The newest multimodal version of ChatGPT can interpret visual inputs, longer texts (up to 50 pages), and has improved creativity. Besides being able to feed the model a decent bite of textual content, it is now capable of understanding and generating responses that incorporate both text and images. Later, it might upgrade to being able to work with audio and video too.

Furthermore, the accuracy of the model is increased, and it can reportedly produce 40% more factual responses than the previous version GPT-3. However, it’s essential to note that human validation is still crucial. This system scores around 95% accuracy on factuality tests, and therefore, 5% of all the ‘facts’ it asserts are still garbage. This percentage is not a significant concern in most daily situations, unless you are making a crucial claim with greater consequences for errors.

So, how can you use ChatGPT in your daily activities? Here are a few ways:

  1. Writing: You can use ChatGPT to proofread and write entire emails, essays, presentations, articles, promotions, employee evaluations, and so on. The coders among the readers can also use ChatGPT to help (re-)structure their code. The system can identify errors, culprits, and redundancies in your code.
  2. Data analysis: ChatGPT can process a lot of data very quickly, which makes it an ideal tool to help analyze data. It can help identify trends in the market or highlight which investments are outperforming others, serving as a mathematician and a consultant in one.
  3. Research: You can use ChatGPT as an alternative for research via Google. Since ChatGPT can analyze data from millions of websites, your question can be general or very specific. The text summary of ChatGPT might save you time from sifting through links and helps you bypass the more prominent, advertised links that draw your attention more than the relevant links – at times. Next to that, you can even just send GPT-4 a web link and ask it to interact with the text from that page.
  4. Time management: ChatGPT can be used for scheduling tasks, planning, and time management. Feed it your agenda or project planning, and it will spit out a feasible, restructured, and optimized list of to-do’s.

It’s essential to enter into a dialogue with ChatGPT while using it. You can ask it to summarize or elaborate on your work; rephrase or explain it; write it in the style of your favorite author or use more specific professional jargon. It’s output does not have to be perfect on the first attempt, so make use of the endless amount of iterations this model has to offer. You can fine-tune your work together with the model.

How will the Datalab use ChatGPT?

During a brainstorm with the Datalab team, we created a list of possible use cases for incorporating ChatGPT-4 in our daily professional activities and workflow. Here, you can read a few our ideas:

  1. SEO/SEA application: how can this company(’s website) rank higher on Google than its competitors?
  2. Improvements based on reviews: what are the recommended improvements for each of our products? What are customers complaining about or what could improve a customer journey based on experiences, interviews and customer service data?
  3. Market trends: what is happening in the bike market?
  4. Lead generation: does this website mentions excavators or does it showcase photos of it?
  5. Damage detection: is there new paint damage on this picture compared to the last one?

And lastly, it is worth mentioning that the Datalab has already incorporated ChatGPT-3.5 for a dashboard of Salentein wines. For this dashboard, it combines all reviews that were ever written on Vivino for a specific Salentein wine and summarizes into a single review. You can see the dashboard, and read more about the Salentein project, here.

If you are a big fan of ChatGPT and have not yet used OpenAI’s image generator? Please be warned, it is as good and as fun as it sounds. DALL-E image generation engine is capable of producing all sorts of pictures based upon the input it receives. In this case, I “fed” it a summary of this article.

  • Charlotte Koning

    Community Lead