No ratings yet.

Have you ever had the feeling that the AI models you are talking to such as, ChatGPT, Gemini or Deepseek sometimes are a bit too agreeable? Not only are they agreeable, but they also rarely question your logic and often accept your assumptions. This raises the question, how helpful is GenAI if it always agrees with you? Are they helping us in structuring our thoughts and being critical or do the just say what we want to hear?

I have used GenAI in a variety of ways, by asking quick questions about how to repair something in my house, by teaching me specific things I am learning for school or by advising me on business idea’s. For this last use case, I quickly noticed that I rarely received critical feedback. Instead, I always got answers that reinforced my assumptions. This experience is in line with what we discussed in class about credible analytics (Vidgen et al., 2017). Here it was shared that data driven insights are only useful if they are; accurate, critical and transparent. Just as that incorrect analytics can result in bad business decisions, AI tools that are overly agreeable can reinforce bias and also result in bad business decisions.

This insight can also be linked to the EU’s AI act on transparency and human oversight. AI tools that are designed to please use could unintentionally amplify misinformation. So how can we make AI more critical? One idea is to design a mode that switches the AI from a yes man into the devils advocate. Another idea would be to recommend, the user to include an additional section in their prompt that the AI could use a guide for that convo.

Would you prefer an AI that is overly agreeable or one that challenges your ideas?

No ratings yet.

Network effects are common in platform businesses. However, are there network effects in crypto currency and what do they look like?

When talking about currency in general, the following is true: the ones using the currency give it value. If the whole world decides to stop using the dollar as a form of currency, the value will drop to almost 0. Therefore, more investors / speculators in a crypto currency increase the value of that cryptocurrency, which in turn results in a people wanting to use / buy that cryptocurrency. This network effect can be classified as a direct demand side network effect.

There are also indirect network effects in cryptocurrency. An increase in the usage of this cryptocurrency will also create new businesses and service providers. An example could be a consultancy company that helps companies in accepting payments in that cryptocurrency or a business that develops secure crypto wallets. These indirect network effects are however quite common in most industries.

For cryptocurrency there are 3 main feedback loops.

1. Price-Adoption feedback loop (Direct network effect)

As already mentioned, a higher price results in more demand which again results in a higher price. However, this feedback loop isn’t always positive. If price fall the opposite could take place as well: price drops results in less interest which in turn makes the price fall further.

2. Liquidity-Utility feedback loop

More users and traders increase liquidity results in lower transaction costs which attracts more users and traders.

3. Developer-ecosystem feedback loop (Indirect effect)

More developers result in more apps and services that are built on that cryptocurrency, which in turn increases the demand to that underlying cryptocurrency.

Most of these feedback loops are present for all sort of currencies. Out of these three, the third is most relevant for cryptocurrency