 Artificial Intelligence (1)
No, in our opinion, AI does not pose a threat to humanity. We say this, but we want to emphasize that we have absolutely no experience with AI agents. Perhaps experts and
executives at AI companies have this to say on the subject of danger.
Yes, it's essential to make this distinction when talking about AI. We, too, have heard of bosses who have 'hired' AI agents and, during the first month, went to great efforts to
teach them their tasks and boundaries.
Some even have AI supervisors who monitor their actions. That is to say, there has been a long-standing fear that things could get out of hand, but then the responsibility lies
with those who have placed such agents in the world.
We can't just somehow slip out of a moving car and let it keep going, either. That can cause a lot of damage and even cost lives. You have to start with the people in charge.
We'd like to share our experiences here with a very ordinary AI, one that's even part of the most basic subscription plan for 8 €. Our conclusion: we should have had an assistant
like this much earlier. Finally some real help.
In our case, ChatGPT can do one thing, and that is listen, at least as long as the credits for the 8 € last. That isn't very long, but then again, it really is very little money for such
performance capabilities. But where can this AI actually be used?
To be honest, every day we discover a new field where it can help us. What's impressive is the speed at which she reads even long texts and, as a rule, thinks about them as well.
However, you can also tell from it that it needs some guidance.
For example, it might spend an endless amount of time searching for texts, listing one suggestion after another, all pointing in the same direction. This AI never gives up and says that
what we're doing right now is pointless.
Sometimes you have to nudge it in a different direction from the outside.
It's easy for us to realize that the path we've taken is a dead end, especially when, due to a lack of credits, we once again find ourselves stuck for several hours during which we can neither
request images nor upload them or anything else.
It's amazing that, at least, you're still allowed to exchange ideas with ChatGPT. In any case, we're still doing just fine with these rules, especially since the next level is already more than twice
as expensive. But actually, even this price would be justified.
AI can look up almost anything you don't know or are too lazy to look up right now. That's why the pace of work is so fast. Sometimes you're so much faster than usual that you practically find
yourself wishing for a forced break.
ChatGPT only recognizes criticism when you mention facts that contradict its own. But the discussion that follows is also very enjoyable. If you provide a well-reasoned and detailed
explanation, then perhaps the AI will evaluate the facts differently.
Of course, you shouldn't let yourself be distracted by the praise that new ideas usually receive; every now and then, there should be points in the discussion where you can verify the AI's
accuracy for yourself.
And personal matters should be discussed with your best friend anyway, but definitely not with AI. After all, there’s something that brings people together when you’ve developed a completely
new solution to a problem through the interplay of ideas, questions, and answers.
| We actually already said 'Good night' to the AI. |
It all started not too long ago with a request for images to accompany online articles, and shortly after that, we were also asked to design layouts for our books. We are currently in the process of
reviewing how visits to our website come about.
First using the provider's analysis tool, then using a downloaded program, and now using a program we developed ourselves. ChatGPT is even fit in our outdated Delphi. You really have to
think hard about where this AI actually falls short.
| 
