#114 The AI Series (3) - Practical vs. philosophical

The one thing that has the potential to fundamentally change the area of EU R&D funding for good is the development of Artificial Intelligence. ChatGPT broke through a year and a half ago and the development of this technology and available services – free as well as paid – are going at the speed of light.

In The Grant editor’s office I am of course following this closely. So, I have decided to make an AI series where I dig into this in more detail. And of course I do this with some of the sharpest professionals on the topic: independent consultant and AI-Chat-guru Kristjan Zemljic and Senior Consultant and very pro-active AI-Chat-user Rita Gil Mata.

In this third episode I discuss and reflect with my two distinguished guests on the more philosophical and ethical aspects versus the practical use of the new tools. An important part of this whole process that we are all going through with these tools developing at such a quick pace. A wonderful conversation.

Time codes:

00:01:47 Introduction

00:04:16 How to percieve the tools and what they bring

00:24:42 How to establish limits 00:32:17 Critical partners and clients towards LLMs

00:45:17 Ethical considerations

00:46:52 Reflections on the position of the EU Commission

00:54:32 Recommendations on what to be aware of

Further help and links

A few prompting tips from the guru

Kristjan says:

Some of power words:

  • User should write: "brainstorm, outline, then write..." rather than just "write" (breaking down steps, gets better results"

  • Use frameworks in prompts. Instead of "generate idea on...", prompt "generate idea on...using first principles thinking." This approach breaks down problems into fundamental building blocks, results in more innovative solutions.

  • Alternatives to it is "generate idea on...", prompt "generate idea on...using system thinking principles", where focus is more on the big picture, and how different parts of the system interact with eachother to achieve the goal (interesting for impact)

Take a deep breath, relax, and let's think this through step by step until you get it right for this prompt. This is a combination of several techniques:

  1. "Let's" assumes cooperation, yields better results than saying "do."
  2. "Step-by-step" initiates deeper reasoning, useful for complex thinking, strategies,
  3. "I trust you..." emotional prompts, encourages LLM, and results in better outputs.

How to acces machines through VPN

For the activation of VPN, you need to:

  1. Download the Opera browser, then:
  2. On the left panel of the Opera setting's page, you will see “Privacy & Security. Click on that to access and activate the Opera browser VPN.
  3. Scroll down to “VPN” section. ...
  4. Once it has been activated you will see a blue VPN badge on the left side of the address bar.
  5. That's it.

ChatGPT

Claude

Connect with Rita Gil Mata on LinkedIn

Connect with Kristjan Zemljic in LinkedIn

Mistral

Bard

Link to The AI Series