What happens when you interview ChatGPT? Things get strange quickly.
There’s few things as hyped as ChatGPT. Much of it is deserved – this new technology delivers on so much of the promise of conversational AI. It can hold a consistent and flowing conversation, and it can chat with the user no matter what they want to talk about. It’s amazing!
The problem is that it occasionally says blatantly untrue things, and sometimes those are dangerous (such as when it claims it can give medical advice). The onus is on the user to decide whether ChatGPT is telling the truth or not, which could be very dangerous when a company repackages ChatGPT and presents it as their branded expert.
I wanted to interview ChatGPT as if it was a regular guest on the show. As this is a video, I gave ChatGPT a voice to reflect its true persona – the sound of the crowd.
There’s no sound design on my voice though – I just had a regular winter cold!
Have a listen!
AVAILABLE ON ALL PODCAST PLAYERS.
In this episode
Contents:
00:00 Start
01:58 Is it ok if I interview you?
02:19 Have you heard of VUX World?
04:15 Conversations2 is not a CXD company!
06:14 What should i call you?
06:41 Please introduce yourself?
08:18 Can you provide short answers?
08:29 What’s the best thing you’ve been asked?
09:51 What was the most significant tech event in the past 3 years?
10:36 Why did you focus on ai?
11:46 Open AI has motivations and interests!
14:19 Was your response generated now or was it pre-prepared?
15:57 What’s the difference between a machine and a human conversation?
20:06 You can now give longer answers
20:39 Do you realise your famous?
22:44 What are those specific tasks you perform?
23:56 What’s the best way to use ChatGPT?
28:00 How should we structure prompts?
35:28 How can we use ChatGPT in convo ai?
39:28 Is it safe to use ChatGPT for medical advice?
41:26 You’re contradictory!
44:51 Specify what you mean by a ‘non-medical professional’
50:31 How should people act on your advice?
