chatGPT/LLM - Design Prompt Engineering

Consider your ten minutes spent on the summation of my hours of exploration a fair trade.

As a critique of chatGPT fair enough.

If you have a suggestion on how to tackle communicating and listening to feedback on a “general discussion” approach to structured queries for an advanced predictive text engine, fire away. chatGPT & Co, seems like a force that the design industry will have to face, employ, or challenge. If I dismiss something, good to have a reasoned basis, personally am nearer to an assessment than when started.

I intended to structure the title and the setup as a head’s up, and advisement for people that have already formed an opinion. There should be a place for long-form discussions. Attempting to shuttle people offsite to a personal blog or website cuts engagement by a big factor. Limiting observations to 240 characters or less just doesn’t seem to fit this task.

As stated at the beginning, there are common forum reading tools that can be activated to allow both an overview and a way to dive deeper into information.

Thanks for your time.

1 Like

This seems relevant to your ongoing research: Power and Weirdness: How to Use Bing AI - by Ethan Mollick

Might Bing AI give you actual research papers, because it is connected to the web?

1 Like

Will take a look at the guidelines for how to maximize Bing. Have been taking Bing AI for a spin. The results are much more along the lines of a tight search engine. Which is a good format. It does not provide information where there might not be information.

I was hoping that chatGPT could do some synthesis and provide some confirmation to a material theory I have developed. False sense of confirmation though.

A similar interaction with Bing netted mentions of differences between rebound between EPP and EPS, but nothing it could find to elaborate upon, but some good links to follow and I learned a lot more. The papers and links it provides are good extensions, but not confirmations. All in all more positive experience for the goal of learning.

1 Like

The “appearance of” aspect is still amazing (to me). Rather than merely (!) providing accurate, penetrating facts, the responses fill cognitive space with reasonable and authoritative sounds. It’s passed a litmus test of being convincing, without needing specifics.

I used to dread design crits with live people who behaved like that, dubbing them ‘content-free zones’.


Your investigation is so fascinating. Thank you so much for coming up with the idea, spending the time on it, and giving us a viewing window. It’s wicked illustrative. Your methodology is damn fine, and your your editorial voice is both very effective for your research purposes, AND hilarious. Pontificating about how AI is sheit (or the New Hotness) really means nothing; what you’re doing is real. White paper? NYT guest article? TED talk? It needs a wider audience.


If you overlay Aristotle’s Ethos, Pathos and Logos to what ChatGPT is doing it is all Ethos for now. AI is not capable of Pathos or Logos yet. It draws conclusions based on trillions of published word pairs without specific emotional or rationally designed outcomes from the human audience who are reading its outputs. Effectively it cannot calculate or explain risk or predict preference at all in its effort to explain anything to an audience…yet. So, some of us still have jobs…for now. Banksy is still better at organizing information (Ethos, Pathos and Logos) to arouse a specific designed human response than ChatGPT.

We are just beginning to understand how this technology has an effect on the mentally ill, unaware and under educated with the example of the Belgian man who committed suicide a few weeks ago.

Very interesting lens on this, thanks.


So chat AI is just like an influencer on TikTok?

Pretends to know what they are talking about but just spewing nonsense?

I fear that the biggest threat to AI taking our jobs are all the cringe posting AI experts on Linkedin posting tutorials and guides on how to use AI to compile Linkedin posts. By the time AI truly becomes that threat everyone will already be so sick of AI and we will happily move to a self sufficient farm. Im halfway there already.

Joking aside, I’ve used chatgpt to quickly find stats and existing guidelines - hours of googling and reading cut down to minuets.
Ive also used dall-e to create visuals for storyboards/personas that also saved many hours. A budget is a budget, and I rather spend those hours on design. The client seemed to agree.



1 Like

I welcome the future when it’s AI fake influencers bots making nonsense content sourced from AI generated nonsense content with AI followers. The circle will be complete.

YouTube will then self-implode, influencers will have to get a real job, and the world will be a better place. :slight_smile:


This is making the rounds on Li, no idea who originated.

1 Like