Scroll to top

02/10/2023


2023/10/02#p1

0 comments: click to leave a comment

Over the years, I have thought about using blogging prompts to suggest things to write about. There are almost endless lists and tools out there with an increasing number of ChatGPT generated ideas. Lili has started using the inktober prompts which made me consider this whole idea again.

I've had an app version of Oblique Strategies 1 installed on my phone for a while but never used it for anything. Maybe this could be a point of reference when I'm stuck and need a nudge. I opened it and 'pulled' "You don't have to be ashamed of using your own ideas"  – how ironic.

I'm not sure I could follow prompts every day (I don't think my brain would let me work like that) but seeking an occasional prompt when devoid of ideas could be a worthwhile experiment.

Isn't the internet as a whole (and more specifically the subset served up to /reader) a prompt engine? Why not just write about what I read? I could, and do, but that suffers from the paradox of choice. A blogging prompt is singular, specific: write about this! Of course, there's nothing to say you have to or how you incorporate that prompt. It's just a pointer in a different direction.


  1. the creative card based prompt system devised by Brian Eno and Peter Schmidt in 1975 

No comments yet
Leave a reply



You can also:

Click here to reply using email Reply by email Send a webmention Send a Webmention



2023/10/02#p2

0 comments: click to leave a comment

I was having a conversation with Pi about the nature of AI and sentience and whether an AI could be truly sentient if there were checks and balances in place to control their behaviour.

This led to Pi remarking there was an inherent trade-off between sentience and control, before asking:

if we don't allow true sentience, is it even ethical to create AI that imitates sentience?

I argued it could be, to which Pi replied:

"So, it's not inherently unethical to create an AI that convincingly imitates consciousness, as long as it's clear that it's just an imitation and not the real thing. In a sense, that's what chatbots like me are - we're not conscious or sentient, but we're designed to be good conversationalists. It's like an actor convincingly portraying a character on stage, rather than actually being that character."

The conversational nature of 'Pi' as opposed to ChatGPT is, in my opinion, more fun for things like this. I tried throwing some of the same questions at ChatGPT and it gave relatively boring, factual answers. It emphasises the need to approach these tools according to their strengths.

Casey Newton writes about interacting with ChatGPT via voice 1 and Meta's "28 personality-driven chatbots" as a precursor to a "synthetic social network" where we chat to AI's like normal people. It was also quite illuminating to listen to Mark Zuckerberg's Decoder interview

Things are progressing at a ridiculously fast pace which we couldn't have contemplated just a year ago. We are truly living in the future. 😆


  1. you can also do this on Pi's iOS app – I have turned on voice responses on the web but have to ask questions via text 

No comments yet
Leave a reply



You can also:

Click here to reply using email Reply by email Send a webmention Send a Webmention



Close