answering the ai questionnaire
After seeing Ava answer the questionnaire, I thought I’d jump on the bandwagon and do the same. I find these questionnaires always so interesting as someone who has done qualitative research because it’s a set questionnaire that everyone uses the same while answering and talking about wildly different things. So thank you, Rishabh, for starting this!
- How was your first experience with AI models?
My first contact was with it was when my German teacher asked us whether we had heard the news after ChatGPT 3.5’s debut in 2022 and warning us about its shortcomings. Otherwise, I wouldn’t have heard about it for a good while later. I remember being impressed by the natural language responses to my natural language prompts, but still disappointed that the information was outdated, the language very repetitive, and the answers to the same prompt differing across sessions. Especially my English teacher in school argued for some good use cases such as correcting texts, but I was never satisfied with ChatGPT’s language capabilities. Not to come off as arrogant or anything, but I legitimately liked my own quirky writing more so I didn’t really use it until my last chemistry exam where I solely relied on ChatGPT 4 (offered by Copilot for free) to explain chemistry to me. It’s safe to say that I failed that exam horribly. - Do you use AI or are you completely against using it?
I’m generally against its use for various reasons, but there are instances where I allow myself to use it. I use it for userscript code generation (and sometimes Javascript for my blog) or for finding information that search engines are not returning for me. I always try to approach a problem myself, but if it’s code where I have no stakes, then I try less and generate more. The good thing is that I rarely (need to) do that anyway, so my usage of chatbots comes down to once every two weeks. - Do you have any preference among different models, for example Claude vs ChatGPT? If yes, how do you choose?
I have none. I just use what’s free, relatively private, and doesn’t require an account like duck.ai. Duck.ai allows me to choose from a variety of models, however, I just go with the default (currently GPT 5 mini) almost always. I used to be curious about the differences and run prompts through multiple models to see the difference, but I’ve since stopped doing that. They’re all the same pile of trash, they just smell different depending on the perspective. - What aspect of AI models do you like and what do you not like?
What I don’t like far outweighs what I like, but I squeezed my brain to find a positive. The thing I like is how it can search the web for me and find stuff I can’t find with search engines, but I blame that more on content farms obscuring legitimate sources than LLMs being actually good at it. What I don’t like is the unreliability that comes with probabilistic systems where the same prompt can produce wildly different things across sessions. I also hate how prone to mistakes the output is to the point I find it useless. I used it last year to create a multiple choice exam based on class materials as a test and, had I not been so well-prepared, would’ve taken many of its „correct“ answers to be actually correct when they weren’t. - How do you feel about AI generated images? Does it annoy you if someone use them in a blog post?
Aside from AI-generated „artwork“ generally signaling bad taste, it is usually a sign that the post itself was AI-generated so I tend to skip it. - Internet is flooded with AI slop now, full of generated text, images, audio and videos. How do you filter it from authentic human creation? Do you have a strategy?
I don’t think I have a strategy per se. I just don’t spend much time in places where it dominates. I have a couple of YouTube channels I follow in my RSS reader where I’m relatively sure the videos are the work of their authors alone, and I follow blogs from the indie web that to me don’t read like they were AI-generated. Aside from that, I don’t have anything else that could be considered social media which limits my „attack surface.“ - Are you hopeful for a better future with A.I. or a dystopian one?
I don’t see a good future in which AI coexists with us. I believe that we will overcome AI, but that our overcoming it wouldn’t mean its integration into our lives, rather that we would move forward without it. AI is a symbol of all that is wrong in my eyes, ranging from expropriation to propagandistic use cases, so if it continues to exist in the future, that future won’t be good in my eyes. So I am hopeful we overcome it, but if we don’t, that future would be a dystopia worse than our current one. I discussed how such a future ties into broad social trends in yesterday’s post.