If you’ve never been there you may have little creativity outside of your domain… Maybe that is a skill that is as yet untapped?
How much are AI hallucinations generated by humans?
AI is a flawed mirror ie it has model ‘deficits’ and also prompting human ‘deficits’…
These create feedback loops that can really confuse and destabilize normal people and coders alike…
While we have a tendancy to blame ‘AI’ for these issues, it’s probably worth remembering that maybe a good proportion of the problems are US asking the wrong questions…
Indeed, I have corrected a few prompts for some of the most experienced coders on this forum (naming no names)…
Best not to blame AI or the users for the learning process I think!
Or to assume that a product that ‘thinks’ has no impact on human thought beyond technical coding…
I suspect it’s more a product of choice anxiety in light of possibility space opening up the possibility for multiple possible replies for certain topics/questions, rather than a direct product of human misunderstanding… though I also supect it’s a bit of both… and I’m no expert either
1 Like
In GPT-4.5, I’ve been focusing on aligning interests and goals. During the hundreds of hours I’ve spent with GPT-4.5 (Unlimited on Pro Plan), I’ve noticed that by dedicating more time to aligning interests and goals and managing personalized memory, the interactions feel more in line with what to expect in GPT-5. GPT-5 is anticipated to combine models with enhanced reasoning and thought, along with expanded memory from previous chat sessions. I found that GPT-4.5 responded better when this setup was optimized.
1 Like