Are you interested in knowing more? Keep reading.
We are still in the gravy train part of the rug pull scam. AI is
heavily subsidised by billionaires and massive corporates, who have
been sold on a false promise of replacing you, the person, replacing
you, the worker, with a perfect machine slave that never rests, never
complains, and does everything it’s told.
But what will happen once the investors start demanding returns, the
free money runs out, the true cost of running AI models comes forward,
and those that have been on the hype train realize the true
limitations of AI? Those who have become dependent on it will be
forced to start paying big. And the already obscene fortunes of AI
Robber Barons will grow even further.
From self-prompting to multiple agents talking to one another until
they both agree, countless resources have been spent on trying to
correct AI hallucinations. However, hallucinations are inherent to the
way AI, or Large Language Models more specifically, work, and they
will never be fixed.
Large Language Models are nothing but a fancy autocorrect. Where they
differ from your phone’s keyboard is the size of their memory, which
they can use to predict the possible next word, based on a huge
database of previous words. But that’s all it is, a prediction.
The AI doesn’t know or understand what it’s writing. Give the same AI
the same prompt, and watch it give you two completely different
answers. It can guess, but it will never know. Fixing this would
require a complete rebuilding of the way the Large Language Model
operates.
“Funny how when I ask AI about something I know, it’s completely
wrong, But when I ask it about something I don’t know, it gives a
great answer.”
This is a paraphrasing of a post I saw on Reddit a couple of weeks
ago. In addition to what was written above, corporate executives love
AI because it is extremely good at one thing – making you think it
knows what it’s talking about, and masterfully avoiding any
accountability.
Just last week, just for fun and to keep up with the newest
developments, I tried asking one of the leading AIs, Claude
specifically, to add a simple feature to a search button. The
requirements were simple: automatically focus the field when it
appears so the user can start searching right away.
At first, it gave me something that seemed like exactly what I wanted,
and it even seemed to work properly… at first.
When I inspected the code more closely, I saw a massive chunk of code
that did seemingly nothing – variables being reassigned to themselves,
delegates and coordinators being created and never used, pointless
two-way bindings that referenced one another convenience initializers
that initialized one variable out of ten, then silently failed.
Just for the fun of it, I removed all the pointless code that Claude
conjured up, and asked: “Even without this code, the feature works as
it should. What is the point of all the useless code?”