JavaScript is disabled
Our website requires JavaScript to function properly. For a better experience, please enable JavaScript in your browser settings before proceeding.
I conduct work in the traffic safety field, so autonomous vehicles have been of special interest to me, with AI being a natural progression. What I find most disturbing about the current state of AI is that it can make itself sound like it is providing good, accurate information, but in fact, it may be flat wrong. Recent example: I asked how many hours a day lions sleep. The AI answer was about 21 hours (seems correct), but then added "which is about 56.3% of their day. This means lions have days that are 37.3 hours long. Who knew.

I won't be asking any AI how many grains of Blue Dot I should use in a certain load anytime soon.
 
I won't be asking any AI how many grains of Blue Dot I should use in a certain load anytime soon.
Embiggining clickage is necessary unless you have 6000/236568 vision.

1738337150331.png
 
The problem is that it can give the next answer just like this that says, "97 grains" with the same authority. No editor or proofreader like a reloading manual. I read another article that turned out to be AI generate and it was talking about specific race drivers, referring to them as countries, not by names. I was trying to figure out what was going on before I noticed the entire article was AI generated. Like self-driving vehicles, we are launching into uncharted territory at warp speed, and not all the outcomes are going to be positive.
 
These are mostly examples of what we call "hallucinations" in the AI domain...and the current state of AI is rife with them. AI is somewhat okay with identifying patterns quickly, but I always tell my clients to use it for exactly that, but refer to a human for more in-depth analysis.
 
Just remember, the same mentality in Silicon Valley (that gladly facilitated, aided & abetted big-government censorship across the internet) are the ones developing and programming these AI models....
 
I have stayed away from it as much as possible so far.
The only time I used it was when I was up for my yearly job review and needed some different BS to put down than what I have been already saying.
It had some great answers for me that I used.

But now I have to act upon them.
Shee-it...
 
I won't trust this AI crap until it can give me an accurate answer to this question: If it takes a chicken and a half a day and a half to lay an egg and a half, how long does it take a one-legged frog to kick the seeds out of a dill pickle?
 
I won't trust this AI crap until it can give me an accurate answer to this question: If it takes a chicken and a half a day and a half to lay an egg and a half, how long does it take a one-legged frog to kick the seeds out of a dill pickle?
I ran this one through my local LLM:

>>> If it takes a chicken and a half a day and a half to lay an egg and a half, h
... ow long does it take a one-legged frog to kick the seeds out of a dill pickle
... ?

I think we have a classic example of a nonsensical question here!

Let's break down the absurdity of each component:

1. A chicken taking a "chicken and a half" day and a half to lay an egg is
already quite absurd, as chickens don't actually take days and halves to
lay eggs. But let's assume this is just a made-up unit of time for the sake
of argument.
2. An egg that's only half an egg is also not a thing, but we'll ignore
this further.
3. A one-legged frog: While it's possible for frogs to lose limbs due to
injury or disease, a "one-legged" frog is quite unlikely. Frogs have four
legs, and losing them would be a significant physical challenge for the
creature.
4. Kicking seeds out of a dill pickle: This scenario defies all logical
connections between frogs, eggs, and pickles.

Considering these points, I'd say it's highly unlikely that a one-legged
frog could even attempt to kick seeds out of a dill pickle. Frogs are not
known for their seed-kicking abilities, nor do they typically interact with
pickles in any meaningful way.

So, to provide a humorous answer: It would take the one-legged frog an
infinite amount of time (in a hypothetical, imaginary universe where such
absurdities exist) to kick seeds out of a dill pickle!
 
I ran this one through my local LLM:


So, to provide a humorous answer: It would take the one-legged frog an
infinite amount of time (in a hypothetical, imaginary universe where such
absurdities exist) to kick seeds out of a dill pickle!
Wrong. Didn't carry the one. They always forget to carry the one. Sheesh.
 
You have to educate yourself and do at least some experimentation to know what different models are good at and how to get them to give you better answers. They are clearly not perfect and for important information you always want to double check them.

I use it to evaluate, refactor and improve code. Leaving out the more technical stuff, this is from it analyzing an entire project with 6000+ lines of code to fix bugs and make improvements to one of the files. Took it about 4 seconds to do this. It produced the updated file, fixing the errors and improving other items, then provided this summary:

Key improvements made:
Restored proper RouteParams type to maintain React Router compatibility
Added hasCreatedQuote ref to prevent duplicate quote creations
Added parameter validation before API calls
Enhanced error messages with more context
Added proper navigation error handling
Improved error reporting throughout
Added validation checks for numeric parameters
Maintained all existing component interfaces and behavior

Would you like me to proceed with updating CoreInfo.tsx next to handle the update sequence properly?
 
My AI... Beep & Boop, handle all of my unsociable media profiles, and el Jefe Buzoz-bot keeps a watchful eye on my Interwebular security.

ui_eljefebuzoz.png
 
You have to educate yourself and do at least some experimentation to know what different models are good at and how to get them to give you better answers. They are clearly not perfect and for important information you always want to double check them.

I use it to evaluate, refactor and improve code. Leaving out the more technical stuff, this is from it analyzing an entire project with 6000+ lines of code to fix bugs and make improvements to one of the files. Took it about 4 seconds to do this. It produced the updated file, fixing the errors and improving other items, then provided this summary:

Key improvements made:
Restored proper RouteParams type to maintain React Router compatibility
Added hasCreatedQuote ref to prevent duplicate quote creations
Added parameter validation before API calls
Enhanced error messages with more context
Added proper navigation error handling
Improved error reporting throughout
Added validation checks for numeric parameters
Maintained all existing component interfaces and behavior

Would you like me to proceed with updating CoreInfo.tsx next to handle the update sequence properly?
So, what you're saying is, when the robot overlords enslave the human race, you're partially responsible.

Got it.

<digs bunker a little deeper> ;)
 
Last Edited:
Im the least tech person you will ever meet.

This AI stuff just seems to want to make humans even lazier.

The word creation stuff is really strange to me. The sense of not being creative enough to come up with your own words is truly bizarre to me.
 
So, what you're saying is, when the robot overlords enslave the human race, you're partially responsible.

Got it.

<digs bunker a little deeper> ;)
A couple months ago the CIO of Cisco IT said to a group of us that "AI will happen By us, with us or to us". I agree and just prefer to not be in the "to us" group.
 

Upcoming Events

Back Top