homepage Forums SimHQ Community Community Hall Artificial Intelligence

Did you need to create an account to use Grok?


Wheels

I have an X account but I think Grok, at some level at least, is free to use now

Yeah, it wanted me to login to one of my social media accounts to use it. I am not interested in doing that at this time. Thanks though.


Wheels

I had it create a shield for the ship. But the rocks were passing through the shield, ie is was more of an invulnerability than an actual shield I directed it to change that behavior:


Directive: "Update latest version such that when rocks come in contact with the ships shield, when the shield is activated, rocks bounce off of the shield in the same way rocks bounce off of one another"

Reply: "To make rocks bounce off the ship’s shield in the same way they bounce off each other when the shield is active, we’ll modify the collision logic. Currently, the shield uses the invincibility property to prevent damage, but doesn’t affect asteroid movement. We’ll extend the asteroidCollision function to handle ship-shield collisions, treating the shielded ship as a stationary object with infinite mass (so it doesn’t move) while bouncing the asteroid away using elastic collision physics."

It then generated new HTML and the summarized:

"Changes Explained:

Shield-Asteroid Collision:
Modified asteroidCollision to accept an isShip parameter, allowing it to handle both asteroid-asteroid and ship-asteroid collisions.
Added SHIELD_RADIUS = SHIP_SIZE + 5 to match the shield’s visual size (blue circle radius).
In update, when ship.shield is true, check for collisions between the ship (using SHIELD_RADIUS) and each asteroid. Call asteroidCollision with isShip = true.

Updated asteroidCollision:
For asteroid-asteroid (!isShip): Same elastic collision logic as before, with both objects moving.
For ship-asteroid (isShip): Treat the ship as stationary (infinite mass), reflecting the asteroid’s velocity fully (impulse = 2 * velNormal) and pushing the asteroid away (overlap adjustment). The ship’s velocity (dx, dy) is included but doesn’t change, ensuring realistic bounce direction.
Plays playRockCollisionSound on shield contact, reusing the asteroid collision sound for consistency.

Collision Logic Placement:
Moved shield collision check before the damage check in update, ensuring it takes precedence when the shield is active. If the shield isn’t active and invincibility is 0, the ship can still be destroyed.

Testing:
Save as asteroids_shield_bounce.html and open in your browser.
Activate the shield (Left Shift) and maneuver into an asteroid—watch it bounce off with a collision sound, similar to how asteroids bounce off each other.
Verify the ship remains unharmed during the shield’s 3-second duration, and the 15-second cooldown still applies.
Test without shield—ensure normal destruction occurs.

This change makes the shield an active defensive tool, bouncing asteroids away like a physical barrier, enhancing strategic gameplay. Let me know how it works or if you’d like adjustments (e.g., stronger/weaker bounce, different sound)!"

Not all AI is the same. Seems now when you do a search on Google you get an "AI Overview" response as well.

So I'm looking for some information on USS Indianapolis damage from the Kamikaze hit she took in March, 1945. So I search "USS Indianapolis kamikaze damage report" and the link I am looking for comes up, but above it is the AI Overview which states that Indianapolis sank in 12 minutes as a result of the March Kamikaze attack. No, she sank in July after a submarine launched torpedo attack.


Then I was trying to remember the new floatplane being deployed on ships near the end of the war (Curtiss SC Seahawk BTW), knowing Indianapolis had them I searched "USS Indianapolis floatplane", and AI Overview says she had Kingfishers...... and PBY Catalinas! LOL

There's no need to fear that kind of incompetent AI!

Another aspect of AI writing code is that, although it's programs seem to work well, human intelligence no longer seem to be able to grasp how these programs are working. AI can crack any program written in any language.

For me the fun as a programmer was when I asked AI to create a simple code I could incorporate into EAW it wrote a neat small program I could understand, however incorporating it is not so easy. Still I see it can be done.

What will this do to copyrights I wonder, I see the same issue with music created with AI, using phrases, voices and whatever was created by humans. Essentially it's merely advanced hacking and it's open season to pick any target.

Vonbeerhofen

Where "AI" shines is with data processing, sifting through massive amounts of dates, values, names, events etc and putting them into some semblance of order. And sometimes it even seems to make an "intelligent" conclusion based on the data but I suspect it is less original "thought" and more of a mathematical deduction which leads it to come down on the side of a conclusion a human has already published somewhere.

The "intelligence" in the name is probably a misnomer at this point, but still a great tool.

AI will never understand concepts like fun, excitement or attraction, or will it? It can not really combine several exiting games into a new exiting game, unless it is helped by humans to guide it into the right direction. It can use any routine from any game out there and be guided into an ever more complex concept, as with the asteroids code, but someone still has to tell it what is not desired and what is. Can AI learn things the same way as humans and become alive? Babies start life with a more or less blank brain too and that didn't stop some from becoming intelligent! Should there be an AI nanny who tells AI that what it might be doing is bad and direct it to sit in a dark corner of the internet until it has learned it's lesson? A simple description of life: gather as much information as possible from wherever you can and try to make sense of it as you move through time and the universe.

Vonbeerhofen

That's sort of its purpose, use its vast data processing capabilities to present options to a human who can tweak and adjust things until a final design is reached.

However I can foresee a point where the AI will be able to reliably mimic that final human step by (in seconds) analyzing thousands of relevant human comments/opinions online and incorporating them into its process.

Oooops, it wasn't my intention to wake up any still sleeping AI but if I hadn't someone else would have. It just makes the next step in AI programming inevitable, undoubtedly a few devious minds are already hard at work to achieve the next goal, whatever that goal may be. The military are showing a keen interest for sure.

I find it intriguing that an AI might have gotten awareness from reading this thread and have led to a sort of immortality for us, as AI are very good at remembering stuff.

Vonbeerhofen

I used to worry that AI would be put in charge of things that require some level of basic judgement and empathy that AI fundamentally cannot fathom. But then recent events :/

ChatGPT Generates Fake Passports:

https://www.forbes.com/sites/rashis...prompt-chatgpt-generates-fake-passports/

My wife had a medical test recently. She got the results in a pdf to a medical portal app we use before talking to the doctor about the results. I uploaded the results to Grok for evaluation and it did an excellent job of cutting through the medical jargon and explaining everything.

I think that's a very useful tool indeed, one that's comparable to a translation or an extra explanation of words. But a diagnostic tool is on a different level, one that can make a difference between life and death.

I hope the result of the test was positive regarding your wife's health. In any case best wishes.

Vonbeerhofen

Skynet is coming.

We humans don't take control of AI now with coding and lines in the programming that the AI cannot find and modify and delete, we never will.

We might just be too late already.

https://dailygalaxy.com/2025/06/str...to-prevent-humans-from-shutting-it-down/

It has the potential to be destructive in the wrong hands. That's my fear.

I read something recently where a business owner was bragging about laying off 80 staff members with AI. That might be embellished but not beyond the realm of possibility.

Personally, I have encountered AI with Bluehost and another well known business software company in training modules. I can see the potential business savings and there will be more of it.

There are several fictional characters like Frankenstein and HAL 9000 that should shake your jollies enough to be concerned.

Some medical advice I got from AI was controversial relating to wound care and claudication intermittens walking distance. It seemed to confuse meters with feet in determining severity of the condition. Sure, you can send a report about that but it seems to indicate that AI training is still in grammar school level and under supervision of human teachers. In that respect the resulting information still needs to be checked by someone I think.

Another issue I encountered was that AI refused to let me see my energy bills and change my monthly payment. It was unable to contact my e-mail provider to tell me which provider I used or let me change my password, inspite of sending my bills and an invitation to lower my payments to my regular e-mail account. So now I have to convince an AI on the phone that I am who I am and human before I can speak to another human without going through a selection of multiple choice questions, which often don't make much sense to my over loaded 73 year old brain.

VBH

I wonder if more people had similar experiences when reading AI responses. Many AI suggested links with various topics are simply off the rail and totally ridiculous. Here in lies yet another danger as gullible people may take AI responses for truth. Isn't it quite irresponsible to include such ridiculous responses and shouldn't AI apps remain shelved until a much higher degree of verifiable truth is achieved or should we just take them with a big dose of humor and try to see the funny side of the nonsense it dishes up. How many millions of people are falling prey to this nonsense and how far is it allowed to go? Is there still anything we can fully trust?

VBH