Are you ready for AI?

[sparky]

Have you noticed it? No matter who you call or when you call, EVERYONE is currently experiencing higher call volumes than usual. It never fails.

they're relentless....

~S~

 

[Lockback]

Overall, I'm not a fan of AI. I see far too many potential dangers.
However this is the best use of AI I've ever seen.

 

[trevorjohnson83]

when they use vision and cameras in robots and the robot sees a pair of scissors, does it have to review many images of scissors? what is the problem that data centers need to store so much information? is it because they use vision?







The way robots "see" and the reasons data centers are overflowing with information are closely linked, but the process is becoming more efficient than it used to be.

How Robots "See" Scissors​

Traditionally, for a robot to recognize a pair of scissors, it would indeed need to be trained on thousands of images. This is because a robot doesn't inherently understand what "scissors" are; it only sees pixels.

• The Dataset Problem: To ensure the robot recognizes scissors in any context, developers show it images from different angles, in different lighting, and in various states (open vs. closed). This is called Supervised Learning.
• Feature Extraction: The robot learns to identify "features"—the pivot point, the loops of the handles, and the straight edges of the blades.
• Modern Shifts: Today, many systems use Foundation Models (like those used in GPT-4o or Gemini). These models have already "seen" the entire internet. So, a specific robot doesn't necessarily need to review thousands of images of scissors again; it leverages a pre-existing understanding of the world.

---

Why Data Centers Store So Much Information​

You are correct that vision is a massive contributor to data storage needs, but it isn't the only one. Data centers are under pressure for three main reasons:

1. The "Weight" of Visual Data​

Visual information is "heavy." A single high-definition image contains millions of pixels, and a video is essentially 30 to 60 of those images every second. When millions of users upload videos or when thousands of autonomous systems (like self-driving cars) upload logs to the cloud, the storage requirements become astronomical.

2. Training Artificial Intelligence​

To make AI smarter, companies keep massive "data lakes." This includes not just images, but text, sensor logs, and scientific measurements. To train a model to be accurate, you need diverse, high-quality data, and you often need to keep the original data to "re-train" the model as new techniques emerge.

3. Redundancy and "The Cloud"​

Most data in a center is stored multiple times. This is called redundancy. If one hard drive fails, your photo or file shouldn't disappear, so data centers keep several copies across different physical locations.

The Cooling Challenge​

Because these servers are constantly processing "heavy" visual data and complex math, they generate immense amounts of heat. This creates a secondary problem: thermal management. Data centers often spend as much energy on cooling the computers (using massive fans or liquid cooling systems) as they do on actually running them. This is why many modern engineering efforts are focused on finding more efficient ways to "vacuum" heat away from these dense server clusters.

 

[Muhammed]

What about your conclusions? You mean you trust a machine more than your own judgement and decisions?

No.

I spend a lot of mental energy attacking my assumptions, reasoning and conclusions from every angle.

I often prompt AI to debunk me and find flaws in my reasoning. It saves me a lot of time and mental energy. It's a force multiplier.

 

[toobfreak]

I spend a lot of mental energy attacking my assumptions, reasoning and conclusions from every angle.

Too bad not more people do that. Especially the Left.

I don't think they ever test their assumptions.

Just yesterday I had some drip here give me a red 'Disagree,' so I asked her disagree with WHAT? And why? And to offer whatever evidence she had to support that view.

I'm a logical man who is used to molding my views to whatever the prevailing evidence suggests is most right.

She refused. Said it was a waste of her time, she'd never change my opinion, so does not bother discussing nor debating anyone on anything unless it is only to agree with someone who already thinks like her.

Now there is a real challenging outlook on life.

 

[trevorjohnson83]

Too bad not more people do that. Especially the Left.

I don't think they ever test their assumptions.

Just yesterday I had some drip here give me a red 'Disagree,' so I asked her disagree with WHAT? And why? And to offer whatever evidence she had to support that view.

I'm a logical man who is used to molding my views to whatever the prevailing evidence suggests is most right.

She refused. Said it was a waste of her time, she'd never change my opinion, so does not bother discussing nor debating anyone on anything unless it is only to agree with someone who already thinks like her.

Now there is a real challenging outlook on life.

you can get sassafrastinastic

 

[shockedcanadian]

Have you noticed it? No matter who you call or when you call, EVERYONE is currently experiencing higher call volumes than usual. It never fails.

You see, it used to be that companies had staff manning CSR lines and they would figure out the number of calls coming in per time and day and have a staff there to deal with it. Sometimes you might wait a couple minutes to get a person or that person might sit free a few minutes before getting a call. Dear God, that was waste having someone sit idle for even a second.

So they convinced everyone to buy cellphones, told them they wanted and needed one and even steered apps so that you NEEDED a cellphone--- now you presumably carry your own phone company around in your pocket now and they realized better to have you wait for them than their waiting for you.

So now, wait times to reach support that used to be seconds to 1-2 minutes vary from 5-10 minutes to 20 minutes to even their calling you back in 1-3 days and by gum, you better be free, available and ready for that call whenever, no matter what the caller ID says or whether you are sitting on the crapper.

Now they use machines at all cost programmed to deal with most common needs so that you don't tie anyone up at all. Instead of a team of customer support, now they might have one person and when you call, you wait until that person is free or free to call you back, hours or days later.

And if they miss you, they will never try again. They have better things to do than wait on customers.

The savings are incredible because basically, they've eliminated a major department, worse, they have in effect shifted a lot of the expense onto your shoulders; companies now expect, depend and rely on you owning a cellphone, usually needing texting too.

You will also note that cust. service tends to be better with private companies whom might lose your business and buy your product off someone else if too pissed off, vs. public companies like utilities, gov. agencies, etc., where they are a monopoly and like their cust. service or not, you are STUCK dealing with them.

Bottom line is that customer satisfaction is a low concern to them and they want to condition customers to getting used to and accepting suck-ass cust. support, even better, they try to steer you only their website and go hunt for your own answers and solutions. Now you need internet service and a computer too.

Interesting take and I cannot disagree.

I would say this, the declining customer service we are experiencing is in line with the new Chinese global market in particular and the global marketplace in general.

You can still easily get your money back from a purchase at Wall Mart, but if you buy something from Amazon if you want to send it back due to false representation of it is not what you expected, a produce that costs $20 will charge you $11.99 to "restock".

Or, consider the outsourcing of CS to places siuch as India. They are a nation of a billion people, customer service is not in their DNA, yet, since they can speak English you are treated like a pain in their asses for a refund.

The global market will decrease ournstandards. As volume replaces quality and loyalty, this will not improve.

 

[The Duke]

Overall, I'm not a fan of AI. I see far too many potential dangers.
However this is the best use of AI I've ever seen.

But those are the best political ads I've ever seen in my life! Bar none!
Holy crappo!

 

[trevorjohnson83]

shouldn't robot made stuff be cheaper? isn't that the advantage we are looking at?

 

[Captain Caveman]

Good old AI .... I use it daily. How good is it? Will you make errors using AI? This great human being addresses it and more. This woman is a daily content creator and has excellent topics. Try watching her on YouTube and see how you think she does.

I don't entertain it, I try to avoid it.

I'm from the 60's/70's, so I don't need it. I can see why current generations need it for help for basic things.

 

[Nobody911]

If you can keep the virtual world separate from the real world, you can definitely enjoy AI.

Do you think the video down below is real or not? lol.

VaD1uXKhAvtCzPov

 

[trevorjohnson83]

I don't entertain it, I try to avoid it.

I'm from the 60's/70's, so I don't need it. I can see why current generations need it for help for basic things.

maybe someday mr caveman they will photocopy all books and use it for an AI memory