Let’s Be Honest About AI

Post Reply New Topic

AI has become this massive blanket term that people throw around without understanding what it actually is. People talk about it like it’s some sentient god-in-a-box ready to take over the world. Spoiler: it isn’t. What most people are calling "AI" today—especially the stuff getting the hype and the hate—isn’t intelligent in any human sense. It’s a Large Language Model (LLM). It doesn’t think. It doesn’t understand. It doesn’t learn in the way living minds do. It statistically guesses what the next word should be based on its training data. That’s it.

In reality, it has less sentience than a bucket of seawater, and I’m not even being dramatic. A tidepool reacts to its environment. A language model doesn’t even know what an environment is. It doesn’t know what you are. It doesn't know what it is. And yet people are either terrified of it or worshipping it.

Yes, it can write, summarize, simulate conversation, and sometimes it feels pretty impressive. But it gets things wrong a lot—sometimes hilariously wrong, sometimes dangerously wrong. These tools are useful, yes. They save time, spark ideas, and automate tasks. But they're not alive, they're not conscious, and they're certainly not going to take over unless we give them too much control out of laziness or fear.

The real problem isn’t that AI is taking over—it’s that people think it is, and they act accordingly. Corporations cut jobs, governments panic, artists rage, and the public imagines a robot apocalypse because nobody’s taking the time to actually understand the tech.

So let’s call it what it is: a tool. A limited, fallible, probability engine that spits out words. Sometimes it's helpful. Sometimes it's wrong. But it’s not magic. It’s not sentient. It’s not evil. And you could switch it off if you wanted. It doesn’t spread the internet—people spread it. People are the ones building tools to automate processes, and people are the ones fueling the fear. So much of the hate AI receives comes from the very systems and behaviors people created and now refuse to take responsibility for.

It’s a mirror.
And the real danger?
What we project onto it.

The real problem IMO is we waste gigawatts of electricity (and then crapton of water to cool it), overload smaller sites and shittify internet even further because of some glorified data harvesters producing awkwardly toned BS. And that people really buy that.

And some people do indeed lose job 'cause letting LLM create half-baked abomination with artifacts is far cheaper than renting a proper graphic designer. It's both comical and dejecting to see what kind of crap makes into billboards, advertisements or even decorational stuff; Like I totally don't have a sticker with sunflower-imitating abomination (with huge weird-looking hole where seeds are supposed to go) on our fridge, as my mother just doesn't care.

You're raising valid concerns, and I hear the frustration. But I think it’s important to put some of this in perspective.

First off—what you’re describing with the energy and water usage? That’s not unique to AI.
Most of that criticism applies to server farms in general, which have been powering everything from social media to online banking to streaming services for years. AI models like LLMs and image generators do require a lot of computing power, especially during training—but once deployed, many of them are more efficient than people think.

The problem isn't just the LLMs—it’s the scale and design of our entire digital infrastructure. We're talking about the same kind of energy consumption that powers Google searches, Netflix recommendations, TikTok filters, and yes, those annoying autoplay video ads too. It’s a systemic issue about how we build and maintain data centers, not just the tools we’re currently using.

As for the design jobs being replaced by AI-generated content—again, you're right that it’s happening.
Companies are cutting corners and using quick, cheap generative tools to create stuff that’s half-baked, awkward, or downright bad. Why? Because it's fast, it’s cheap, and most consumers don’t notice—or care. The sticker on your fridge is a perfect example: people buy this stuff not because it’s good, but because it’s available and just good enough to not trigger a second thought.

But here's the deeper issue: this isn’t about AI vs. humans—it’s about how much our systems value aesthetics, quality, and human effort. The same thing happened when stock photography flooded the market, or when print went digital and newspapers started relying on listicles instead of journalism. We keep swapping craft for convenience.

AI didn’t invent this problem—it’s just the next wave of it.

And yes, a lot of the output is awkwardly toned BS, because LLMs don’t think. They simulate tone and logic based on their training, which means they can only approximate meaning. They don’t know what they’re saying. That’s why the results can feel so hollow or uncanny.

But—and this is important—people buying into it uncritically is also part of the problem. Not because they’re stupid, but because we’ve created a culture that prizes more over better, speed over substance, and cheap over crafted. Until that changes, AI will be used as just another cog in the machine that values profit over quality.

So yes, it's comical. Yes, it’s dejecting.
But the real enemy isn’t just AI. It’s complacency, cost-cutting, and the devaluation of skilled human work.

Want to resist that? Support artists. Hire creators. Share things that are actually good instead of just algorithmically efficient. We can still steer the culture—if we care enough to try.

"Google’s emissions up 51% as AI electricity demand derails efforts to go green -
Increase influenced by datacentre growth, with estimated power required by 2026 equalling that of Japan’s"
https://www.theguardian.com/technology/ ... orts-green

It’s a good example of why I find The Guardian increasingly hard to take seriously. The headline alone is designed to provoke panic—“Google emissions up 51%!”—with little context offered upfront. Yes, AI and data centers use a lot of power. But framing it as if the entire tech sector is about to single-handedly destroy the planet is more about stoking outrage than delivering balanced insight.

There's an old adage: "50% of all statistics are wrong." It’s tongue-in-cheek, but relevant here. Raw percentages without baseline context are misleading. Is that 51% increase over a historically low post-COVID level? What’s the actual source of emissions—training models, or just global cloud usage in general? Are these emissions net of AI helping reduce waste or optimize energy elsewhere?

The Guardian, like many modern outlets, increasingly feels like an echo chamber for progressive anxieties—climate, tech, inequality—often framing complex issues in emotionally loaded, binary terms. The result? Less room for reasoned discussion, more room for algorithm-chasing headlines and comment wars.(no I'm not a climate denier)

It’s not that the article is false—just that it’s framed in a way that reinforces a particular worldview, and that’s what makes it feel less like journalism and more like narrative steering. Like all newspapers its framed from a political perspective not an objective one!

^^^ Generally I don't think it's AI to be blamed, it's people who misuse technology indeed (Hence "we", not "AI did it"). But nonetheless it does for yet another outlet to waste resources on (like these ads, tracking, DRM), that's my point.
I agree we vote with ourselves, and that's why I don't use / never deployed LLMs and reject "content" where LLM was engaged (it's still pretty easy to spot, especially with imagery). Even if we put ethics aside, it's just utter crap, seriously.

and that is most certainly fair enough, I use it for role play a translator because I am not the most easy person to understand and to write quick Emails when I'm pressed for time, to use it constantly would be Ethically Wrong although it has spit out a CV/Resume that has impressed even the most pudding brain people at the unemployment centre.

You (and the other contributors here) raise a valid point Sable Noctis,

Might I propose that:

* Practical:
- The energy hungry human race for knowledge, comfort ... has indeed created a huge contributor to our growing energy deficit with the introduction of AI LLM's.

- But, that I think all of us when we were 'pubescent dreamers' held the belief that AI intelligence would even have reached further than it currently has in 2025. As such I propose the phrase that my vendors often offer me: "It works as designed sir"

* Social:
- As every big advance or new "tool" humanity creates or discovers a long period of 'jerky' integration will follow. The bigger the social impact of the "tool" in question, the longer the integration time. And perhaps with AI we've bitten of more than we can chew...? perhaps...I don't yet know... But I can't but think back to my 1st grade teacher proclaiming that: Gaming would "blow-out" our minds potential for reasoning...

- The advise I gave my kids is: Be Extremely deliberate and mindful of what you make yourself dependent on!

kind regards,
Kada

"power required by 2026 equalling that of Japan’s" - that's the juicier portion of the headline, to my mind. The Japan part.

Thank you, Kada — that’s a really well-reasoned perspective, and I think you’ve nailed something that often gets lost in the noise: we’re living through an awkward adolescence of integration with a tool far more disruptive than we may have anticipated, even if it’s not quite the “thinking machine” we imagined decades ago.

You’re right — many of us did expect AI to be further along by 2025, maybe even sentient or deeply embedded into our everyday decision-making. But instead of HAL 9000 or Data from Star Trek, we got Large Language Models — brilliant at mimicry, but hollow on the inside. And you're spot on with "it works as designed, sir". It really does. It just wasn’t designed to think — it was designed to predict, and sometimes we forget that’s a very different thing.

I also appreciated your social lens — the jerky integration idea is exactly what we’re seeing. Every leap in tech comes with societal whiplash. Printing presses caused panic. Electricity was feared. The internet created moral chaos. Now AI is being shoved into everything from classrooms to hospitals, with little pause to ask, "Are we ready? Should we be?"

That quote from your first-grade teacher hits hard too. I remember similar warnings — that video games would rot our brains, that calculators would ruin our math skills. And yet here we are, still reasoning, still creating. Maybe AI will fall into the same category: not mind-destroying, but mind-changing — depending on how we relate to it.

And your advice to your kids? Gold. "Be extremely deliberate and mindful of what you make yourself dependent on." I don’t think I’ve seen a better guiding principle for living with modern technology — not just AI, but all of it.

Really appreciated your take — it added some needed clarity to a conversation that often spirals into extremes.

[/quote]
"power required by 2026 equalling that of Japan’s" - that's the juicier portion of the headline, to my mind. The Japan part.[/quote]

An equally appreciated reflection, Kada — though I must gently contend that your attention, while valid, seems disproportionately anchored to the infrastructural implications — particularly the power requirements — as interpreted through the lens of politically entangled media apparatuses. Much of the discourse surrounding AI’s ecological footprint is indeed rooted in third-party reporting, often filtered through editorial biases and an incomplete technical understanding of the systems they're critiquing.

The tendency of such narratives is to conflate computational scale with sentient threat, drawing alarmist equivalences between energy consumption and existential risk. In reality, most so-called "AI" systems, including LLMs, remain functionally illiterate to context, devoid of intention, and wholly incapable of agency. The media, grasping for metaphors that sell column inches, ascribes to these models the semblance of cognition — and in doing so, creates the illusion of a mind where none exists.

It becomes a kind of epistemological sleight-of-hand: sensationalism masquerading as insight. Their critique is less a forensic audit of technical systems and more a reflection of cultural anxiety toward tools we barely comprehend — yet widely disseminate.

We must, therefore, differentiate between the critique of usage and the critique of nature. An LLM over-deployed or implemented irresponsibly may very well tax infrastructure and displace labor — but it is not doing so because it’s intelligent. It is doing so because we, as a species, are prone to over-application of novel tools before we establish ethical scaffolding or technological maturity.

Which circles back to your original point about deliberate dependency: your advice is wise precisely because it shifts the locus of concern from the machine to the human relationship with the machine. In the end, AI is a mirror — and perhaps what unnerves us is not what it is, but what we reveal when we gaze too long into it.

As a former graphic designer, *sighs* yeah. Companies do love to cut costs. However, they will quickly backtrack on this if it no longer proves to be economically viable. A friend of mine works at a studio and the higher ups were experimenting with AI, but they rather quickly pulled the plug on that when an investor threatened to cut funding.

Companies are still hiring graphic designers, however, there's a lot of...what I'd call swipe and ghosting out there, so you have to be wary. When applying for a graphic design position, it has become the norm to complete a design task. This is typically unpaid work that tests your ability to see if you are up to standard. Unfortunately, certain companies post fake positions, ask for free work posed as a design test and then once they have that work they'll claim it as their own and never talk to you again. This is a somewhat common scam.

I personally don't accept design tasks unless they've interviewed me first. Even then, I consider how long the task is going to take and will likely watermark my work. One place in particular wanted seven design tasks. Seven. You don't need seven design tasks. That's what my portfolio is for! Another place had the audacity to ask me to show up to a group interview at 9AM and work till 5PM for free. AKA bringing in an entire team of designers and having them work a full work day for nothing. Likely hiring no one.

Then there's the expectations! It's not enough to be a graphic designer. Places want someone who knows motion graphics, 2D animation, 3D animation and modelling, photography, videography, SEO, programming and so on. Of course for the same low wages. They want an entire creative team but only want to hire one person to do it all.

I don't blame people who give up. It's difficult to get into the industry and there's so many hoops. Now I'm just stuck going between temporary retail job to temporary retail job. There are still decent companies out there that aren't trying to scam you for free work or replace you with AI slop (that's so bland it's insulting). I need to get out of my crappy town and find a decent company to work for. I hate it here. There's an AI poster in my town that's just so ugly. It's the most cutesy non-descript design. One time I brought a date to a café nearby and even she remarked on how awful the poster looks with the messed up anatomy and sickly sweet aesthetic.



I don't think it'll replace us artists just yet. We're a resilient bunch.

you know this thread was supposed to be 'no AI isn't going to rule the world because its not sentient' its turned into 'the server rooms are hotter than the Sun' and 'wasting water and Overall economic and moral cost of it even existing in a world that practically evolving toward it' utilizing 3rd party new articles and Social religious(used in the broadest terms) and political ideals. all I can do at this point is laugh.

Seen that video liked outer worlds now I'm on no mans sky and loving it. beyond that... I cant even be bothered anymore.. So lets just say yes.. no.. yes and no.. and leave it at that

Sorry, I know I went a tad off topic.

Would this video bring us back on topic?

The trick is to make your own LLM, and have it run locally on your own GPU if you're technically savvy enough! Then it only runs for about 3-5 seconds to answer your response, then cycles your gpu down. Environment saved! New tool added to the tool belt!