Can everyone please chill? Thoughts on AI from a moderate perspective

Posted on 2026-08-30 by [terra]

This is the opinion of one person. One young, naive, software engineer who just wants everyone to chill out. This is by no means comprehensive.

Can we all just, like, relax?

Can the AI hype-rs making all those startups and talks about how LLMs are going to become AGI just chill? Claude is at the level of a very talented intern when it comes to both coding and general intelligence.

Undeniably very very smart, but an intern who makes silly mistakes and needs guidance/supervision

Can the AI haters who talk about using AI for writing code/chatbot purposes as being morally evil also just like...relax?

Can we recognize that a tool can be both bad for the environment and very important/productive for one's life? You know, like the automobile???

Can everyone just like...chill tf out

. . .

Let's start with the pro-AI camp. The silicon valley hype train conductors.

LLMs are revolutionary tools for writing code

This is very hard to argue against.

Anyone who's used one can attest to how good these LLMs can be when directed carefully. The code they write is, frankly, good in most situations. With supervision and direction, you can basically stop thinking about the line to line details/syntax of your code and think purely at the function/module/type levels of abstraction.

It's very hard to deny that this is a big shift in how code is written, and will be written in the future.

But its capabilities are like a talented intern

A framing device that I've found extraordinarily useful in contextualizing my AI usage and my trust level is treating it like a very smart intern.

This is based on my own experiences using LLMs for coding, as well as some secondhand experiences.

LLMs are currently at the level of talented interns

Possibly smarter than any human alive in terms of breadth of knowledge and strict algorithmic recall.

But as an engineer? Undeniably entry level at best, intern level in most cases.

In my experience, LLM coding capability matches this description exactly

Doesn't that line up with a very talented intern?

Can we just chill and realize that this is all that LLMs are capable of right now?

With how they are now, does this massive, massive wave of LLM based tooling really make sense?

Automated "engineers" for coding, would you trust an intern with that level of control?

Handing off important office tasks like writing emails, making slides, analyzing spreadsheets...would you let an intern do that unsupervised?

Would you trust an unsupervised intern as a medical professional?

LLMs are revolutionary for rote, mechanical, mundane tasks, the kind you can safely hand off to an intern. The revolution is that everyone was handed one of these extremely smart interns who will work around the clock and at a moment's notice.

Asking an LLM to work on its own without supervision is just asking for trouble.

So can we just like...chill on the AI startups and crazy hype?

Can we recognise the tool's strengths while not ignoring its weaknesses?

Now let's talk to the other camp.

LLMs are just not that bad in the grand scheme of things

Let's address the main arguments that are typically raised.

I'll be arguing that these points are valid, however missing nuance.

1. Environmental and other moral arguments

It's undeniable that the training of LLMs is extremely power hungry and uses a lot of electricity, often to the detriment of wherever the datacenter is located.

This is cited as a common argument against the usage of AI, even if usage of the tool is environmentally cheap, supporting said companies will support them in performing bigger, more expensive trainings, leading to more environmental harm.

This is a fairly sound argument in my opinion.

In a vacuum anyway.

How does this fit into our reality of capitalism?

There is no ethical consumption under capitalism.

I'm not going to sit here and pretend to know the philosophy involved here, that's more nuance than I'm capable of as a silly, young software engineer.

But like, we don't go witch hunting after car drivers, do we?

Cars are undeniably damaging to our environment, right?

Why do LLM users get this disproportionate amount of judgement thrown at them?

Why the higher moral standard?

If you judged someone's morals by their every purchase, shouldn't you be blocking anyone who eats Lay's? Drinks Coke? Eats meat? Buys jewelry? Uses anything with a lithium-polymer battery, like the phone you're probably reading this on?

Doesn't seem so fair now, does it?

There is no ethical consumption under capitalism.

My understanding of that statement is: when existing under capitalism, moral purity is impossible.

Using LLMs doesn't appear to be demonstrably worse than the other examples by my count, so can we just like...chill on the moral purity argument?

2. Societal impacts

This one is for the poor open source maintainers.

Or really any maintainer of public knowledge.

There are two undeniable effects that have come about as part of the LLM uprising.

  1. Free knowledge is being harvested and exploited, to the detriment of the maintainers of such knowledge. Similarly for free tools. Examples of this: Wikipedia, copyright infringement, LLM traffic to the internet at large.
  2. The massive wave of "contributions" to open knowledge and codebases from LLMs is causing their maintainers to be overwhelmed. Ask anyone in these spaces and they'll tell you.

Both of these are actively happening, and to ignore them would be to put one's head in the sand.

But are they a reason to say that LLM usage is unethical?

I think both point towards two different directions of the same thing: regulation.

LLMs need immediate regulation

At a legal level, and at a community level.

I'd like to argue that under capitalism, there isn't much we can do about companies being bad actors. Our only choice is to boycott, but that's really only possible from positions of privilege.

But we don't exist under only capitalism

In a functioning democracy, our collective desire for our free resources to not be exploited should be reflected by our representatives.

In a functioning democracy, laws, lawsuits, fines, sanctions, injunctions, and other tools should be wielded to protect our public interests.

One can dream, right?

Community

But what we do have control over are our little communities.

And many have adopted stances that stem the bleeding: blanket bans on LLM based contributions.

However, I want to argue for more nuance.

LLMs have vast vast latent knowledge embedded within them, and thus have a lot of potential for insight when pointed in the right direction.

LLMs are very good at internet research, being able to read papers and google search at a rate much much faster than we can.

LLMs can hold so much more knowledge than our human brains. Their ability to correlate information from various sources (like source code for debugging and the aforementioned research) is so undeniably useful.

And finally, like I argued earlier, LLMs are excellent at rote tasks, the kinds of things that a smart intern could do.

A blanket ban on LLM contributions kills all that, so I just want to see more careful policies.

I really like the @mycoliza AI usage policy in her project Mycelium

your code (and any associated prose, such as commit messages and documentation) must not look like it was written by a large language model.

This makes a ton of sense to me, and she elaborates in the doc further details that you should read.

But doesn't that policy just...work?

LLMs are allowed to be used to make contributing easier and make insights that a human brain just can't (or at least mine can't), while also blocking the main source of toil from maintainers, dealing with a deluge of unsupervised slop with the goal of resume building.

If this policy was followed, wouldn't it just work?

Now a new can of worms: AI art.

3. "Killing" art

This argument is, to me anyway, pretty baffling.

Those that argue using LLMs for writing or art are killing the practice of...making art?

I get the capitalist argument: why spend money on an artist when you can generate it with AI, thus forcing artists to change careers.

But like, did fast food kill the food industry?

In the same way that a consistent, mid, okay burger from a fast food chain isn't what you always want, can't AI art be treated the same way?

Recognizable in style, often pretty mid, and not always what you want?

Like, if I just want a cute drawing of me and my cat in matching bee costumes, using AI to generate an image isn't taking money away from an artist. I just...wouldn't pay money for that in the first place?

If I want a nice painting I can put on my wall of me and my cat in matching bee costumes, I'll reach out to an artist instead of trying to generate it.

If a company wants to serve me the equivalent of fast food as the front page of their business, isn't that just their choice?

You can just choose not to shop there if you see that they're serving slop.

Now, nuances.

But what if it's unrecognizable?

First of all, that feels like a strawman argument. No such AI generator can make art of the same caliber as a talented artist. As far as I'm aware anyway.

And if there is one day, shouldn't the world change accordingly?

So many careers have died due to new technologies, why should art be any different? (Again, accepting the straw-AI in this argument)

But to create art is a human endeavour that brings joy!!!

Yes, it is!!!!

One thousand times yes.

Which is why I do it for fun!

But losing it as a career means we'll never develop new artists and new art!!!

Aha, gotcha.

Remember the straw-AI that led us down this argument? That there exists an AI that can create indistinguishable art?

If artists become worse at art than machines, then that just means the career will shift.

If one can create novel art that is better than the AI, then they can forge a career

Besides, there's another clear career that defies this line of reasoning: professional chess players.

Chess engines became better than humans a long time ago, but pro chess players are still around...why?

Because we recognized the inherently human talent and skill involved in being a great chess player, and chose to foster it

If AI is to become better than humans at creating art, that doesn't mean it'll die as a career, it just will change.

And much like how chess engines didn't kill playing chess for fun, AI won't stop you from composing a nice song or writing a beautiful poem.

Or in my case, rambly blog posts.

In conclusion

To the people making startups or vibe coding: just chill please. LLMs are at the level of very smart interns and should be treated as such

To those arguing moral arguments: no ethical consumption under capitalism

To those maintaining open knowledge: First of all, thank you for your time and effort. Second of all, reconsider blanket bans on AI, there's good that you're missing out on

To struggling artists: I think your careers are fine. And if not, the practice of humans making art isn't going to die.

I just had to get my thoughts down somewhere and this is just how I do things. By writing my silly blog posts.

. . .

Stay safe out there and thanks for reading!



Liked this post? Please share your thoughts on fedi/bsky!

Subscribe to get emailed about new posts.