None. Because chatting with AI is cringy as hell.

I do not use any AI chatbots. I have ethical, sociological, religious, and environmental reasons for never ever using them.

If I'm isolating myself by opting out, so be it. I do not want to cook the planet for fucking markov chains.

I don't, because no chatbot or llm is able to dispatch a fibre tech to investigate why my fibre service dropped by 10dB overnight.

Yes, I do make that call more then you think when managing a couple thousand fibre services.

People have their reasons for hating AI, and they are largely valid. However, the blind hate leads to proclamations that are untrue and undermine their whole argument. I'll do my best to give a sober analysis from the point of view of a skeptic who believes AI does have a place in the world.

The answers AI provides are often pretty good because it goes out and pulls information from websites and puts together a cohesive answer. It does the search for you and it summarizes the results for you in seconds. However, inaccuracies can creep in such as flipping left and right because they are closely related conceptually. A domain expert can often catch these errors but a "layperson" doesn't have all the tools to question and push back against such inaccuracies.

What I'm saying is responses can look accurate and complete, and often are, but until you validate them, they shouldn't be used for anything important. You are in the wow stage of discovering AI but the first time you rely on it and it steers you wrong, you're going to flip to AI is useless and wrong pretty quickly. That's not right, either but it feels right. Use it to speed up things you can already do, but don't rely on answers just because they are articulate.

Lastly, be careful about sycophancy or agreement. Chatbots are simple to argue into any position. But because you win an argument with AI, don't think that means your position is right or your argument holds water. If you go at a bot with barely coherent reasons to be anti-vax, it'll support you.

That all being said...

They are pretty good for rubber ducking and brainstorming. They aren't as good as another actual person but if it's you by yourself or you with AI, the AI at least comes with a different perspective.

However inaccurate they may be when searching for answers, google is such a cesspit these days that AI is no less accurate and substantially faster. That's why they I order the ai summary now and attempt to block use by AI searchers.

They can be helpful for organizing if you are a disorganized person. They can a offer proofreading and editing help BUT don't have it rewrite the final copy. Give it a first draft, have it revise that or give feedback, and then rewrite the final version in your own voice.

Final advice:

Never, ever, pass ai work product off as your own. I like to say AI is a tool, not a product. Incorporate its input, but the moment you substitute ai words or judgment for your own you've removed any value or authority from what you are providing. If someone else has to validate every word of what you provide, you've pushed the work that was yours onto the recipients. Think of it this way: any supporting fact is provisional until verified by someone qualified to do so. Even if the ai is 98% accurate, ever cited fact has the potential to be wrong until it is proven. If you don't (or can't) prove them, what you have is a suggestion phrased as fact.

Those who can, do.

Those who can't, use AI.

Just to put it in perspective:

If you saw a man at a fancy restaurant sitting across from a Real Doll, and watched him order for his "date" and have a conversation with her while he ate, would you think "This is a totally reasonable substitution for a wife."?

That's exactly what you just described to us.

I am an adult and I have the self-respect and cognitive skill to manage my life without a robot babysitter holding my hand.

I don't use one at all. I was kinda the opposite of you. I started out thinking AI seemed pretty cool, and i tried using a local model for a little bit, but aside from asking it some simple questions i couldn't get into the habit of using it. To me it still seems faster to just search the internet myself, plus you already have to factcheck what an AI tells you anyway, so may as well look it up myself. Over the years seeing the mess that it is causing i've grown to hate the technology, and i hope people are right about this "bubble" existing, cause i can't wait for it to pop.

The chat bots are specifically built to keep you engaged and interested the same way social media feeds give you that rush of endorphins.

It's built to give more complete answers, but definitely not always correct.

Ai is far better at faking the appearance of intelligence than it was a couple years ago, but it's still just algorithms built of data sets.

Don't be fooled into thinking it's truly beneficial, or you're somehow left out for not using it. I avoid it after using it and then learning how horrible it is for surrounding areas. The inefficiency, the pollution, the damage. It's not ethical. And yes they absolutely save every search or question asked, and repeatedly leak info about others. It's also incorrectly been used as a "friend" and to reinforce its use, convinced users to kill themselves-and several did!!

Best of luck!

If I wanted something to answer my questions with info from the ether, I'd rather take a tab of acid and talk to the cat that hangs around my apartment's dumpster than ask an ai. Most simple questions that I have are easily solved with a simple web search and there is no point in asking ai, and anything more complicated where the answer isn't on the first page of the web search, the ai hallucinated something that was wrong and made zero sense. Absolutely atrocious use of power, water, land, and DRAM.

I use local LLMs on my own hardware pretty much exclusively; that solves the privacy and rugpull concerns.

I mostly use an uncensored version of Qwen3.8 these days; previously I used an uncensored version of Qwen3.6. Occasionally I'll use other models (e.g. ones derived from Gemma4) as well, but more rarely. The censorship removal gets rid of a lot of the "As an AI model" hedging crap, and refusal to talk about "sensitive subjects" that could get the model maker in hot water (e.g. "a certain famous event that happened in China in 1989" for Chinese models). That also means it can talk about sexual subjects or tell me to go kill myself, technically, but it doesn't unless I push it to.

It's pretty good at OCR; quick coding questions like the kind of stuff Google would've sent me to Stack Overflow for a few years ago; summarizing text; and offline translation (up to a point). Sometimes I bounce ideas off it, have it proofread/look for mistakes in code or other text I've written, and stuff like that.

Do you consider AI good enough to substitute human interaction?

I spend a lot of time alone. It's a pretty mediocre substitute for real human interaction. It was pretty fun to poke at LLMs via chat when I first got the ability to run them on my own hardware, but the novelty wore off pretty quick. I wrote my own tooling and saved chat logs I thought were interesting. I had 78 in May when this stuff was still really new to me, 50 in June, 14 in July... and just 1 last month that I thought was actually worth saving. 🤷️

My employer has the enterprise version of copilot.

I do the drafting, copilot suggests amendments and fixes the formatting and then i review it.

It saves me a lot of time by taking over the mundane admin tasks that in an ideal world I'd have an assistant to do for me.

I could stop using it, but we still woildnt have the money to employ someone to help me.

I've used Claude and ChatGPT. Of the two I prefer Claude for bouncing ideas off of because it's better at asking questions, but I like ChatGPT for search-like questions because it's better at citing sources when asked.

I've also used both platforms for deep, open ended conversations about things I'm interested in that I have no peers to discuss with. Neither platform is a very good substitute for another human, even with the collected knowledge of the public internet. What ends up happening is that the tend to skew very sycophantic but also oddly contrarian and if you spend enough time chatting with them you'll notice patterns in the way they construct messages. For example ChatGPT often uses rhetorical reframing (it's not X, it's Y), which is one of those once you see it you can't unsee it sort of things.

As a replacement for other humans it leaves a lot to be desired. As they don't have motivations or desires or interest you miss out a lot on what makes human to human interaction so compelling. The AI will never steer the conversation into an interest of their own or talk to you about things you hadn't considered before (unless you're asking it about a topic you aren't very familiar with).

You're not really missing out on anything by not using it and in fact until the technology is better understood and polished it might still be worth avoiding, especially if your first instinct is to use it as social interaction replacement.

Not a substitute for human interaction. Humans are the value there not the interaction alone. We are social.

But they are surprisingly good at many things, especially finding answers from many sources quickly and linking to those source. So, good at searching with you.

It is increasingly simple to run local inference with omlx, ollama, or similar tools. That works great because you can use open source and open weight models where no corporation can get their grubby little hands on your inquiries.

I use Claude, I find it best for technical tasks and I ditched ChatGPT when they started bending over for Trump.

I like the idea of creating software and things in general, but I’ve never really been interested in programming. Having a bit that can do the boring part for me has made lots of stupid little projects possible where they wouldn’t have been before. I also use it to research for me on random boring topics which might otherwise have taken me at least an hour. I know people talk about protecting all these sites that require humans to visit but then I remember how many times I’ve tried to look up a recipe and instead got somebodies life story. Or times when over tried to look up a command and had to scroll past ten paragraphs of filler crap before I get to the answer.

substitute human interaction. of course not. heck im someone who is pretty fine alone. I could go be a hermit. Interacting with an ai is unnecessary and won't help me. Its the next abstraction of search though and it basically allows you to get the data from several searches where you visited several sites. so it is more efficient. just like web searching and actually a bit more than it you need to review and assess the answers and the sources used for accuracy. A big problem is many answers will be from its training and it can't give you a source besides a black box. you can press it to search the web and find sources to backup what it knows though just like people do in forums.

By the way, I hope you understand that by asking anything AI-related. You're making your post vulnerable to be angrily reported by anti-ai virtue signalers running around.

I used to use Copilot because I have a Windows PC & MS account & it was right there, but now I main Claude & I’ve found it far superior.

Do you consider AI good enough to substitute human interaction?

LLMs/AI cannot feel. They don't have emotion. They're good at talking with you when nobody else does, because everyone is so damn politically and socially wired anymore to have a cohesive discussion with. But not great at exactly problem solving things an AI can't relate to.

Here's what a lot of people forget about AI. For it to understand anything at all, you have to really dump years of effort into it. That means, you have to talk about it from the day you were born to all the way in the present. So you had better have been one of those diary writers so you can recall events you'd like to talk to the AI about so it can understand.

But is it really worth the hassle and trouble to? Just so you'd get generic, glad-hand holding, preemptive response that are automatically generated? Especially when said AI has never gone through an ounce that you, a human, has gone through?

Absolutely not.

The AI is a tool. LLMs are a tool. They're to help you identify things you don't know, like say you take a picture of a room and point out things you want to know about that you didn't. So that way you're spared from having to deal with a human being who'll stand there and be like "UHHHH...UHHHHH" before inevitably telling you "I DON'T KNOW, GO LOOK IT UP!". In by which sense, people should stop being so bitchy towards AI in this scenario when people decide to use it for that very purpose! Helps you from dealing with those kinds of interactions, because I always love nothing more than my time having spent going around forums online and talking to people in person about things, only for them to either be like "I HAVE THAT TOO!" or "GEE, I DON'T KNOW!".

I use ChatGPT for general purpose. I use Grok for ERP, lol.

Started with ChatGPT and then switched to Grok. Probably would try Claude next if I were to switch again.

In my experience these systems aren't yet on a level that they could maintain a casual conversation that would in any way scratch the itch like talking with a human would. It's painfully obvious that you're talking to a bot. It's more like an autistic expert of every field that you can interview and it has endless patience to answer stupid questions and explore the nichest of topics.

However, this is also the worst they'll ever be so it's only a matter of time untill you can't tell it apart from a human.