>Alright /g/ents, I know I'm a few months late to the party, but I've recently started using Chat GPT and I'm curious to know what practical uses you guys have found for it.
>Personally, I've found it useful for setting my schedule and reminding me of important tasks throughout the day. I've also used it to ask for fashion advice for a first date at the Wakefield Museum, and it was surprisingly helpful. And when it comes to dating apps, Chat GPT has helped me come up with some first messages when the woman's profile didn't give me much to work with.
>But I'm sure there are more applications out there that I haven't even considered. Have any of you found any other practical uses for Chat GPT? I'm eager to hear your experiences and learn how else I can use this nifty little tool in my daily life.
It's close isn't it. And can be quite addictive to start with when you need something that string together various chains of thought. I am actually quite curious to what you've been able to use it for.
Sadly there's no fucking way anyone here is that smiley in their posts.
I haven't seen or read it but that's the problem with franchising media. It's a film about demons (I assume) so a prequel also needs to have demons in. There's nowhere else to take it while staying true to the artist's vision what fans will pay for.
I've been using Gemini Pro but I feel horribly guilty about it and I'm tempted to move over to Mistral for the open-source element. But I've always stopped short because it has less features.
>>29308 Not engaging with something crap isn't the same as pretending it doesn't exist. Generative AI has just joined the pile of shite flatly don't care about, like Borderlands, Ollie Murs and meal deal savings.
If it weren't for the ways in which AI is, first of all, being used as a tool of oppression and surveilence, and secondly, threatening to tank the US economy (and by extension lots of others too), I wouldn't pay attention to it. That's because it's more akin to the "smart" features in some home appliances than any kind of technological revolution. Unfortunately, fascists and chancers alike have bet the farm on it, forcing my gaze to be drawn to it.
>>29308 In several key ways, it doesn't. How would you define AI? Chatbots, reskinned search engines, and image generators are not exactly "intelligent".
I've got a growing feeling that a significant percentage of Youtube features and documentaries are now not only voiced by AI, but the entire script is pulled out of an AI's arse. I guess an AI voice as such is nothing shocking anymore now, but a lot of them are increasingly just not the way a normal human person would sit down and write up a narration. I can't completely put my finger on it, but it goes beyond just the mispronunciation of words, and of abbreviations in particular that AI still seems to struggle with. Just the choice of words. It's like, when you hear an actual youtuber and enthusiast about a certain topic talk to a camera or do a voice over, you just know it's an actual person who is knowledgeable and whose heart is in it. But when an AI voice recites an AI generated script, it's just off. It has a weird air of detachment from the subject and feigned competence about it.
>>29311 You are correct. 'Intelligence' is a marketing misnomer from the 1950s. We are not discussing AGI or sentience. We are discussing applied statistics: sophisticated predictive models that automate and scale certain tasks. But let's be real: arguing over the word "AI" is a tedious semantic diversion from its tangible impact. A nuclear reactor isn't a "star," but it's still a useful descriptor for a process of nuclear fusion, and arguing the metaphor is pointless when the goal is to discuss reactor design.
Or to point it another way it's ultimately a distraction. It's the 'Chinese Room' argument. It doesn't matter if the machine understands or is intelligent. What matters is that it produces an output that is functionally useful (or dangerous) at a scale that was previously impossible. A high-speed printing press isn't "literate," but it changed the world far more than a single monk who was.
>>29310 You've just made the most compelling case for engaging with it. You cannot dismiss it as trivial and also claim it's a landscape-altering tool of oppression that threatens the US economy. It is either a trivial gimmick or a significant tool of power. It cannot be both. If you accept the latter (which you seem to, given your concern about 'fascists and chancers'), then your position of not engaging is simply an abdication - it is seeing a horde of American barbarians approaching the gates and deciding to "not care" about ballistics.
My original question—about the trade-offs between a closed, corporate model (Gemini) and an open-source one (Mistral)—is the only relevant political question in this situation. Do you prefer the "tool of oppression" to be an unaccountable black box controlled by Google, or a transparent, auditable model that can be run locally? Your non-engagement is just a vote for the former.
It's quite hard to tell, because there are so many desperately shoddy channels run entirely by humans. The output of AI isn't much worse than the shit Simon Whistler and his army of Czech writers churn out every day. Is that awkward phrasing an AI artifact, or just a mistake by someone who doesn't speak English as their first language? Is that mispronunciation a glitch in a text-to-speech algorithm, or just a symptom of someone reading a script that they don't actually understand?
I don't think the issue is as simple as "AI bad, humans good" - I think the fundamental problem is with lazy, low-effort "content" churned out purely to feed the algorithm and earn ad impressions. Human slop is scarcely better than AI slop.
>The output of AI isn't much worse than the shit Simon Whistler and his army of Czech writers churn out every day.
Simon Whistler is really quite bad at feigning competence and authority on a subject. Just pick one of his videos, any video, where you yourself know a good bit about the subject because it's one of your hobbies or you're an enthusiast. He will not only have his facts wrong, sometimes ever so slightly and sometimes glaringly, but you'll spot him faking it in every third sentence, to the point that he'll almost be blissfully unaware what shite he is talking. And I guess that is then also the thing that really irks me about him.
There was a time when most people just knew to keep their mouth shut about things they really had no grasp of.
But yeah. Just the way AI generates entire scripts and then narrates them. It can well boil my piss, and then I will usually stop watching when I notice. Because if you can't even be arsed to put work into making a video that you expect me to watch, then you're not worthy of my time.
>>29314 >You cannot dismiss it as trivial
Perhaps your reading comprehension has been outsourced to an LLM, as I did no such thing. Not playing with AI "tools" isn't an abdication of anything, as it would be a waste of time to do so. I don't need to engage directly with a flattery machine to know how these technologies are being used and abused, no more than I need to install malware on my PC to know about that.
>the trade-offs between a closed, corporate model (Gemini) and an open-source one (Mistral)
Hmm, yeah, guys, do you want to drink cold piss or hot piss?
You're such a tedious cunt. You can sit there and be smug about how right you are without getting involved to begin with, and I really wish you would, because you bring nothing to the actual discussion.
You're just those guys who stuck to "dumb phones". Right, fair enough, you don't need this technology. That's grand. The rest of the world is taking it up and using it.
>>29318 >You're such a tedious cunt.
Say that to me in the Metaverse and not online and see what happens. And it's not my fault, or problem, that the only way you can see a woman with her clothes off is by running her Instagram through a nudify app.
>You can sit there and be smug about how right you are
I've never needed your permission before, but thanks anyway.
>you bring nothing to the actual discussion.
Says the bloke arguing from analogy and calling me a cunt. Despite your histrionics I never actually had a go at you before now, I just think generative AI, at least all this consumer level stuff, is guff. You primarily seem upset because you assumed I thought generative AI was trivial in the sense that it's having no effect, which isn't true, I think it's having lots of ill-effects and a negligible number of positive ones. You've then also had a strop because I don't like something you do, which I can't imagine doing to someone on here on account of them not liking AM4 motherboards or David Cronenberg films, but I suppose that's the difference between me and you - you're a prick.
>The rest of the world is taking it up and using it.
If by that you mean the kind of person who uses cry-laughing emojis, then it's almost true. Oh, and I mustn't forget those fascists I mentioned yesterday, they love this shite. From AI assisted IDF strikes in Gaza, to whatever cruel slop Trump's social media team is putting out today - this cack is all the rage! Also, if it's so popular, why are you getting in a teary about what I think? You sound like Taylor Swift stan.
>>29317 In your previous post (>>29310) you literally grouped it with "Borderlands, Ollie Murs and meal deal savings." You are now trying to memory-hole your own argument. You did dismiss it as trivial, before you called it a tool of oppression.
You cannot hold both positions.
"Hmm, yeah, guys, do you want to drink cold piss or hot piss? I'm going to go with no piss, thanks."
This is a false choice, a luxury belief.
"No piss" is not an option on the table. This technology is being integrated into logistics, finance, surveillance, and law, whether you "play with the tools" or not. Your "non-engagement" is simply a statement that you have no interest in how the boot that is stamping on a human face is designed. My question is for people who do want to understand the boot's design, to see if a different one (open-source) is less harmful or can be defended against. Your "principled" abstention is just ceding the entire field to the "chancers" you claim to despise.
I'm not the lad you were arguing with, for clarity. I should have put NTA/NTL.
I've just been observing the thread without posting and I felt your conduct was cunt-like, and I felt you completely dismissed a couple of good and valid points in the post prior. So that's why I got involved, to call you a cunt.
>>29320 Can you point to the bit where I said it was "trivial"? You seem utterly obsessed with this point, but I never made it. Borderlands, Ollie Murs and meal deals are all fucking massive, but I don't give a monkey's about any one of them. The same is true for your AI gimmick bollocks.
I don't need to engage with generative AI. I'm not going to. I don't need to generate an LLM gf to know how shite and harmful this stuff is. I don't really care if you want to have a debate about open-source vs the massive mobilisation of capital that is clearly going to win out. I don't want to have that debate, because the tech is going to be reliant upon the same pie in the sky energy and data centre requirements, and it's going to primarily used for the same underhanded and malevolent ends, no matter what system is in place.
Consumer level AI "tools" are toys for idiots who post "😂😂😂" on Facebook. Many of the non-consumer level ones exist solely to strip away civil, economic and legal rights. Now leave me alone before I literally beat you to death with my 404 Media subscription, you massive fucking gimp.
>>29322 >the massive mobilisation of capital that is clearly going to win out.
>it's going to primarily used for the same underhanded and malevolent ends, no matter what system is in place.
You are 100% correct. This is the problem.
But this is exactly why the open-source vs. closed-source distinction is the only debate that matters. When "massive capital" (Google, OpenAI) controls the "malevolent" tools, they are opaque, unaccountable black boxes. We cannot see the training data, the weights, the hard-coded biases, or the kill-switches.
When an open-source model is released (like Mistral's), it escapes that direct capital control, even if it was funded by it initially. We can inspect it. We can audit it. We can run it locally, disconnected from their "underhanded" data-gathering infrastructure.
Your argument—that all AI is inherently bad—is a non-starter. It's here. It's already being used by the "non-consumer level" actors you're worried about. The real political question is: Do you want the "malevolent tool" to be only in the hands of the "massive capital" you despise, or do you want to arm the public with the same tool so they can understand it and defend themselves?
My question is about the only axis of control we have: transparency and auditability (open-source) vs. a corporate black box (closed). Your "principled" stance is just you proudly announcing you're too lazy to learn about the tools being used to "strip away" your rights, and you'd rather just read articles about it instead. You are a spectator.
>>29323 When did I say AI wasn't here? Saying something is shit isn't a "non-starter", it's my opinion. I know it's being used by powerful governments and businesses to make everyhing worse, I already told you that several posts ago, thicko. Thanks for dropping the obsession with that time I didn't say it was "trivial", by the way.
>Your "principled" stance is just you proudly announcing you're too lazy to learn about the tools being used to "strip away" your rights
How is not using Microsoft Co-Pilot the same as being "too lazy to learn" about AI, you twat? No one has made a good reason why I should start farming out my ideas and interests to another Silicon Valley data harvesting scheme, nor have you made an argument as to why I'd have any more joy using an open-source AI model. I don't want a chatbot that gives me spurious baking tips or bad mental health advice, it doesn't matter to me if you tell me it's open-source or not. It's like I'm telling you I don't want a car, but you keep saying "but this one's electric!", and it's like, I don't want to pay for it, or the insurance, and I live by the train station - I don't need it!
>You are a spectator.
Oh, I see. And you have a seat at the table because... ? What, exactly? You're just another AI hypeman, whose claims of the inevitability of AI ubiquity should be viewed extremely sceptically, given the money set to be made (not to mention lost), and the failure of AI to find a real use case in the majority of people's lives.
Hey, here's an idea for you pair of geniuses, why don't you, for the first time ever in this thread, tell us all what you actually do with AI all day? Do remember this is a SFW board, mind.
>>29324 "tell us all what you actually do with AI all day?"
You've spent three posts assuming that I'm asking for baking tips. I'm not. My primary use case is complex document analysis - "verify key assumptions in this 300-page report" or "find the 10 weakest points in this proposal." It's an analytical force-multiplier, not a toy.
This is why your car analogy is wrong. You're saying "I don't need a car, I live by the station." I'm saying "I have to move 10 tonnes of material, and I'm asking about the difference between a secure, private lorry (Mistral) and a corporate-owned one that spies on my cargo (Gemini)."
My original question—which you've consistently failed to grasp—is about how to escape the "Silicon Valley data harvesting" you're (rightly) worried about. I need the capability (the lorry) without the surveillance (the corporate driver).
You don't "need it" because you don't have 10 tonnes to move. Fine. This question isn't for you. Best get off to bed.
>>29326 The original question was "Which LLMs are you lot using these days?" Then you, or someone like you, had a pop at two posters who weren't me who said "we don't".
>Best get off to bed.
It speaks to the kind of man-baby you are that you think bragging about not having a bedtime is some kind of own.
And, as I suspected, you are abandoning your critical thinking to the chatbot.
You aren't escaping anything, you're already in it.
>>29326 >My original question—which you've consistently failed to grasp—is about how to escape the "Silicon Valley data harvesting" you're (rightly) worried about.
I'm someone new. I might have posted before to shit on AI, but I haven't posted in this thread for a while. The great thing about open-source AI is the same as open-source everything else: you can put it out there and pretty much just leave it there. The private interests are going to try to suppress it, because they want to get rich off it and if all the hype is true, someone is guaranteed to become the richest human in history when they license their AI model to the entire planet. They might make better AI models to justify charging people for it. But it is inevitable that even if some AI model appeared that was perfect at everything came on the scene, it would get enshittified after a couple of years and people would want to have the open-source alternative. Desktop Linux is a pile of shit compared to Windows, but it's functional and some people love it, and it's not going anywhere despite its negligible market share. Open-source AI models will be the same. And when Claude or Gemini or whatever has its Windows 11 moment, and everyone turns against it, the open-source alternative will still be there, ready and waiting to step into the gap. So your idea that it's one or the other is fundamentally flawed.
Also, >>29327 is right - this was never your original question.
>you are abandoning your critical thinking to the chatbot.
I stated my use case in two kinds of adversarial analysis. It is the opposite of "abandoning critical thinking." I am not asking it "is this good?" or "what should I think?" I am using it as a tireless opponent to stress-test my own critical thinking and find flaws I might have missed.
You are unable to grasp this because you are stuck on the "baking tips" consumer model.
>You aren't escaping anything, you're already in it.
You've finally understood. This is my entire point. We are all "already in it." This is why your no piss / I don't need a car stance is an abdication. Since we are in it, the only relevant question is how we engage.
Your method: Non-engagement (i.e., being a passive victim) and reading 404 Media articles about what the boot looks like.
My method: Actively finding the safest tool (Mistral, the private lorry) to counter the corporate tool (Gemini, the spy lorry) and understand the threat.
You have no answer for my lorry analogy. You ignored it completely and resorted to insults about bedtime because you've lost the argument.
>>29326 >My primary use case is complex document analysis - "verify key assumptions in this 300-page report"
Yes, because the thing you want to use to "verify key assumptions" is a black box that lies and hallucinates like a fucking crack addict.
>>29328 >Desktop Linux is a pile of shit compared to Windows, but it's functional... Open-source AI models will be the same.
This is a good analogy, and you are correct about the long-term political economy. The open-source alternative will always persist as a 'Linux' option.
>So your idea that it's one or the other is fundamentally flawed.
This is where you are wrong. You are conflating a long-term, macro argument (the market will have both) with my immediate, micro question (which one can I use today). My original question stated Mistral has less features. My lorry analogy clarified that I have a high-capability task.
To use your analogy: Linux is functional for basic web browsing. It is not functional if my job requires Adobe Premiere or specific CAD software. That is my problem. The open-source Linux option does not yet have the Adobe Premiere capability that the corporate Windows option does.
So for my specific task, right now, it is a binary choice:
Use the 'spy lorry' (Gemini) and get the job done.
Use the 'private lorry' (Mistral) and fail at the task.
My question is for others who are also trying to solve this practical, present-day capability gap.
>this was never your original question.
This is tiresome pedantry. The motive for the question (the trade-off) was explicit in the post.
>>29331 >The motive for the question (the trade-off) was explicit in the post.
I just scrolled up to see it again in case I missed anything but it's really fucking not. Why not post this:
>Okay lads, that's what I intended to ask but given both the other people on this website aren't seeing it, maybe I didn't express myself clearly. Sorry. Can we address the question I did intend, as it has the potential to be more interesting, now?
The capability gap between closed and open models isn't growing and probably won't grow by much in the foreseeable future, due to an interesting quirk in training dynamics.
Both Anthropic and OpenAI use a cascading training strategy - they train a huge model that's very expensive to run, then they use the outputs of that model to train a succession of smaller models. The big model does most of the hard work of making sense of everything everyone has ever written, while the smaller model has the comparatively easier task of learning to imitate the big model. Claude Opus costs $75/MTok versus $6/MTok for Sonnet; GPT-5 Pro costs $120/MTok, versus $10/MTok for vanilla GPT-5 and $2/MTok for GPT-5 mini.
The leading open-weights Chinese models just skip the first step and go straight to the second. Deepseek is mostly trained on the outputs of ChatGPT, while Qwen3 is heavily trained on Gemini. If you ask Deepseek to tell you about itself, it'll often say that it's ChatGPT and that it was developed by OpenAI.
Unless something fundamental changes in how LLM training works, there's really nothing the developers of big proprietary models can do to stop other people from copying their homework. Anyone who can use a model can use it to train their own model, distilling >95% of its ability at ~5% of the cost. Open models will never catch up with the leading closed models, but they'll never lag too far behind either.
What the fuck, why do they have to deliberately try make it look like something out of a Black Mirror episode where the thing goes rogue and murders a family in their sleep?
>>29335 The whole point seems to be that it does chores but nobody who can afford to spend $20,000 on a robot is doing their own chores anyway. This is just a very expensive way to fire the help.
>Has a 55lb carrying capacity so it can do any of your chores
That's not even 25 kilos what the fuck is the point of a robot that can't carry as much as I can?
>>29336 It's definitely somewhere in the uncanny valley of "Is this real or just a really involved promo for an upcoming TV series?"
>Neo will keep improving
You're going to do a rugpull or go bust in 2-3 years and the robot will stop functioning, leaving everyone with a very expensive, disproportioned mannequin.
>>29336 Because that's what science fiction has conditioned us to see in a humanoid robot.
>>29339 >You're going to do a rugpull or go bust in 2-3 years and the robot will stop functioning, leaving everyone with a very expensive, disproportioned mannequin.
I think the real business plan is to be bought out once they generate sufficient interest and credibility. These things are closer than ever before but it's the price point, refinement and simple public acceptance that are now the barriers - which no start-up can provide.
Obviously the real solution is to develop adult features but we're not allowed that so China will win the robot race by default.
>>29340 >I think the real business plan is to be bought out once they generate sufficient interest and credibility.
On the one hand, I think it's bold of them to assume they'll ever get sufficient credibility. On the other, tech investors, particularly in the US, seem to be almost childishly credulous, so this company will somehow generate an 11-figure valuation on the back of literally fucking nothing like just about every other tech company out there.
I predict two possible outcomes. Either:
* after 2-3 years they'll be rugging; or
* they'll sell, and after 3-4 years the new owners will do the rugging.
(Functionally, from a user perspective, going bust without a succession plan looks identical to a rugpull, and probably has the same motives anyway.)
The whole play is about training data. When the built-in AI on Blanky McBlankface can't work out how to do something, a remote operator takes over via a VR headset. That's obviously quite a silly thing to do if you're trying to turn a profit, but it's an incredibly productive way of generating huge amounts of training data for robot AI. You only get (and only have to pay for) data on tasks that the AI can't do yet, which also makes the training runs more computationally efficient.
It's how Waymo got to fully autonomous drivers - they had human drivers in their cars for years who did most of the driving at first, but gradually did less and less as the AI learned from them until they weren't needed at all.
>>29342 I went looking for how many fatalities they've caused (just to see how they compare to Tesla's impressive body count), but got distracted by this absolute piece of nonsense:
One interesting thing that came up recently was OpenAIs stats on how many users show mental health emergencies:
>OpenAI has released new estimates of the number of ChatGPT users who exhibit possible signs of mental health emergencies, including mania, psychosis or suicidal thoughts.
>The company said that around 0.07% of ChatGPT users active in a given week exhibited such signs, adding that its artificial intelligence (AI) chatbot recognizes and responds to these sensitive conversations. While OpenAI maintains these cases are "extremely rare," critics said even a small percentage may amount to hundreds of thousands of people, as ChatGPT recently reached 800 million weekly active users, per boss Sam Altman.
https://www.bbc.co.uk/news/articles/c5yd90g0q43o
The direction the reporting took was that these issues came from the chatbots but I'm wondering if instead we're at last getting real data on how much of the population will be undergoing or heading towards a major mental crisis at any given time. As we're internet people I'm sure we'll agree that this is likely underreporting but given how one of the biggest use-cases has become therapy it's probably getting close to a pulse-check on the mental state of the population.
There was an interesting comparison I listened to recently that touched on parallels with how social media developed and it made me realise how things are playing out:
>>29341 The people buying them out will be Apple and Alphabet and their interest will be in stripping out the start-ups talent and IP.
The buyers will either get a cool piece of history or more likely their money back.
Assuming they don't somehow pull off hyper scaling anyway.
>>29346 >The buyers will either get a cool piece of history or more likely their money back.
What, like when Revolv was acquired by Google or Pebble was acquired by Fitbit, or with AI Pin acquired by HP? The corporations happily refund all the users out of the goodness of their corporate hearts? What planet do you live on?
None of these are 'cool pieces of history', incidentally. They're less notable than Furbies.
It's sad that people with mental health issues think they have nowhere else to turn to but AI with their problems. As somebody who has struggled with mental health for most of my life, not just my own mental health but also that of close family, I can say that those people are often in dire need of an actual human person to talk to but just feel that there is a barrier that keeps them from seeking actual professional help.
There needs to be much more done still to, well, normalise the idea that when you are mentally unwell, you go see a specialist, same as when you've got a broken foot or wrist.
I would speculate that many of the mental health emergencies also occur in countries with less widespread access to those services, and that probably includes the U.S., with the shit kind of healthcare system they've got there, where there is an additional financial barrier to getting proper help.
Then again, another way of looking at it is that while it's worrying that they have to turn to AI, at least they have that straw to clutch.
Two things strike me about this, in a philiosophiocal sense really, about the nature of mental illness.
For the first, it often seems to me when I've been at my lowest ebb, that it's not just the obstacle of obtaining help and having "someone to talk to", it's the idea that it doesn't really matter if it's not somebody who actually cares. What I really want in those moments is a friend, and even if they do care in a way, you know deep down that a mental health professional is just doing their job. It's like paying a prostitute to experience the feeling of love- You know it's not real.
So in that sense, what's really the difference between a therapist and an AI? Sure a lot of the best qualified human ones will be better, in an objective sense, that the random output of a computer algorithm. But compared to the average mediocre NHS CBT services, it's a toss up.
Secondly, that brings me to the roundabout question if avast majority of people's mental health troubles nowadays are not even really deeply rooted mechanical, inherent illnesses at all. We are categorising them wrongly. They are deficiencies. They are more like a spiritual form of scurvy. A person who isn't getting all of the correct social nutrients comes down with a horrible case of going fucking mental, and the prescription should be an intensive course of mixing them together with other lonely mental cases in their area and getting them to add each other's number on WhatsApp.
Depression is not the same thing as being miserable. Anxiety is not the same thing as being worried. The modern fad of "mental health awareness" has pathologised normal emotions, while further stigmatising people with severe mental illness.
Lots of things in life are shit. It is entirely normal and healthy to feel miserable if shit things are happening to you. It is entirely normal to feel worried and stressed when you're going through major changes or your life has become very uncertain. Those feelings only become the symptom of an illness when they're abnormal - if you become trapped in a pattern of thinking that becomes self-reinforcing, or if you're so overwhelmed that you can't function.
Loneliness is not a mental illness and no drug or therapy can substitute for real relationships. That's not to diminish loneliness, because it does real damage to people's wellbeing and physical health, but it isn't an illness in itself. A minority of lonely people will need professional help to fix their loneliness - if they're genuinely too frightened to go out and meet people, if they're so miserable that they can't summon the energy to try, or if they still feel lonely in a room full of friends.
The majority of lonely people just don't have any symptoms that psychiatry can treat. They don't panic whenever someone talks to them, they don't have any profound deficits in social skills, they aren't so deeply depressed that nothing in the world can cheer them up. There's no underlying pathology that is perpetuating their misery; they just aren't doing the things that might be uncomfortable in the short term, but that they know would make them happer in the long run.
This majority just needs an arm around the shoulder or a boot up the arse. They know why they are unhappy, they know what would make them happier, but they just can't be bothered. That's human nature - we all know that we should eat more veg and get more exercise, but most of us don't. By the same token, if you went to your doctor because you're very unfit, there's really not much he can do for you other than encourage you to go for a brisk walk or suggest a fitness class. There's no medical treatment that would help you, and diagnosing you with an illness would just make you feel more helpless.
It's great that a lot of people feel more able to talk about their feelings these days, but it's a complete disaster that they increasingly tend to think of their feelings as symptoms. Being unhappy about tangible problems in your life doesn't mean you're ill, it means you're human. Negative emotions are an evolved response meant to spur us into action, not a defect that needs to be repaired.
Sorry, I must have phrased things badly. In an attempt to clarify:
Most people who go to their GP about their mental health are suffering from problems of living, not mental illness. There is nothing wrong with their brain or their mind - their problems are real, tangible and external. They will not benefit from psychotherapy or drug treatment, because they aren't suffering from any disease. They have real problems that cause real suffering and they really do need help, but their problems aren't medical and a GP or psychiatrist can do very little to help them.
We have encouraged people to think of problems of living as being signs of illness, through "mental health awareness" that "de-stigmatises" mental illness by redefining ordinary difficulties as symptoms of disease.
This is bad for people who aren't mentally ill, because it reduces their sense of agency and gives some of them false hope that a course of tablets or CBT might fix the fact that they hate their job or their marriage is falling apart or they've got no mates. It's also bad for mentally ill people, because redefining mental illness to include perfectly normal experiences further stigmatises people who have experiences that are very much not normal: psychosis, mania, thought disorders, severe self-neglect and so on.
We're trying to mend the holes in a fraying society - problems like atomisation, alienation, lack of community, insecure work, a lack of hope - with inappropriate, ineffective and counterproductive medical interventions. Psychiatry can only treat problems in the mind, not problems in the world. People who are reasonably and rationally miserable about tangible problems in their lives need practical support to resolve those problems, not medical treatment. The only people who benefit from this falsehood are the government, who get an excuse for not fixing things that are making people miserable.
It comes off like you are trying to disagree with me when overall you agree with the thrust of what I was getting at. I can't be bothered to fully explain how right now because I am on my phone at work, but if I remember by tonight I will probably try.
