Buckle up, me old muckers, it’s blog time and this time it’s about AI and how bad it is when used by pretty much anyone.

There is a constant push these days to use AI, and it’s getting jammed into everything because the corporations have spent so much money on this particular bubble (apparently around $1.6 trillion as of early 2026) that they’re desperate to get us all to adopt it so they can start charging us properly for it. We at MycoPunks are not fans, because for the few little good things it has done, it’s like watching something that doesn’t think try and replicate good ol’ fashioned human creativity, because that’s exactly what it is. Soulless artistic output, laden with errors, and used by people that don't understand that it's just mashing pre-existing data together in a way that looks good enough.

AI is a fuckin’ plague on any creative endeavour, for sure, but it’s even less desirable when it’s applied to something niche but science-based, like mycology. We saw this mad bit the other day from another company where they’d published a post about Lemon Tek, they’d clearly made it using AI and it literally wasn’t a Lemon Tek recipe. The AI had checked what ingredients combine most frequently with lemon in some kind of drink and produced some ginger/lemon hot tea sort of beverage. Now, this just means that whoever posted it knows literally nothing about the world of mushrooms, because it was so obviously wrong that if you’d even spent a month looking into your hobby you’d know (we ain't gonna name names because that feels too mean given how dumb this is). Obviously in this case it’s harmless, stupid, and to be quite frank, the online equivalent of not realising you’ve left the house without your trousers on, but it could absolutely be dangerous as fuck when we’re dealing with mushrooms. 

They’ve pulled a marketing mastermove by calling it “Artificial Intelligence”, because everyone’s seen films with the thinking robots in, so you have a pop-cultural touchstone to inform yourself with and feel familiar with the concept. To briefly summarise what the process is, you ask the clanker to write you something. The clanker then looks at all the words it has stolen from all the people that have written them before, and jams the ones that appear to meet the context into an order that matches the style of the works it has stolen. This could be useful to you, if you already knew enough to check its work, and there was enough pre-written stuff for it to be able to apply the right words in the right fashion, but that’s not really the case in mycology. The "AI" doesn’t know what the words are, or the meaning, it’s just matching patterns together and hoping it’s close enough. There aren’t any thought processes going on, it doesn’t understand what the words are or what they mean, it’s just copying patterns, like a horse that clops the correct answers to basic sums when you cough at it.

This isn’t a very sensible idea when it comes to mushrooms, because whilst some of them can make you see God, some of them can also send you directly to him. It’s too niche a hobby, there hasn’t been enough written about it for the fuckin’ fancy predictive text machine to be able to get things right, and when it’s related to a field where the wrong chomp can end your existence it seems a little irresponsible, to say the least. The way AI stuff is presented is that it looks like it is written with confidence, and is authoritative. This isn’t the case, but will it fool people who aren’t used to critically thinking about their sources of information? Yeah. Would you entrust your life to that little bitch Clippy? Didn't think so.

There’s also the environmental impact - most mycologists give somewhat of a fuck about the environment, on some level: learning about how everything is so interconnected tends to do that to you. We’re not saying this hobby turns you into a paragon of virtue, but it definitely makes you think a bit more about the environment, and the complex interplay between everything that’s part of it because that is what the hobby is, even at the simplest level of growing some wood-lovers. 

AI uses up so many resources that it’s just mullering through fresh water and releasing shitloads of carbon dioxide into the air, constantly - ChatGPT alone handles over a billion queries a day (this means “you typed something into the box”). Trying to find the amount of resources this uses up varies massively, as the pro-AI people give a low estimate of around 2g per query, whereas the anti-AI people go higher at around 4.3g being commonly cited online. Let’s split the difference and call it 3g - that’ll be 3 billion grams of carbon dioxide, so that’s 3000 tonnes of carbon dioxide, a day, from ChatGPT alone, that didn’t really need to be let into the already-pretty-fucked-world. Elon Musk, the world’s richest facsimile of a human, is busy probably giving everyone near his AI data centre cancer, because human lives just don’t matter when you’re a sentient pile of money steering an empty husk around the place.

And obviously because it’s something all the billionaires have got a massive boner about (lots of jobs to be cut, I suppose, the sacred number must continue to go up) it is getting jammed into fucking everything, normal searches, your phones, tellies, all sorts. A solution in search of a problem, it’s now almost harder to avoid than it is to use. Imagine spending $1.6 trillion on something beneficial for the planet instead, being known as the person or company or groups of companies who forever changed the course of humanity for the better - hang on, fuck that, what if we just carry on trying to be the human with the most money? That’s surely loads better. If Captain Planet was real the only powers he’d have left at this point would be a drinking problem and PTSD due to all the billionaires he’d thrown off buildings.

Hilariously, this could all peter out if things carry on the way they are. Some of the pro-AI lobby are reporting that almost all available human-generated data has already been used to train the machines on. This means that AI may well just start eating itself and forming responses based on previous AI responses that have the usual errors in, and then it’ll keep building on this absolute tower of idiocy until it just doesn’t work for this purpose any more. Decent human content is needed constantly in order to keep it from cannibalising it’s own mistakes, but there’s no way we can generate enough to make even a drop of difference so maybe - just maybe - it’ll all fuck off like 3D TVs and “The Metaverse”. Remember The Metaverse? Nah, thought not.

We don’t hate technology, we’re mostly huge nerds around here, and if there was an actual decent use case for this sort of thing then maybe there would be a way to justify it - aeroplanes are good, mobile phones are good, tellies are good, and they’re all inevitably entrenched in exploitation due to the capitalist hellscape we live under yet we see the purpose in these items. There are absolutely some useful areas for the kind of pattern recognition that AI is, but generating a picture of Goku fucking a Minion in the ass is something you could pay a human to do if you really needed to see it happen that much, and you’d put less wear and tear on the planet. Maybe that’s our new side hustle? Might be worth a shot...