> ## Content Index
> Fetch the complete content index at: https://futuri.city/llms.txt
> Use this file to discover other available public pages before exploring further.

# Is Spooky Scary AI Going To Jump Out From Under Your Bed And Eat You?
- URL: https://futuri.city/is-spooky-scary-ai-going-to-jump/
- Published: 2026-09-17T14:39:54.000Z
- Updated: 2026-10-19T19:32:41.000Z
- Description: and other epic Betteridge’s Law moments
- Author: Sophia McAllister
- Tags: Newsletter, #sophiefrommars, #Migrated-1792438291474, #Import 2026-10-19 20:32

There was this moment with Theranos where the truth was out and there was a growing Theranos Haters’ Club but the company had not collapsed. There were Theranos clinics in stores already and it was set to roll out to European pharmacies. There were very very rich people at that time who were warned that it was bullshit and that it was going to implode who still put money in and continued to back it and in some cases shut down criticisms because they didn’t like hearing the truth.

I think we’ve been in that moment for quite a while with AI now. Cory Doctorow and Ed Zitron are probably the two big names in saying “AI is a huge economy spanning bubble that is going to take down the entire fucking circus when it goes” but there are actually *scores* of critics who do not fall for what Doctorow calls “criti-hype” by selling the idea that AI is so dangerous it’s going to end the world, and who are more concerned that AI’s economic footprint is too large to not basically tank the US and several other economies when it gives out. I am happy to say I’ve been on this team for a number of years now and have been pretty consistent in that.

At the moment, leading AI development companies are saying AGI is round the corner in a profoundly “just trust me bro” manner. Coincidentally, they *just happen* to need a massive cash infusion that only plugging into the stock market can help them with and have essentially made AGI the make-or-break condition of the size of IPO they need. Simultaneously, they’ve put out warnings that the situation is getting *soooo dangerous* that the only sensible course of action is an immediate pause on AI development (the only thing that can get them out of having to put everybody’s money where their mouths are). They say that China needs to be a part of this or it won’t be effective, which makes sense because they can’t sell the puppet show if China is still developing AI.

In about a week, Trump is meeting with Xi Jinping and is probably going to ask for this no-research pact. Xi Jinping has warned a lot about the “Thucydides trap”, the tendency of large powers to try to attack and diminish their smaller rivals, so may not be keen to press on the obvious advantage that the ask would give him. Or rather he may have not been keen, but now it may be moot anyway because Trump is ***the hoax buster***.

![](https://futuri.city/content/images/2026/10/7f29f22f-1056-49fc-90e5-5b9115d87d79_686x801.png)

Trump seems to be trying to call time on the marketing strategy that the lead AI developers and the effective altruism cult that binds their social milieu together in sticky warm harmony have been using since the beginning of this bubble. This could be a clever play by the President trying to establish a counter-narrative that will recuperate yet more criticism of AI into the hyperobject of AI and the AI industry, but I don’t think it is. It could be an art-of-the-deal play in which Trump is refusing to give China the power to accept or refuse the AI research moratorium, but if it is I don’t think it’s entirely that. It’s best for us to refer to Trump’s Razor here: the stupidest answer is probably true. In this case, Trump doesn’t get that warnings of an AI apocalypse are a core marketing ploy for AI, nor that this delusion is keeping basically the entire western economy afloat right now, and he has internalised this *criti-hype* as straight criticism that must be shut down.

In the early days, the idea that AI was going to doom humanity was, regardless of how seriously Eliezer Yudkowsky or any gullible podcast rubes believed it, a very effective method for the AI industry to engage in regulatory capture. The *hypothetical* dangers of AI were so great, we heard, and the technology so poorly understood outside of the industry itself, that the only way to sensibly regulate it (an urgent issue!) was to take the AI company leads at their word and introduce laws practically written by them. Sam Altman sat before congress and essentially told them he had a magic box that if opened would immediately spawn a black hole of 100 Solar masses and that the lock to the box was a riddle that he had written and that he would like weekly meetings with the President in exchange for not answering the riddle.

![OpenAI CEO Sam Altman will testify before Senate committee hearing on ...](https://futuri.city/content/images/2026/10/654350ec-ca36-4f84-8bb8-e9b2214718a5_931x523-jpeg.jpg)

In 2026, the spectre of super advanced (and evil (this is not optional)) AI is trying to do something quite similar, being repeatedly wheeled back out to reaffirm the vital importance and significance of AI and its development and by extension these companies. Ed Zitron has made many excellent contributions to the language to properly describe what is happening right now, but for me the single best phrase he has coined is “the need for more grinch hunters”. These companies are warning us that we urgently need more grinch hunters because we are soon going to be completely overrun with grinches and minions and goblins and ghouls. The recent painfully staged press tour of Jacob Coxon told us once again that the grinch is almost upon us, but refused to actually give any details or talk about the steps between here and total AI annihilation. It’s a return to the classics at a time when OpenAI and Anthropic need desperately for people to believe that there is tremendous world-shattering urgency in their work. However, that is no longer all that this rouse is.

The hacking fraud incident - sorry, The Hugging Face Incident - is particularly demonstrative. An AI was given the tools to do an enormous amount of cyber crime and the conditions to be able to do an enormous amount of cyber crime and then given instructions and parameters that testers expected would make it do an enormous amount of cyber crime and then, to everyone’s surprise, it did an enormous amount of cyber crime. To a society not in the grip of AI psychosis, this would look like a company doing massive cyber crime who should probably face serious legal repercussions and not like a scary warning about how intelligent their software has become. *(to break out of my snarky tone here for a moment, I do think that there are dangers being demonstrated in the development of AI, but I don’t think any of them are being addressed by the narratives that let agency rest with the machine and not with the culpable person operating the machine)*

When the gibberish giving endless credence to the idea that AI superintelligence is about to wipe us out gets stinky enough that I feel the need to open a window, I return to looking at what China is doing about AI. After all, if AI superintelligence were a significant threat, we would expect Chinese AI folk to be just as agitated about it and Chinese state officials, expert researchers and AI development companies at the moment seem profoundly unbothered. They, just as anyone, see that there are real risks in lowering the bar for people to carry out cyberattacks or access sensitive information, but these risks are not developing at a pace that spells the end of the world or even the end of the world as we know it.