> ## Content Index > Fetch the complete content index at: https://spyglass.org/llms.txt > Use this file to discover other available public pages before exploring further. # AI's "No Fate" Future - URL: https://spyglass.org/ai-end-of-humanity/ - Published: 2026-09-15T12:26:47.000Z - Updated: 2026-09-15T12:31:40.000Z - Description: Is AI about to end humanity? No one knows – most notably those proclaiming such things... - Author: M.G. Siegler - Tags: ai, rsi, agi, openai, anthropic, deepmind, google, demis hassabis, sam altman, dario amodei, bill gates, yann lecun, tech, donald trump, politics It was the best of times, it was the end of times... We can't even make it a few hours now, let alone an entire day without [someone](https://www.bloomberg.com/news/articles/2026-09-15/google-deepmind-staffer-says-ai-may-kill-us-all-in-exit-post?ref=spyglass.org) at one of the leading AI labs saying that what [they're building](https://x.com/dkokotajlo/status/2099600298855829616?ref=spyglass.org) is going to kill us all. Not just be bad for the environment in some way. Not just be bad for society in some way. But *kill* us. It's to the point where this is no longer just on fringe forums or in dinner discussions in the Bay Area. This fear is going mainstream. My family is constantly pinging me about it. It has [hopped](https://x.com/mgsiegler/status/2099449661274698080?ref=spyglass.org) across the pond. The President of the United States is resorting to [staged call-ins at tech conferences](https://www.nytimes.com/2026/09/14/technology/trump-ai-safety-jensen-huang.html?unlocked%5Farticle%5Fcode=1.BFE.6RVR.cSzQJON-bJf4&smid=url-share&ref=spyglass.org) to try to dispel such notions and calm such fears. That ends up doing the opposite, of course. The good news for the AI companies is that this is making the data center [debates](https://spyglass.org/ai-perception-problems/) look completely [trivial](https://spyglass.org/ai-steve-jobs/). The bad news is, well, [everything else](https://www.techmeme.com/260915/p6?ref=spyglass.org#a260915p6). So what's up? Are we toast? I mean, maybe! But also maybe not. As an outside observer in all of this who simply conjures the dark arts of triangulation and adheres to the principle of [Occam's razor](https://en.wikipedia.org/wiki/Occam%27s%5Frazor?ref=spyglass.org) whenever possible – in other words, someone who tries to be rational, *but decidedly* *not a* [*Rationalist*](https://en.wikipedia.org/wiki/Rationalist%5Fcommunity?ref=spyglass.org) – my simple, obvious read is that no one actually knows. That doesn't sound comforting, but also it perhaps *should* if all of these reports are freaking you the fuck out. I'm not telling you not to trust all these literal doomsday accounts from these people within the labs because they are largely from more junior, less tenured people, but I am telling you to look at their own biases. Just quick, cursory scans often reveal the clear and obvious ties to said Rationalist movement. While [Effective Altruists](https://en.wikipedia.org/wiki/Effective%5Faltruism?ref=spyglass.org) often get thrown under the AI doomer bus, as [Cal Newport succinctly outlined in an essay last week](https://www.nytimes.com/2026/09/05/opinion/ai-silicon-valley.html?ref=spyglass.org), it's actually Rationalists (by way of [Extropians](https://en.wikipedia.org/wiki/Extropianism?ref=spyglass.org)) that planted these end-of-times seeds a couple decades ago. Unsurprisingly, a lot of people now working in the AI field are adherents. They self-selected to do this work, undoubtedly believing they could help protect the world from this demise. But as we [creep closer to AGI](https://spyglass.org/agi-2026/) – but really, [RSI](https://en.wikipedia.org/wiki/Recursive%5Fself-improvement?ref=spyglass.org) seems to be the [particularly](https://spyglass.org/inklings-ai-overkill-anthropic-walks-back-siri-ai-just-works-elons-diamond-hands-apple-tv-is-hbo-openais-rsi-clause-siri-is-not-dtf/#:~:text=M.G.%20Siegler-,I%20Quote...,-%22The%20faster%20the) [problematic](https://spyglass.org/thoughts-on-back-to-the-frontier-ai-races-netflix-out-youtubes-youtube/#:~:text=for%20thoughts%20on...-,Back%20to%20the%20Frontier%20AI%20Races,-OpenAI%20Technique%20in) acronym now (fears of which also [date back long ago](https://x.com/davidchalmers42/status/2099528302398447742?ref=spyglass.org), well before the Rationalists) – a lot of these people clearly believe the end, which again, they've long believed in, is near. Again, [I'm not saying they're wrong](https://spyglass.org/what-if-anthropic-is-right/). I'm saying that I don't believe they actually know. And I'm saying to just keep in mind their biases in making such proclamations. And I also might point out that as this rhetoric ramps, there are undoubtedly people who are trying to take advantage of the literal FUD – on both sides. Anyway, one level up from these literal doomsayers, we have executives at the various AI companies not quite echoing these sentiments, but also not completely shooting them down. [This is the real problem](https://spyglass.org/thoughts-on-ais-maniacal-mixed-messages-siri-ais-potential-developer-problem/#:~:text=Please%20Get%20Excited%20for%20the%20End%20of%20the%20World) in all of this because it does seem to lend credence and gravitas to the arguments. You might [expect this](https://spyglass.org/anthropic-becomes-the-villain/) from Dario Amodei given [his background](https://spyglass.org/anthropic-vs-administration/) and [known beliefs](https://spyglass.org/ai-nuclear-weapons/), but Sam Altman – his mortal enemy – also now [calling for a slow-down](https://spyglass.org/ai-frontier-race/) certainly raises eyebrows, if not fears. Is that simply because a slowdown would also be in the interest of the leading AI labs such as OpenAI? I mean, you'd hope that take is too cynical, but it can't be completely discounted either. I still mainly chalk it up to no one knowing what is going to happen as we approach RSI. If we sort of [accidentally trip into it](https://spyglass.org/claude-mythos-quandry/), it very well could be bad! It could also be fine. But again, is that worth the risk? So when even more respected leaders in the field, like Demis Hassabis, are also pushing for a pause, I believe it's an acknowledgement that no one knows where exactly this is heading. I mean, how could anyone? Even [Amodei basically admits this rationale in his most recent essay](https://darioamodei.com/post/we-must-pace-the-frontier?ref=spyglass.org) calling for the 'pacing'. Others, like Yann LeCun, think [all of this is "bullshit"](https://x.com/ylecun/status/2043673634363851250?ref=spyglass.org). But he has his own biases too! In this case, [against LLMs](https://spyglass.org/meta-ai-world-models-lecun-agi/). He has long believed, like many others, that while [LLMs are useful](https://x.com/ylecun/status/2099806809431081258?ref=spyglass.org), they're ultimately [too rudimentary to get us to AGI](https://spyglass.org/llms-vs-world-models-ai/). And if that's the case, you certainly wouldn't believe that their ability to mimic human speech will bring about the end of the world. But again, as illustrious as [his background](https://en.wikipedia.org/wiki/Yann%5FLeCun?ref=spyglass.org) may be as one of the "Godfathers of AI", the reality is also that he too doesn't know for sure. Nobody knows for sure. We're in uncharted waters here, and I think even the LLM skeptics would have to say that the technology has taken us far past what many originally thought possible. But, but, but Bill Gates! Yes, [is there anything about his background](https://spyglass.org/thoughts-on-bill-gates-doomer-the-mac-mini-morphs/#:~:text=Bill%20Gates%2C%20AI%20Doomer) that makes him particularly suited to predict this future? I'm sorry, but no. He too might be right, but it's not because of some secret knowledge he has in this field. So yes, I'm saying there's a chance. But also a chance that there's no chance. Which isn't particularly helpful, except perhaps in calming nerves somewhat from the folks constantly saying this is the end of times. There is nearly an infinite range of outcomes here between being fine and being dead. Also, many of them – probably half! – have outcomes which end up being good, if not *great*, for society. And that's how many of these folks obviously rationalize – I hate to keep using variations of such a loaded word here, but it's also fitting – continuing work in this field. If the end of humanity is even just possible, wouldn't you quit immediately?[1](#one) Yes, unless perhaps you thought that you could save the world with your work. So that's the obvious tension here. That, alongside the other elephant in the room: China. This fear that if we even just pause, let alone stop work on AI, China will zoom ahead. Obviously this is what fuels President Trump's rhetoric and annoyance here. And that too seems rational! If China lacks the Rationalists undercurrent in the field, they're likely to keep pushing forward unless they determine [their own risks](https://www.nytimes.com/2026/09/14/world/asia/china-ai-security-risks-anthropic.html?ref=spyglass.org). There's [some inkling](https://www.reuters.com/legal/litigation/how-china-is-preparing-risk-ai-escaping-human-control-2026-09-14/?ref=spyglass.org) of that happening, but it also [seems less about](https://www.nbcnews.com/world/china/china-ai-slowdown-trump-amodei-altman-threat-cold-war-rcna597631?ref=spyglass.org) the end of the world and more one level of rhetoric down, on the [catastrophic](https://spyglass.org/claude-mythos-quandry/) security front. So where do we go from here? I don't know because again, nobody knows. Certainly that alone could and perhaps should put me in the cautious camp, pointing to a "pacing" of the technology. But I'm also trying to think realistically here. Not just on the China fears, but also around the fact that many of these companies are [racing towards IPOs](https://spyglass.org/ai-ipo-race/). The two things don't have to be mutually exclusive, but they're also [inherently tied together](https://spyglass.org/thoughts-on-openais-potential-comeback-hollywoods-movie-tickets-sold-disaster/#:~:text=Is%20OpenAI%20Set%20for%20a%20Comeback%3F) in ways. What I'm saying is that the incentives here are all, yes, *misaligned*. At the very least what we have right now is [a full-blown communication crisis](https://spyglass.org/thoughts-on-ais-maniacal-mixed-messages-siri-ais-potential-developer-problem/#:~:text=Please%20Get%20Excited%20for%20the%20End%20of%20the%20World). There's a line between whistle-blowing and pulling the fire alarm in a movie theater, and I think we're now testing it. I, of course, think you should be able to voice your opinion that you think AI is going to destroy the world. But I'm not sure you should be employed at a frontier AI lab if you feel the need to project such opinions outward in ways that are not constructive and, in fact, may end up *destructive*. I'm sure these people think it's adding fuel to the pacing case – if not an outright *ban* on AI – but it could ultimately backfire. The reality remains that until [Kyle Reese](https://en.wikipedia.org/wiki/Kyle%5FReese?ref=spyglass.org) – or the [T-800](https://en.wikipedia.org/wiki/Terminator%5F%28character%29?ref=spyglass.org) – comes back in time, no one knows the future. 👇 ****Previously, on** ***Spyglass** **...** [AI’s Mixed Messages • Siri AI’s Potential Developer ProblemDo you feel like every other headline and talking point coming out of the companies at the AI frontier contradict each other? You’re not alone there. The future is so bright... because of the nuclear blast that the rogue AI agents will trigger. By accident or on purpose. But![](https://storage.ghost.io/c/af/ca/afcaa655-46e2-45b8-889a-2881de5cce69/content/images/icon/Spyglass-Rings-----Multi-9716c2d4-cacd-40fe-8b46-569cebd120cb.png)SpyglassM.G. Siegler![](https://storage.ghost.io/c/af/ca/afcaa655-46e2-45b8-889a-2881de5cce69/content/images/thumbnail/ChatGPT-Image-Sep-8--2026--01_50_34-PM-1-345f1d1d-8469-48d0-b4dd-742d9d4fa5b0.png)](https://spyglass.org/thoughts-on-ais-maniacal-mixed-messages-siri-ais-potential-developer-problem/) [Bill Gates, Doomer • The Mac Mini MorphsUh, Bill Gates sounds a lot like... Dario Amodei in writing up his thoughts on where he thinks the world is headed with regard to AI. It’s a good essay, worth the read, but it’s certainly not going to help the AI perception problems! (And it’s also![](https://storage.ghost.io/c/af/ca/afcaa655-46e2-45b8-889a-2881de5cce69/content/images/icon/Spyglass-Rings-----Multi-aaf07995-5a63-43b5-b287-79ecadc12667.png)SpyglassM.G. Siegler![](https://storage.ghost.io/c/af/ca/afcaa655-46e2-45b8-889a-2881de5cce69/content/images/thumbnail/ChatGPT-Image-Aug-26--2026--03_56_25-PM-1-e6aade88-1169-4805-b658-29ccfe45d01c.png)](https://spyglass.org/thoughts-on-bill-gates-doomer-the-mac-mini-morphs/) [What if Anthropic is Right?What if we’re dangerously close to tripping into oblivion?![](https://storage.ghost.io/c/af/ca/afcaa655-46e2-45b8-889a-2881de5cce69/content/images/icon/Spyglass-Rings-----Multi-f113f83c-a702-4f9e-be60-d1b4fd2647a5.png)SpyglassM.G. Siegler![](https://storage.ghost.io/c/af/ca/afcaa655-46e2-45b8-889a-2881de5cce69/content/images/thumbnail/ChatGPT-Image-Aug-3--2026--09_26_21-PM-52d7698f-a394-4340-88d1-dea56f31765e.png)](https://spyglass.org/what-if-anthropic-is-right/) [AI Am Become DeathAs Anthropic blows up their potential AI usage, the Pentagon goes nuclear…![](https://storage.ghost.io/c/af/ca/afcaa655-46e2-45b8-889a-2881de5cce69/content/images/icon/Spyglass-Rings-----Multi-89a079a9-2635-448f-aace-a04a39d03aef.png)SpyglassM.G. Siegler![](https://storage.ghost.io/c/af/ca/afcaa655-46e2-45b8-889a-2881de5cce69/content/images/thumbnail/oppenheimer-071723-2-b381da3f9c314ca7b2325ba2056c50c3-ef42adc4-f8ae-4a93-b3ca-1eebed10eb30.jpg)](https://spyglass.org/ai-nuclear-weapons/) [Pushing or Pacing or Pulling Back From the AI Frontier?Is this a breather in AI development, or something more?![](https://storage.ghost.io/c/af/ca/afcaa655-46e2-45b8-889a-2881de5cce69/content/images/icon/Spyglass-Rings-----Multi-5ca23cb0-13c0-46d7-a5ea-e6fd98a9fa62.png)SpyglassM.G. Siegler![](https://storage.ghost.io/c/af/ca/afcaa655-46e2-45b8-889a-2881de5cce69/content/images/thumbnail/ChatGPT-Image-Aug-19--2026--04_40_32-PM-a4faab04-255f-4afe-a4cc-f0cbdb143f98.png)](https://spyglass.org/ai-frontier-race/) --- 1 Which it *seems* like Bilal Chughtai [did](https://www.linkedin.com/feed/update/urn:li:activity:7505380570303062016/?ref=spyglass.org) in this most recent instance. But, well, you can read his background and beliefs [in his own words](https://bilalchughtai.co.uk/?ref=spyglass.org)... [↩](#)