Humanity Could Survive AI If We Stop Within 10 Years | Nate Soares artwork

Humanity Could Survive AI If We Stop Within 10 Years | Nate Soares

MTS

November 13, 2026

Nate Soares, president of the Machine Intelligence Research Institute (MIRI) and co-author of If Anyone Builds It, Everyone Dies, joins MTS to discuss MIRI’s pivot from technical alignment to public advocacy and their new theater ad campaign, Do Not Accept.

0:00

Nate Soares

The main way that I expect to get out of this is humanity putting a stop to the most dangerous AI research. I haven't said we can't, you know, if we survive 10 years mostly, my guess is that we put a stop to the most dangerous research.

0:15

SPEAKER_2

And we're back. We are live with Nate Soares, who's president of the Machine Intelligence Research Institute, MIRI, and co-author, along with Eliezer Yudkowski, of If Anyone Builds It, Everyone Dies. So, MIRI just released their new campaign yesterday called Do Not Accept, which was a very, very moving, very emotional video. So, Nate, welcome to MTS.

0:40

Nate Soares

Thanks for having me.

0:42

SPEAKER_2

So, tell us a little bit more about this new comm strategy for MIRI, Do Not Accept and the other things around it.

0:50

Nate Soares

Yeah, you know, it's part of a broader strategy to get more people to understand that, you know, these AI companies really are gambling with our lives.

They're sort of really trying to build tech that threatens human extinction. I happen to think the extinction risk is quite high, but even people who disagree with me say the extinction threat is, you know, 10 percent or more than 10 percent if they sort of don't want to be pinned down too much. So, you know, this campaign, I think it's a great short film. It's going to show in theaters between previews or before previews across the nation. And this is sort of one of many things we're trying to do to help people understand what's going on in Silicon Valley.

1:39

SPEAKER_2

It seems like there's been a pretty tremendous rise in salience about AI over the last few months, especially because of Jacob Coxon resigning from Anthropic. How long do you think it will take to make AI an issue on the scale of, say, climate change?

2:00

Nate Soares

Yeah, it's hard to say. We've sort of seen polls from the election where, I don't remember the exact numbers, but data centers are an issue that voters talk about. And with sort of further polling, I think we see a lot of people feel like data center protests are a way that they can protest what's happening with AI. And we hear a lot about people who care about data centers from an environmental perspective. My impression is that a lot of people care about data centers because they sort of have a sense that this AI stuff is not going in a direction that's going to go well for them. I think my impression talking to a lot of people who sort of aren't in the AI bubble is that AI extinguishing humanity feels pretty plausible. A fully automated economy run by billionaires feels pretty plausible. Various AI dystopias feel pretty plausible, but they don't really buy these utopic pictures. I think we are seeing a lot of people concerned about a possible extinction threat here. I think salience is rising. I think there's a good chance that the 2028 elections will be around AI.

If things like this meter report saying that maybe we don't even have until the 2028 elections turn out false, which hopefully they do, or hopefully the data turns out misleading.

I hope it doesn't go quite down the same road as climate change. I think the climate change took a lot of wrong turns, but I think it'll probably get similarly salient, potentially quite fast as the tech keeps developing.

3:36

SPEAKER_2

You preempted my next question, which is how do you prevent the AI safety movement from essentially becoming another version of climate change, where it gets hyper-polarized, it gets dragged down into a lot of sort of less substantive populism, and you end up in a situation where the goals, the original goals of the movement are essentially defeated by things that you don't want them to be?

4:08

Nate Soares

Yeah, and I mean, I would even say, I think that the environmental movement sort of shot itself in the foot with bans on nuclear power, where I think, you know, it's a lot more people have died from coal dust in their lungs and have died from nuclear power fiascos. So, yeah, you know, I think it's an important question. I don't feel like I definitely have the answer.

If we could experiment on, you know, 100 planets, we'd, a whole lot of things about this whole AI situation would be easier. One of my ideas here and one thing that I try to do is sort of say laser focused on the extinction threat issue. I think a lot of those sort of populist forces will pull things towards like, you know, we want to stop AI because of the effect on artists and because of the effect on the environment, because of the effects on education, and also because it might kill us all. And from my perspective, you know, one of those things is not like the other.

Read the other 20 minutes of this episode here.

$15 gets you this episode and 99 more. Credits never expire. After checkout, one link brings you back here to read it.

Not what you expected? Full refund within 14 days.

Thousands of transcripts fetched by people building searchable podcast archives

Fetch transcripts with code

The same credits work from a script or an AI agent. The demo key returns a sample episode in full, no card needed:

request
curl -H "x-api-key: pt_demo" \
  https://spoken.md/transcripts/1000651996090

Markdown with the speakers named, for your notes, your knowledge base, or anything that makes HTTP calls.

From $0.10 per transcript. No subscription. Credits never expire. Prices exclude VAT, added at checkout for EU customers. Not what you expected? Email us within 14 days with 20 or fewer credits used and we refund the pack in full.

Using your own key:

request
curl -H "x-api-key: YOUR_KEY" \
  https://spoken.md/transcripts/1000801737667