Wow, I like scientific mindset a lot more than "truthseeking" (what does it mean??) or scout mindset!

I think you are right that there is too much jargon in the new set of principles and the old set is much nicer.

I also agree there should probably be a consultation with the community on this.

But, even if I accept that, it doesn’t change the fact that a radio broadcast is unnecessary due to points (4) and (6). Doing the radio broadcast or not wouldn’t change anything.

Thank you for your response!

This is an important point to consider. OpenAI is indeed exploring how to put ads on ChatGPT. 

My main source of skepticism about this is that the marginal revenue from an online ad is extremely low, but that’s fine because the marginal cost of serving a webpage or loading a photo in an app or whatever is also extremely low. I don’t have a good sense of the actual numbers here, but since a GPT-5 query is considerably more expensive than serving a webpage, this could be a problem. (Also, that’s just the marginal costs. OpenAI, like other companies, also has to amortize all its fixed costs over all its sales, whether they’re ad sales or sales directly to consumers.)

It’s been rumoured/reported (not sure which) that OpenAI is planning to get ChatGPT to sell things to you directly. So, if you ask, "Hey, ChatGPT, what is the healthiest type of soda?", it will respond, "Why, a nice refreshing Coca‑Cola® Zero Sugar of course!" This seems horrible. That would probably drive some people off the platform, but, who knows, it might be a net financial gain.

There are other "useless" ways companies like OpenAI could try to drive usage and try to monetize either via ads or paid subscriptions. Maybe if OpenAI leaned heavily into the whole AI "boyfriends/girlfriends" thing that would somehow pay off — I’m skeptical, but we’ve got to consider all the possibilities here.

There’s a longer discussion of that oft-discussed METR time horizons graph that warrants a post of its own. 

My problem with how people interpret the graph is that people slip quickly and wordlessly from step to step in a logical chain of inferences that I don’t think can be justified. The chain of inferences is something like: 

AI model performance on a set of very limited benchmark tasks → AI model performance on software engineering in general → AI model performance on everything humans do 

I don’t think these inferences are justifiable.

Thank you very much for your reply. I especially want to give you my profound appreciation for being willing to revise how your results are described in the report. (I hope you will make the same revision in public communications as well, such as blog posts or posts on this forum.) A few responses which I tried to keep as succinct as possible, but failed to keep succinct:

• I don’t think the "best matching" qualifier mitigates the reasonable concern (not just mine, but others’ as well) with the three-scenario framing. The concern is that the design of question may create an anchoring effect. The sophistication of the respondents also does not mitigate this concern.
• I don’t think the disclaimer you and Benjamin Tereick both quoted makes much difference to our discussion here, on this forum. It’s just an elaboration on what "best matching" means, which is good to include in a survey, but which is already intuitive to readers here.
• You seem to be saying that something significantly less than the slow progress scenario would be an "extreme possibility". Is that correct, or am I misunderstanding? If so, I strongly disagree. If you and the other authors of this report view that as an extreme possibility, I would worry about you baking in your own personal AI forecasts into the design of the survey.
• If I were designing a survey question like this, I would make the lowest progress scenario one in which there is very little progress toward more powerful, more general AI systems. For example, a scenario in which LLM and AI progress more or less stagnates from now until 2031, or a scenario in which only modest, incremental progress occurs. If you wanted to stick with three scenarios, that could be the slow progress scenario. (I’m no expert on survey design, but I don’t see how you could reasonably avoid having something like that as your lowest progress scenario and still have a well-designed question.)
• My complaint about framing the results as experts’ "probabilities" is that this directly contradicts the "best matching" qualifier. I didn’t raise a complaint with the intersubjective resolution framing of the question.
• That being said, I do find the intersubjective resolution framing counterintuitive. I didn’t bring this up until now because I find it difficult to wrap by head around, I wasn’t sure if I was misunderstanding the logic behind it, and it seemed like a distraction from the more important point. The reason I find this framing counterintuitive is best explained through an analogy. Let’s say you want to ask experts about the probability of the volcano Eyjafjallajökull in Iceland erupting again between now and 2050. Let’s say you ask, "In 2050, what percentage of vulcanologists will say Eyjafjallajökull has erupted?" This is confusing because you would think the percentage would be around 0% or 100% in virtually any realistic scenario. If someone thinks there’s a 51% chance the volcano will erupt, then they should say 100% of vulcanologists in 2050 will think that. If they think there’s a 49% chance it will erupt, they will say 0%. I don’t understand how you translate this into probabilities, since the only numbers the respondents are telling you are 0% and 100%, and neither is their probability of an eruption. Even if there is a logical way to do this that I’m not getting, can you rely on all your survey respondents to understand the question in the way you intended and answer accordingly?
• With regard to edit #3, I was indeed only pointing out an example of the broader phenomenon of framing bias and question wording bias. Another example I raised in the comments was the AI Impacts survey that found a 69-year difference in the median date of AGI essentially just by changing the word "tasks" to "occupations". Most would not expect differences so large ahead of time and most people have expressed surprise at the amount of difference in responses to the different versions of these questions. It’s relevant to bring up because Benjamin Tereick’s argument was that the results to the one version of the three-scenarios question don’t seem to him to indicate a question wording bias. My counterargument is that you can’t tell how much bias there is from just one version of the question; you have to ask at least two different versions and see the results. For what it’s worth, the concern with the three-scenarios framing is a potential anchoring effect and I think it makes sense to understand percentage probabilities as causing an anchoring effect for the reason titotal explained here.

Thanks again for a helpful, cooperative, and open reply.

To be clear, just a few of the reasons why this doesn’t make sense are:

1. We don’t know that humanity will create superintelligence soon.
2. We don’t know that a superintelligence is likely to be malicious.
3. We don’t know that there are alien civilizations in our galaxy or elsewhere in the universe.
4. If many alien civilizations in our galaxy achieved superintelligence thousands or millions of years ago, there are ways of them monitoring Earth. A civilization that achieved superintelligence 1 million years ago has had plenty of time to set up probes throughout the galaxy. A civilization that achieved superintelligence 1,000 years ago hasn’t, but, as you mentioned, it can tell which planets have conditions conducive to life. It could notice the rapid increase in carbon dioxide in the atmosphere, which is evidence of industrialization. It could also monitor for unintentional and intentional radio broadcasts. And this is only what we can imagine with our current science and technology and our current intelligence.^[1]
5. It’s unclear how our Earth-originating superintelligence would be a threat to alien civilizations that achieved superintelligence thousands or millions of years ago. This sounds implausible on its face and needs to be justified, which it probably can’t be.
6. If alien civilizations have been dealing with many other alien civilizations with superintelligence for thousands or millions of years, they don’t need a warning from us about it.
7. The simulation stuff is totally off the rails. That doesn’t even have anything to do with SETI, METI, or the Fermi paradox. Plus, it’s extremely bizarre and sounds like complete madness. I don’t think the simulation argument/simulation hypothesis makes any sense in the first place — on its own, it’s quite mad — and this takes the simulation idea up a notch to an even stranger place, with no attempt to logically support it (not that such an attempt could ever succeed, anyway).

All in all, this is just a whimsical sci-fi idea and not a serious or credible proposal for how to allocate resources that instead could, say, save the life of someone in sub-Saharan Africa.

1. ^{^}

   Edited on Nov. 17, 2025 at 3:50 PM Eastern to add:

   This post seems to assume that alien civilizations would have some sort of alliance or diplomatic relations. If true, and if there are many different alien civilizations in our galaxy, presumably with quite different ages, older civilizations with probes hidden everywhere could communicate information about Earth to younger civilizations without probes. This communication could happen either via radio broadcasts that Earth could not pick up on or via other communication technologies that Earth does not have.

Oh, cool! Thank you!

Interesting project! Is there a Creative Commons open textbook for social work you could add to the mix?