Methodology: I'm not a fan of political talking heads, so I'm essentially operating by entering the handles of influencers and bloggers I've heard of alongside keywords like "successfully predicted" into search engines.

Results:

 - Michael Knights (a foreign policy writer) predicted the inability of the U.S. campaign in Yemen to subdue the Houthis within a reasonable timeframe, arguing that they could continue to put pressure on the U.S. indefinitely.

 - George Behizy (a fairly standard Conservative Twitter/X influencer) predicted that Maduro would be abandoned by his political allies and removed from power in a covert operation within a matter of months back in November.

- Andreas Krieg (another foreign policy analyst, some appearances on CBC) predicted the lack of major Iranian retaliation after Operation Midnight Hammer, in favor of pursuing a face-saving offramp.

I have no special love or dislike for any of the four people whose predictions we've mentioned. Speaking objectively, all of these claims are about on par with each other in terms of impressiveness. They saw a pattern and predicted that the pattern would continue.

I should note that the bar is much lower than this, of course. If you randomly guess at geopolitical outcomes for a year, and your target audience knows very well that you share their political tribe, then, by the end of the year, you will have an audience that remembers all of your correct predictions and has forgotten all of your incorrect ones. Qualitatively, I've observed sentiments (and justification) identical to the ones you expressed whenever I have talked to anyone about a political influencer they're a fan of.

"boring calibrated"

A google search returned no results for me. What does this mean?

called out the possibility of snatching bolsonaro; then they snatched bolsonaro.

Do you mean Maduro? If so, lots of people were predicting a stand down of military forces followed by Maduro being removed from power. In the weeks leading up to it, there was a very public campaign by the U.S. along the lines of "Hey, Venezuelan elites, we'll take good care of you if you allow the removal of Maduro." I wouldn't use it as a singular example of profound prescience; I think any fan of any political talking head could point to a couple dozen similar predictions.

Funny enough, every political faction I've ever seen claims that the image in your post is representative of what's happening to them.

I also continue to find Belle of the Ranch to be good at finding news that I wouldn't otherwise find and which turns out to be relevant later.

A cursory look at the channel yields video titles that look a lot like partisan clickbait to me. It's the same sort of rhetoric I see on Facebook. "<Politician's> Moves BACKFIRE in the most HILARIOUS way!".

I do not think I would trust this channel to tell me anything other than whatever is maximally favorable to her preferred political party. If your priors are that anything favorable to her party is certain to be true and anything unfavorable to it is certain to be false, then I'd expect you'll find her insightful, but, if that were the case, you'd be able to save substantial time by just writing those priors on a rock and consulting the rock.

Looking at the data collection, region and country are fairly straightforward, but "Europe" could mean a wide variety of things to a wide variety of people, meaning that the identification numbers might overestimate attachment to the EU as a political structure. Someone might feel attached to the historical Europe, as a cultural and technological force, yet not consider the EU to be meaningfully connected to it.

More directly, I think the core issue that the EU faces is that of interests. Every strong or rising power today can point to a clear set of interests motivating its base of power (usually a majority of citizens) to support it:

• The PRC can advertise that it has made its citizens wealthier by facilitating nation-wide industrialization and negotiating trade agreements that have greatly increased living standards. Likewise, it can point to China's pre-PRC experience under the influence of foreign powers, and the more recent experience of some nearby countries, to claim that this newfound wealth isn't possible without a strong central government to keep other powers from using the threat of force to rent-seek. Moreover, the PRC can claim that the interests of China are the interests of the overwhelming majority of its current citizens' descendants - even if they happen to live elsewhere.
• Russia can point to the early post-Soviet years as an example of what happens when there isn't a strong, independent government serving as a check on foreign investors and local oligarchs. Moreover, there is certainly a perception - starting with the Bush administration appearing to spurn early attempts at reconciliation -  that the U.S. has committed to being an enemy of Russia, and that without central coordination of resources to deter them, they would forcibly bring back the bad old days of state-permitted looting of everything not nailed down by anyone with a carpet bag full of money. Also, as with the above, it is generally understood that people can be intrinsically Russian regardless of their citizenship, and that their fates are intertwined with the fate of the Russian government. <Backlash / Discrimination, depending on your political leanings> against overseas ethnic Russians post-2020 has cemented this.
• America, in part due to ongoing polarization, has two separate proposals to two somewhat-separate bases of power. 
  • To a degree, it is still a nation-state, and can claim that a wealthy America provides opportunities for people who are descendants of the historical American nation. This is a bit less intense than in the above two countries, since these Americans could renounce citizenship and become a fully-assimilated Australian/Spanish/Austrian within the span of one generation, if so inclined, but there is still a perception that the umbrella of American hegemony provides leverage to ordinary citizens. It's this perception that keeps the pointy end of the U.S. military well-staffed.
  • The second proposal is as a somewhat international collective of elites, where they can come together to make sure the world remains a good place to do business, and negotiate with "rogue" actors from a shared position of strength. Everyone can attend the same fundraisers, invest in the same companies, and send their children to the same universities to network with each other. Notably, it's this perception that allows the U.S. to arrange things like the stand-down that enabled the <Arrest / Kidnapping, depending on your political leanings> of Maduro by appealing to local elites who might want their kids to go to Harvard.

A shared language and some cultural overlap isn't enough to create a nation - there have to be shared interests that motivate people to be willing to work and fight in the interest of the collective. 

From what I've seen, there is a core political class in Europe which serves as the EU's base of power and is glad to provide it with support in exchange for coordinating the funding of friendly NGOs and helping to oppose non-aligned political movements, but this is a relatively small group that does not enjoy universal popularity. To meaningfully counterbalance America, Russia, and China, there would need to be a common cause that a large number of people are willing to go above and beyond for - the sort of thing makes workers do more than they're paid to do and makes soldiers willing to fight rather than rout. The usual motivation is something along the lines of "This nation is the exclusive property of all of your descendants, forever, and the blood and sweat you invest will never be lost to them", but the Venn diagram of people advocating that for Europe and people who support the EU is very close to two nonintersecting circles.

I don't know that it necessarily advocates cooperation so much as coexistence, or tolerance. The message isn't "we all want the same things, and we should work together to get them", it's "once you decide that it's evil to allow the other half to continue their way of life - regardless of whether they want the same things as you do - you've guaranteed a war that will end your own way of life just as thoroughly".

I think the ending makes that clear. You don't get to just wipe out all of your enemies for free, they'll fight back. The choice you get to make is whether the 'good' of your own survival outweighs the 'bad' of their survival, or whether you'd rather the two cancel each other out.

So this would forego our ability to assess how well they do autonomously and make the scenario more similar to having custom or dynamic prompting per task.

I considered this, but, in the wild, we'd expect to see LLMs using a baseline level of generic prompt engineering for the interfaces they have access to. I wouldn't suggest per-task custom prompts, but looking at the SOTA for general scaffolding might get more true-to-life results.

Looks like a neat idea, though things do sort of converge to vague positivity rather than actually getting anything done. You mention that the scaffolding is a sticking point, and I wonder if you wouldn't get better results by just grabbing the top system prompts for similar tasks and hard-sticking them onto the models, maybe with a bit of manual prompt engineering, such that e.g. Gemini always sees a very explicit order to assume user error when something goes wrong.

The other issue is that open-ended real-world tasks are a bit unfair to give to production LLMs, on the basis that anything these LLMs could do on their own would've already been done by an enterprising human using the very same LLMs, but with the intent to succeed^[1] rather than the intent to evaluate the model's performance.

1. ^{^}

   Characterized by more direct prompting, more frequent intervention, and a general willingness to 'cheat' on the LLM's behalf by doing things manually when it's bad at them.

It's interesting that Facebook/Meta fell so far behind in AI despite the substantial resources on hand. 'Metaverse' was an inherently flawed idea that they thought they could make work through market leverage, but well-scoring LLMs have been done successfully by a wide variety of organizations, from Alibaba to X to OpenAI to Anthropic. 

Is it something organizational? Does Facebook have any successful spinoff initiatives?

I'll have use for a model with a time-horizon 100x bigger than now in three years, I don't know if I will have use for a model with a time-horizon 10,000x bigger than today in six years

This is where I get lost, here. Isn't "there will be a model with a 10,000x bigger time-horizon" equivalent to "the singularity will have happened"?

Some people argue that the time horizon won't keep growing at the same pace, and it will plateau, and others argue that it will and we'll get a technological singularity, but if an LLM can do anything that would take a moderately competent human five years, then that does seem like the end of our current mode of civilization.

In other words, I don't see a set of possible worlds where LLM time horizons get too long to be marketable to hobbyist engineers and that lack of marketability is still a concern.

Conversely, if gorillas and chimps were capable of learning complex sign language for communication, we'd expect them to evolve/culturally develop such a language.

We've seen an extreme counterexample with octopi, which can be taught some very impressive skills that they don't pick up in nature because they aren't sufficiently social to develop them over the course of multiple generations. I think it's within reason that gorillas could have the ability to learn more complex language than they use, so long as it's not economical for them to spend time teaching their offspring those complexities as opposed to teaching them other things.

I will say that I'm very skeptical of Koko, though, for other reasons.