The root cause tree on page 452 gives a good overview of how complex the behavior can be.
The good news is that the grid operators have a good idea of what the problem was/is and it's well understood how to fix it. The downside is that it will require quite a bit of both time and money to reinforce the grid infrastructure.
Wisdom of the crowd has some fatal flaws that are especially important when it comes to things like IPOs.
- Most significantly, most scientific research focuses on things that are actually amenable to guesses with a normal distribution, like "amount of jellybeans in a jar" or "length of the border between country A and country B". An IPO is a binary choice where it either goes public or not. There is no correct value to converge to.
- It has been shown that as bettors gain more information about the bets of others, predictions lose accuracy and bettors converge to a consensus value instead. It seems to me that online prediction markets would be extremely prone to this as the bets of other people are all there in the market price.
- Prediction markets generally become more accurate as the diversity of the bettor pool grows. The users of polymarket and Kalshi heavily skew towards young men from certain socioeconomic groups, who may be biased towards one or the other outcome.
In the case of an OpenAI IPO, it seems likely multiple of these would converge as people start to fall prey to groupthink because "everybody knows that they'll IPO soon" in their local media bubble.
The question isn't "will they IPO" the question is "when will they IPO" which is not a binary question. The rest of your point about Polymarket users now being mostly degenerate gamblers is true though.
The relative timing, valuations (and float sizes) of the expected SpaceX, Anthropic and then OpenAI IPOs would still be highly correlated. Even allowing for moral degeneracy among most of the gamblers on this particular Polymarket market.
It’s two binary questions about whether they will IPO by specific dates. It’s not obvious to me that this maps to a more granular “when will they IPO?” question.
> It has been shown that as bettors gain more information about the bets of others, predictions lose accuracy and bettors converge to a consensus value instead.
This makes intuitive sense to me; is there a name for this phenomenon?
A prediction platform’s biggest value is publicising information from possible insiders, who at some point will work harder to maintain secrecy not to lose their informational advantage. So all that remains are people gambling on public info.
That said, greed from insiders looking to make a quick buck will always skew the price towards ‘truth’
> An IPO is a binary choice where it either goes public or not. There is no correct value to converge to.
Of course there is - they are betting on the "when".
> It has been shown that as bettors gain more information about the bets of others, predictions lose accuracy and bettors converge to a consensus value instead.
I dunno how to reply to this - that is exactly my point, but it appears (to me, anyway) that you are saying this in disagreement?
Let me clarify - my point is that wisdom of the crowd converges on to a value that is quite near the actual value.
> In the case of an OpenAI IPO, it seems likely multiple of these would converge as people start to fall prey to groupthink because "everybody knows that they'll IPO soon" in their local media bubble.
Sure, if everyone is in the same local media bubble, that once again, that is unlikely, because these are people who don't make money from the result, they make money from correctly predicting it, hence they are exactly the demographic that will seek out more and more information outside of any bubble they may be in.
It's one thing when proponents of $FOO spend time boosting their PoV/wishes/hopes on a forum. It's quite another when they have to put their money where their mouth is: then they are open to new information!
> I dunno how to reply to this - that is exactly my point, but it appears (to me, anyway) that you are saying this in disagreement?
Yes. The research shows that they do (almost) always converge, but that they DON'T always converge to an accurate value. In particular, there can be behavioral biases at work that warp the perception of bettors in one direction or the other. A well documented case of this is when fans of a sports team pile in and bet for their favorite team, causing the price to shift too much towards the more popular team. Exposure to the predictions by other bettors then causes the total market to converge to the biased price. Interestingly enough, people still do this even though this phenomenon is well documented and they have to put their money on the line. People just don't care enough about their $10 bet to do thorough research.
In a similar way, I would not at all be surprised if some people are such fanboys of OpenAI that they start to display cultlike behavior. You can easily find such people online even on this very site. It's not such a weird thing to consider that people at the peak of a hype cycle don't always behave in rational ways, especially when they're just betting $10 when drunk on a Saturday evening.
if you can identify where and how prediction markets are wrong, why aren't you applying that and making millions?
> - Prediction markets generally become more accurate as the diversity of the bettor pool grows. The users of polymarket and Kalshi heavily skew towards young men from certain socioeconomic groups, who may be biased towards one or the other outcome.
Citation? If your small population is high IQ, accurate predictors and you diversify to average IQ population, won't the accuracy go down not up?
You could do this whole plan without needing to do the assassination part, as long as you are willing to throw Ukraine under the bus. Conversely, even if the assassination scheme goes exactly as planned, there is no way of guaranteeing that the new Russian leader would be willing to restart gas deliveries until the war in Ukraine has wrapped up. Given that especially the eastern EU countries have absolutely no intention of allowing Ukraine to lose, this seems like a very tall order.
Finally, both Nordstream 1 AND Nordstream 2 still have a gaping holes in them from the bombings so restarting deliveries will probably take several years at least.
All in all, this plan gets only a 2 out of 10 for being impractical, too slow AND depending on factors outside our control. 1 point because it does at least sound spy-ish and proactive.
The problem for Europe is that they have nothing to trade with Russia. They can get better prices elsewhere.
Before 2022 we had the big EU auto companies in Russia, we also had nice handbags, shoes and outfits to sell to Russia. Plus we could have them hide their money in London property. Machine tools were also a brisk trade.
Nowadays Russia needs nothing from Europe. Nothing apart from peace and their 300 billion back. But we have gone past that stage. The Russians have never broken any energy contracts in this, the West has cut themselves off.
Regarding the EU not wanting the former Ukraine to lose, there is a difference between what the officials want and what the people want. From Finland to Portugal I am sure most people would want no war and cheap energy, however, their 'leaders' are just doing what Washington tells them to do.
Delusional. In Europe people want not to just see Ukraine win, but also Russia to lose. It's the leaders that are timid, the people would like to see much wider ranging support for Ukraine.
The Russian energy will not be accepted again in Europe for the foreseeable future, ending the war will not change that.
Huh? Even in the US the majority are sympathetic to Ukraine in every survey. Sure we all want no war - but we are not really in favor of no war at any price.
I agree, but also find it funny that by that standard the DRM in the original Google video streaming product was not hacked before the service was shutdown, after about 2 years :)
It was unhackable while it mattered. It was hacked 5 years after it no longer mattered. And all but the effectively beta release remain unhacked even now.
To the community it was unhackable, until very recently.
It's security measures held up so long that it appeared to be unshakable. There were no obvious flaws.
In hindsight it was hackable, but keep in mind how long it took. This console has long been obsoleted.
Well obviously? It's literally being broadcast in the news when diplomats talk to each other. What do you think they are talking about if not policy discussions?
So it seems normal that a bunch of politicians, in the current climate, got together and decided that the weakest form of age verification imaginable absolutely had to get passed everywhere?
I'm not saying there's definitely no coordination, but nobody had to get together to decide that 2026 was the year for 90s fashion to make a comeback. Human society is very prone to fads in all areas.
> They found previously unknown race conditions. Not in obscure corners of the codebase. Not in exotic edge cases. In the kind of code that every Erlang application depends on, code that had been running in production for years.
If these race conditions are in code that has been in production for years and yet the race conditions are "previously unknown", that does suggest to me that it is in practice quite hard to trigger these race conditions. Bugs that happen regularly in prod (and maybe I'm biased, but especially bugs that happen to erlang systems in prod) tend to get fixed.
True. And that the subtle bugs were then picked up by static analysis makes the safety proposition of Erlang even better.
> Bugs that happen regularly in prod
It depends on how regular and reproducible they are. Timing bugs are notoriously difficult to pin down. Pair that with let-it-crash philosophy, and it's maybe not worth tracking down. OTOH, Erlang has been used for critical systems for a very long time – plenty long enough for such bugs to be tracked down if they posed real problems in practice.
Erlang has "die and be restarted" philosophy towards process failures, so these "bugs that happen to erlang systems in prod" may not be fixed at all, if they are rare enough.
As of now, the post you're replying to says "bugs that regularly happen ... in prod"
Now, if it crashes every 10 years, that is regular, but I think the meaning is that it happens often. Back when I operated a large dist cluster, yes, some rare crashes happened that never got noticed or the triage was 'wait and see if it happens again' and it didn't happen. But let it crash and restart from a known good state is a philosophy about structuring error checking more than an operational philosophy: always check for success and if you don't know how to handle an error fail loudly and return to a good state to continue.
Operationally, you are expected to monitor for crashes and figure out how to prevent them in the future. And, IMHO, be prepared to hot load fixes in response... although a lot of organizations don't hot load.
That is obviously true but doesn't mean as much as you seem to think. Washing laundry is also not much work but it adds up to a lot over the years, especially if you skip a few weeks of laundry every once in a while. That is not an excuse to not do it.
The answer is the same in both cases: acquire some discipline and treat maintenance with the respect it deserves.
reply