michal_dubrawski

I know that Laplace's Rule of Succession is overused in judgmental forecasting, but the formula might be some reference point if we can treat the process behind the event occurrence as having a constant probability in the analyzed period.
It is likely not constant, but as I see it, this is a decision which can be made at any point of time, so treating it as something with constant probability may be reasonable - do you agree?
I mainly object to using it in cases of forecasting related to evolving processes, where we in fact have more information indicating the progress of such process. Like for example, the company is developing a game, and it has been announced to be released on date X - then using this rule and taking the months or days between its initial announcement and today as time when it has not been released as a basis for our calculations of probability of its release is absurd. The probability cannot be the same for the first days after the announcement and the days around the date of its planned release. It would be more appropriate to look at the usual delays for previous productions of that studio, or delays of similar game projects by similar game studios. We could include cases of games which were never released for a separate probability of project being abandoned or the studio going bankrupt during the work on the project.
You may see this as obvious, but I have seen people I highly respect as forecasters using the Laplace's Rule like that - good forecasters can make mistakes too, I make mistakes not as rarely as I would want 😉 and I am very happy to learn from all of you, please let me know when you think you see a mistake in my thinking, knowledge, assumptions or calculations.

So, we've observed 78 months where the event didn't occur (since February 2018 when O-RAN started)

The number of successes (s) is 0

The number of trials (n) is 78

Applying the formula:

P(event occurs in a single month) = (0 + 1) / (78 + 2) = 1 / 80 = 0.0125

We want to know the probability of this event occurring at least once in the next 13 months (the time till end date). This is equivalent to 1 minus the probability of it not occurring in any of those 13 months:

P(event occurs in next 13 months) = 1 - P(event doesn't occur in next 13 months)

= 1 - (1 - 0.0125)^13 ≈ 15.09%

So, I don't think we are at the time of peak probability of that happening, the highest motivation for Huawei to join O-RAN was before they started to be removed from NATO countries (see this article: https://www.lightreading.com/5g/europe-s-inaction-on-huawei-may-have-come-at-the-worst-5g-time), and they were dismissive about O-RAN in the past in their official comments.  Now the probability still might be a bit higher than for a random month because of what is happening in Germany - https://www.infer-pub.com/comments/134520 - but the lack of comments about that from Huawei in more than a month after the article still makes me think it is more probable that this is Huawei clients initiative which Huawei may not support. So that part of my last forecast stays the same: https://www.infer-pub.com/comments/137717

I changed my mind, updated my beliefs on several things:
1. The resolution criteria are quite broad and inclusive, and the threshold for a  resolution as "yes" is lower than the title of the question suggests. 

This question will be resolved as “Yes” if Huawei or the O-RAN Alliance announce a collaboration between them on or before 1 October 2025. Such a collaboration might include, but is not limited to, Huawei joining the O-RAN Alliance, testing or developing equipment with O-RAN standards, or making financial or resource commitments.

So, as I see it it can be triggered by Huawei announcement that they are testing or developing equipment with O-RAN standards, or making financial or resource commitments. It might be enough if Huawei declared that it is testing or checking the feasibility of such a solution, for example as a response to its German clients' demand, to resolve the question as "yes", since checking means resource commitment. I don't think it was an intent of the INFER Question Team to be that inclusive as I just described, but, I think we should take these possibilities into consideration. 

2. This article https://www.lightreading.com/5g/europe-s-inaction-on-huawei-may-have-come-at-the-worst-5g-time makes me update on the argument that Huawei may not want to resign from software because they would lose the opportunity for spying or sabotage: 

Software is a feature of not just the management systems but also the baseband units that process mobile signals and are usually found at the bottom of a mobile mast. Even if Huawei was evicted from the management system, it could still introduce malicious code into the baseband software, said one industry source on condition of anonymity. Lum agrees. "The baseband unit is a security issue," he said.
Huawei maintains that its software has undergone more rigorous scrutiny than that of any other vendor. The UK even established a dedicated facility, the Huawei Cyber Security Evaluation Centre, that puts the Chinese vendor's source code under a microscope. No other vendor is subject to the same examination. And despite this, no one has produced conclusive evidence of any wrongdoing by Huawei. 

3.  Interesting article shared by @DimaKlenchin   https://www.telecoms.com/telecoms-infrastructure/open-ran-is-failing-to-deliver-on-its-ultimate-promise showing that the idea of O-RAN might be losing its market attractiveness as it has not delivered what was promised:

The big hope for Open RAN was that it would flood the market with a tsunami of innovative, competitively-priced vendors, leaving operators spoilt for choice as to whose products they could use in their networks. By extension, it was also hoped that Open RAN would make it less costly for Western operators to replace equipment supplied by China's Huawei and ZTE.
The reality today is markedly mundane. Operators are still interested in deploying Open RAN, but most of them are using products from familiar names like Ericsson, Nokia, and – albeit to a slightly lesser degree – Samsung.
With the notable exception of CSPs in Germany, most telcos that needed to replace Chinese-made kit are well on their way to doing so, and as such, one of the driving forces behind Open RAN is ebbing away. 

Great find by @ScottEastman here: https://www.infer-pub.com/comments/134520 but more than a month has passed from that news and I haven't seen any Huawei remarks that they consider such a move as a solution (likely technical side of the process would take a long time).
That may not mean anything, but It looks to me as this talk about Huawei joining O-RAN to be able to stay in Germany came from Huawei clients rather than Huawei itself. However, it is important signal, since Huawei clients must have talked with the company about that option and it was publicly suggested as a way for Huawei to stay in Germany as a hardware provider.
According to this article western markets may not matter that much for Huawei and its biggest shareholder - the Chinese Government because the global market is big, and the company already handled the problem of being kicked out from the US and a few other western countries and is on the growth trajectory. One perspective is that for the Chinese government, with their global ambitions, Huawei might be worth more as an asset which can be used for espionage than just a source of income. From that point of view, keeping their own close source proprietary software seems like a much better move than joining O-RAN and allowing other countries to use other software providers with their equipment (see this article about technologies supposedly developed by Huawei - see the quote below since this is behind a paywall). Of course, I don't know Chinese decision makers preferences on that, but cases in which Huawei employees were accused of espionage activity (for example: link, link) or when the company was investigated for potentially malicious intent (link) strongly suggest that they do not see it only as a business. 

Quotes from the Washington Post's article  "Documents link Huawei to China’s surveillance programs" from December 14, 2021 mentioned above :

A review by The Washington Post of more than 100 Huawei PowerPoint presentations, many marked “confidential,” suggests that the company has had a broader role in tracking China’s populace than it has acknowledged.

These marketing presentations, posted to a public-facing Huawei website before the company removed them late last year, show Huawei pitching how its technologies can help government authorities identify individuals by voice, monitor political individuals of interest, manage ideological reeducation and labor schedules for prisoners, and help retailers track shoppers using facial recognition.

and:

The Huawei slides shed light on the company’s role in five surveillance activities in China: voice recording analysis, detention center monitoring, location tracking of political individuals of interest, police surveillance in the Xinjiang region, and corporate tracking of employees and customers.

They still have open job postings for the AI Research LAB in China (both posted two days ago on August 21st) - @PeterStamp already posted about that here: https://www.infer-pub.com/comments/137603): 
https://jobs.careers.microsoft.com/global/en/job/1755631/Principal-Applied-Scientist-(WPO-team)
https://jobs.careers.microsoft.com/global/en/job/1743227/Researcher

There was recent information published by Reuters that China state controlled entities used Amazon and Microsoft cloud service to bypass restrictions on access to high-end US chips. So the US government may update a law to remove that way of bypassing this ban, but that should not affect the Microsoft AI lab in China.
https://www.reuters.com/technology/chinese-entities-turn-amazon-cloud-its-rivals-access-high-end-us-chips-ai-2024-08-23/ 
https://www.reuters.com/technology/list-chinese-entities-who-have-turned-cloud-access-restricted-us-tech-2024-08-23/

With the clarification provided by @dante here, and especially this part "something like a sabotage mission or covert operation we wouldn’t consider large enough for an invasion. ", I reduce my probability for invasion of Lithuania to below 0.5% (so it rounds to 0%).

Even if by 2027 Russia has achieved their military objectives in Ukraine (however they will define them at that time) and declared victory, on a rational level, I think it is very unlikely they would want to start another confrontation this soon, and with a stronger opponent. The only scenario which comes to my mind of a large-scale invasion happening which could be considered as not violating the rational actor theory, is the one when it happens shortly after the start of the US-China military confrontation, most likely over Taiwan, especially, if China would be winning. However, a large-scale Chinese military confrontation with the US happening in this timeframe, likely would not fit into rational actor theory even if China would be confident about their military having the upper hand, because of the risks to their economy (I am not an expert on this, but I read that in Krzysztof Wojczal's book "Trzecia Dekada" and it makes sense to me). So that scenario itself seems very unlikely because there are a lot of "if" conditions to be satisfied first. Even if the US would leave NATO (again, something with very low probability - Metaculus community median is currently at 97% that the US will remain in NATO until at least January 1, 2029, I would probably say 99% here), I doubt that the US would just be watching the Russian invasion of NATO countries in Europe and doing nothing. There might be of course some other scenarios I have not thought of. 

But my point is that the rational perspective (at least if we define rationality as in rational actor theory) is not everything. Before full-scale invasion of Ukraine by Russia the consensus of most of the Superforecasters was for a long time that this will not happen. I remember that I initially agreed with their rationale, that this wouldn't be a rational move for Russia, but facts on the ground made me change my mind. I remember feeling bad going against the consensus of such a wise group of people with such a great track record (people I really respect and have a few friends among them), but I decided that to learn anything, after considering all the information I have, I need to follow my thinking and my judgment wherever it leads me. Some of my Superforecaster friends also changed their view on this, and the consensus eventually changed as well. There were also superforecasters and other great forecasters who were convinced about this happening very early, but they were a minority, an exception. I remember that Metaculites have done quite a good job there, but as a community we also haven't seen that months in advance. Some members of the great Samotsvety Forecasting Team won good money on prediction market betting that this will happen - I wonder if part of the success could have been non-western centered perspective and insights of Misha Yagudin - a great forecaster who is Russian.
My point here is that the consensus of so many great minds was for quite some time in its essence that Russia won't do it because it is not the rational thing to do for them.

As Robert Jervis wrote in his book "Perception and Misperception in International Politics":

"A state that is not willing to run major risks may misperceive or miscalculate and undertake very dangerous actions. The state’s behavior would not correspond then to its basic intentions. For example, the Russians probably grossly mis-estimated the risks they were running by putting missiles into Cuba. And at many points Hitler may have been reckless not because he willingly tolerated a high probability of war but because he was certain that the other side would back down. When his generals opposed his policy on the grounds that it was too dangerous Hitler did not argue that the risks were worth running. Rather he told them that the risks were slight. Indeed it may be that states that use force to alter the status quo often differ from others less in their willingness to run perceived risks than in the fact that they perceive low risks in situations where others perceive high ones."

and:

"(...) because differences in firsthand experiences lead to differences in perceptual predispositions, the chance of misperception is increased by events that one actor experiences but that others with whom he interacts know of only indirectly. Furthermore, each actor may be unable to understand how the other is perceiving and why it is behaving as it is."

Miscalculations and miscommunication happen and quite often lead to armed conflicts. I think it is safe to say that Russia miscalculated in 2022 with their initial goals, but we are yet to see how it ends for them in the long run.
Other example which comes to my mind is Iraq invading Kuwait. Here is the long quote from "Debriefing the President: The Interrogation of Saddam Hussein" book by John Nixon a CIA analyst who interrogated Saddam Hussein after his capture in 2003:

Saddam was forever puzzled by his country’s relationship with the United States. When we talked about U.S.–Iraqi relations, Saddam often got a perplexed look on his face, as if he was still trying to figure out where the relationship went wrong. “The West used to say good things about Saddam,” he said. “But after 1990 all that changed.” (Interestingly enough, this comment was echoed by officials from the George H. W. Bush administration during an NBC roundtable discussion on the twentieth anniversary of the Gulf War. Brent Scowcroft, who was national security adviser, said that after 1990, Saddam just changed. James Baker, the former secretary of state, shook his head in agreement. They couldn’t understand why. Things had gone along well in the 1980s, but somehow Saddam changed. Saddam had remained remarkably consistent in his governing and his penchant for doing the unpredictable. The Bush 41 administration had been caught unaware by Saddam’s foray into Kuwait. I strongly doubt that if Washington had made it clear to Saddam what it was willing to do to reverse any hostile Iraqi move against Kuwait, he would have crossed that red line.) Pointing out that America had supported Iraq during the Iran-Iraq War, Saddam said, “If I was wrong, why did the U.S. support me? If I was right, why did they change?” In Saddam’s mind, it was the United States that had suddenly and inexplicably changed course. “Iraq had a good relationship with the U.S. in [President] Reagan’s time, but it took a wrong turn during the [Bush] father and son era,” he said. “I saw the day in the 1950s when Iraqi youth would line up for information about America. Now what does it look like? The embassies all have guns.”

So my main point here is that I think we should be looking at how the current war between Russia and Ukraine can affect these misperceptions between Russia and NATO in the future.

We all talk about how Russia doesn't have red lines, and how the western politicians are gradually overcoming their mental barriers regarding what types of weapons can be provided to Ukraine and how Ukraine can use them against Russia. However, we should also be thinking about how this affects Russians perceptions and how it can cause misperception.

Russians are told by their government and media that they are fighting NATO, and they not only see NATO equipment fighting them in Ukraine, but now after the Kursk offensive also in their own territory (not to mention that many of them undoubtedly perceive occupied parts of Ukraine as Russian territory, but this still must be kind of next level of confrontation for them). And in a way, they are fighting with NATO, as NATO provides weapons, ammunition, intelligence, money, training and other supplies and support to Ukraine. So, I think for Russians and Russian soldiers and officers, armed confrontation with NATO soldiers may be easier to imagine and accept as an act, than before the full-scale invasion of Ukraine.
Would a kind of mirror-imaging cause them to think the same about NATO red lines as we now think about theirs? Would their own experience lead them to doubt the resolve of NATO countries to follow Article 5 over things like soldiers crossing Baltic country border? I don't know, but I believe we should be thinking about that.

I always look for new signals before I update of confirm my forecast, but in this case, I do not expect any new significant signals for months and likely years.

However, it may be very useful to create a list of indicators to watch. Lets start with macro factors before we may want to try creating the list of indicators on the smaller, more detailed scale (like Russian propaganda and rhetoric, efforts to destabilize Baltic States, troops build up etc). The key enabler for this to happen IMO would be some form of military collapse of Ukraine. Can something like that happen before April 2027? Unfortunately, I think it is something we shouldn't dismiss as a possibility. Ukraine has lost the core of their experience soldiers, and there are reports coming from the Ukrainian side that those soldiers who are now on the front lines are often poorly trained.

There are also other risk factors. I am not an expert on Systems Theory, but I see this conflict as a complex system in which a relatively slow rate of change is the result of either it not being far from equilibrium or the state where at least forces match each other. There are also forces working on both sides that prevent this situation from radical change (like the West mobilizing themselves to send more support quicker when the situation worsens for Ukraine). At the same time, this all depends on the lack of critical failures of these forces to balance the situation. If, for example, there is a critically low level of artillery ammunition and maybe so, that might start a cascading effect and produce a non-linear rate of change. Think about something like the fall of Afghanistan. I do not think this is likely in Ukraine, but we should not underestimate the complexity of this system. And complexity limits its predictability.

Even if by 2027 Russia is very successful in war with Ukraine, I still don't think they would go for confrontation with NATO unless there are some special circumstances limiting NATO effectiveness and especially the power and/or decisiveness of the US. What comes to my mind is some huge political crisis in the US or maybe the crisis caused by the Chinese invasion of Taiwan. This is of course very speculative, but coordinating moves between Russia and China in case of their decision to take Taiwan (with Russia, for example, invading Suwałki Gap) doesn't sound impossible to me.

I have also applied my ACTA Framework to analyze this question. I see source of  resolution ambiguity in the term "significant deployment of Russian military forces". It might be worth reversing this and asking ourselves what deployment of Russian forces to Baltic states would we call insignificant? Since the fact alone has that big significance, I don't think there is force small enough to call it insignificant in case of cross boarder deployment.

"The Russian forces must engage in offensive actions against the country's military, including armed clashes, bombardment of targets, or attempts to seize and hold territory within the country."

Red teaming this with possible scenarios, I think positive resolution may be triggered by hybrid operations with sabotage mission of little green men or military intelligence operatives, or for example some drone attack on the Gas Interconnection Poland–Lithuania (GIPL) pipeline.

I don't think some form of hybrid attack is that unimaginable given the situation, we have on the boarder between Poland and Belarus where Polish soldiers and boarder guards are constantly attacked by migrants with stones, slingshots (these slingshots are particularly powerful because they are stretched between trees, allowing them to launch metal balls with significant force), knives (one Polish soldier died in June from a knife attack by migrants and two boarder guards were wounded), tree branches with nails, stun guns... 

I also found this article: https://www.sciencedirect.com/science/article/abs/pii/S0030438723000224 Full text could be accessed via ResearchGate: https://www.researchgate.net/publication/371948393_The_Baltic_predicament_in_the_shadow_of_Russia's_war_in_Ukraine 

German tabloid Bild also discussed supposedly leaked document said to be produced by the German Ministry of Defense related to NATO exercise confronting the Russian invasion of Baltics States and Suwałki Gap in particular: https://x.com/EHunterChristie/status/1746698331395403987?lang=en
https://europeanconservative.com/articles/news/leaked-german-defense-document-sketches-out-russian-war-scenario/  
I don't trust tabloids, but it might be true.

Good point made by @DimaKlenchin here. IIRC we have discussed so far the following forces/factors which could in theory influence Microsoft to close the LAB:

1. pressure or order in a form of legislation from the US government to leave China (they can force them to do so as Dima mentioned in the comment I referenced above),

2. general risks associated with geopolitical tensions between the US and China, especially prospects of conflict around Taiwan,

3. the prospects of AI/AGI race and in general rivalry between the US and China in terms of research progress,

This article from the Economist highlights another potential factor making research LAB in China, an uncertain endeavor in the medium term:

At the same time, Mr Xi’s paranoid security apparatus is making it harder to move some intellectual property (ip) generated in China outside its borders, ignoring economically minded officials’ concomitant efforts to attract foreign investment. Exports of some AIs, such as speech- and text-recognition software or even TikTok’s recommendation algorithm, now need permission from the commerce ministry. So far this has not stopped most IP from leaving the country. But this may change at any moment, says Alex Roberts of Linklaters, a law firm. “We are at a tipping point.”

In anticipation of tougher American and Chinese IP regimes, some foreign companies have started to move research staff out of China. Even as Apple doubles down on its Chinese R&D, Microsoft, its big-tech rival, is said to be offering AI researchers in Beijing relocation packages as it winds down more sensitive R&D in China. Although AstraZeneca and Bayer remain ostensibly bullish on Chinese research, some pharmaceutical companies are seeking more clarity about the rules for cross-border transfer of data and IP, says an industry insider, and are rethinking new investments. The flying-taxi firms are staying put for now. One day they, too, may have no choice but to take flight.

At the same time, I do not think that China would want to force Microsoft to leave by blocking their transfer of research outside of China. In fact, I am not even sure how this would work. It is one thing to block transfer of technology from companies, universities and research centers which China controls, and an entirely different thing to control what Microsoft is doing in their Chinese LAB, not to mention being able to prevent them from communicating/sharing the results with the rest of Microsoft. Still, maybe this signals, that China's leadership is seeing the value of their research results and is not that happy to share some of it with the world (still, my impression is that they still share a lot, even in AI space), which may lead to implementing some other restrictions in the future. 

I think that it should be already resolved as of 1 or 2 August depending on the time zone based on this: https://www.infer-pub.com/comments/136163 and while I think that this should count I am not certain that the team will resolve it as such so I rise the probability based on the current tensions.