efosong

Very much outside my wheelhouse, but going off a few things:

• I used the arc analysis AI tool to search for and summarise research. It was pretty optimistic: "In conclusion, the convergence of computational biology, automated synthesis technologies, and market demand indicates a strong likelihood that proteins of at least 400 amino acids in length will be successfully produced in laboratory settings by the end of 2030." But as with all AI analysis tools: caveat emptor
• Protein synthesis seems like the kind of tech where people like to build capabilities, and then let other people build on top of those capabilities. Particularly when the ability to predict certain properties of proteins with AI has recently improved. If this is true, we might see sustained innovation and development in this area.
• Naively, someone making a big deal about getting to 164 makes it seem like there is some pull towards developing the ability to create longer proteins. I'd guess that 400 within 6 years isn't completely implausible.
• My AI-assisted search didn't pull up anyone surpassing 164. Given that 5 years have passed (though some of that includes coronavirus disruptions)
  

I'd probably go higher of gut feeling, but I must defer strongly to the crowd on this question, since I'm quite ignorant on the topic.

I think the crowd is too high on most of these, but perhaps I'm being overconfident. My main thought is that the base rates for these (except chemical) are very low, and the incentives also broadly point against these things being used.

Chemical

Since WW2 there have been a few uses of chemical weapons by state actors in armed conflicts, according to Wikipedia:

• Egypt in the 1960s
  
• Iraq in the 1980s
• Cuba in Angola in 1975
• Syria in the 2010s

Several states have used white phosphorous munitions, which some consider to be chemical weapons, but others (including the OPCW) do not, so I'm not counting their use towards this question.

Russia have recently been accused by the US and UK of using chemical weapons (including the explicitly listed chloropicrin) in the ongoing Russia-Ukraine conflict. The OPCW has already concluded that Russia has been using CS gas. I'm not sure why they haven't come to a conclusion on the use on chloropicrin, but assuming the conflict continues, it may be found. Though to complicate matters, the head of the Russian CBRN forces was recently assassinated — this may lead to a change in tactics, and it's not clear in which direction.

Sudan have also been recently accused of using chlorine, though I can only find one news story reporting this, and the reports seem to be 'unofficial'.

On base rates, ignoring Russia-Ukraine and Sudan for now: an upper base rate would be 40% (basically looking back and counting every year since 1945 when this would've resolved 'Yes'). A slightly lower base rate of 26% is obtained by counting at the 'conflict' level and modelling with a Poisson distribution.

I think we should be on the lower end of this now though: the CWC is in place, and if we are to believe the OPCW, many countries don't even have chemical weapons to deploy, which was less true in the 70s/80s. I'd then go for around 22% of usage outside the current Russia-Ukraine conflict.

Taking the Russia-Ukraine conflict into account: I don't know if the OPCW are investigating Russia for use of chloropicrin. If they are, I don't know how likely it is that they find something. I'll therefore just add another 33%p to my forecast. And maybe another 5%p for Sudan.

Biological

There have been few confirmed uses of biological weapons since 1945. The clearest example is probably in the Rhodesian war. Israel also used biological weapons in 1948. A few countries have accused the US of using biological weapons in the 50s/60s, but they're a bit tricky to confirm now.

This would give us a base rate of about 10% — but both of these incidents happened a long time ago, and mostly before the Biological Weapons Convention. Most countries are now part of this. Even North Korea is signed-up. Of those not signed-up, Israel and Egypt are the only ones I think could plausibly have offensive bioweapon programmes. Some of those countries who are signed-up may also have clandestine programmes. But in general — I don't expect any country to actually use bioweapons in the next 6 years, even under fairly extreme (but perhaps not absolutely existential) pressure. While there may still be some 'flexibility' regarding the taboo on chemical weapons, biological weapons to me still seem beyond the pale even for countries often considered 'rogue states'.

I therefore expect it to be extremely unlikely biological weapons are deployed in the next 6 years. I'm going in at 3%. My gut instinct is perhaps even lower, but I don't want to diverge too far from the 10% base rate.

Radiological

I find the use of these weapons extremely unlikely. They seem to serve little tactically or strategically useful purpose outside of total war. As far as I can tell, nobody has actually ever used a radiological weapon in warfare. Such weapons also require access to certain materials, which are presumably harder to get hold of than those required to make chemical or radiological weapons. I'll reserve a 1% probability to err on the side of caution (e.g. account for a very small chance of Russia doing something to one of the Ukrainian nuclear power plants, or a total war between Iran and Israel). I don't think we'd see radiological weapons deployed even in a Taiwan-related conflict.

Nuclear

Of course, nuclear weapons have not been used in conflict since 1945, so we're starting from a pretty low base rate. Famously, we've been close a couple of times though. I don't think at this point that Russia are likely to use nuclear weapons in Ukraine. Outside of that, I don't really see any nuclear conflicts brewing unless something dramatic happens in Taiwan over the next 6 years.

Reserving a couple of percent for the usage of tactical nuclear weapons, but broadly I think the nuclear taboo is likely to hold.

I think this is unlikely within a 5-year timeframe — particularly with the condition that the BCI must be implanted.

With a baseline scenario of "The FDA operates the same way it has in the past", I expect <2% that this will be approved for able-bodied individuals within 5 years. Mainly because the studies for BCIs for treating serious conditions are expected to take years, and I expect the bar will be even higher for devices with no (or less) clinical benefit.

We do, however, have the complication that the new administration seems quite hostile to the FDA and other regulations, so perhaps we can imagine an FDA which is influenced a lot more by politicians.

• On one side, you have Elon Musk who founded Neuralink and therefore is likely to apply pressure specifically to get BCIs to market. Though even Neuralink is focusing on blindness and quadraplegia at the moment, and I imagine they'll find it both more profitable and less backlash-inducing if they continue to focus on those with clear conditions in mind until the public becomes comfortable with the idea of BCIs.
• On the other side, you have many within the new administration who are sceptical of various established medical interventions — I imagine they would find allowing brain chips outside of extreme circumstances to be much more objectionable. I think actually the set of people who find BCIs for non-medical uses to be objectionable is quite big, and I expect a strong backlash to these devices (including protests and political campaigns to ban them) if they are brought to market within the next ~5 years.

I'd tentatively apply about a 30% chance to "FDA deregulation allows quicker approval of Neuralink products during the Trump term", but only 25% chance that anyone actually successfully gets a BCI approved for able-bodied people.

Finally, I think there's some chance of an arms-race dynamic with other countries: in particular, China. Tricky to tell how much of this is just sensational reporting, but there are claims that Chinese BCI researchers are working on BCI for cognitive enhancement. I actually think it's likely that this tech takes >5 years to get to the point that it even works, but I reserve a 1% chance that both this happens, and the US administration are sufficiently spooked that they pressure the FDA to fast-track something.

No real  big changes in outlook for these countries. Kenya and Nigeria have positive outlooks at Moody's, and time has passed since my previous forecast, so I'm shaving a point off Kenya and Nigeria.

For Ethiopia, I'm simply coming in line with the crowd as a 'score conserving' move, as I'm still unsure how the present default status of Ethiopia factors in this question.

Confirmed previous forecast

Sticking fairly low due to the general deescalations in the Middle East, and the (partial) neutering of various Iranian proxy forces. Doesn't seem in Iran's interest to get involved in anything they call a "war" over the next year. I also doubt we'll see aggressions from other countries that will escalate to a war over the next year.

Relaxing back towards mortality baseline.

Actually, I'm making an ad-hoc adjustment to my previous forecast. It seems like the DeepSeek news is having an impact on the market. Via the Financial Times (subscription required): Tech stocks tumble as China’s DeepSeek sows doubts about AI spending. I think the market is picking up on a real signal here.

The three models in my ensemble are giving quite different results: (Yes 48%, 53%, 64%). I'll err on the lower side of this for a couple of reasons:

• I think the "DeepSeek moment" could be very consequential for NVDA stock: at least, in the medium term. This article makes a detailed bear case for NVDA on these (and other technical) grounds. The question becomes whether the market will agree and price this in over the next few months. The stock market famously has a magical 'fundamentals-defying' ability (at least in the short-term.
  
• I think straying too far from a 1/3 prior would be unwise. NVDA is ahead at the moment, and the market is priced to expect higher volatility in NVDA stock, which is reflected in a higher probability of ending up as the largest cap (despite only being marginally ahead at the moment).