geoffodlum

Geoff Odlum
About:
Show more

No Scores Yet

Relative Brier Score

18

Forecasts

7

Upvotes
Forecasting Activity
Forecasting Calendar
 

Past Week Past Month Past Year This Season All Time
Forecasts 3 17 170 18 679
Comments 3 11 44 12 336
Questions Forecasted 3 15 33 15 96
Upvotes on Comments By This User 1 7 55 7 350
 Definitions
New Badge
geoffodlum
earned a new badge:

Active Forecaster

New Prediction
geoffodlum
made their 13th forecast (view all):
Probability
Answer
Forecast Window
45% (+30%)
Yes
Feb 4, 2025 to Aug 4, 2025
55% (-30%)
No
Feb 4, 2025 to Aug 4, 2025
Why do you think you're right?

I am raising significantly to just under 50%.  Netanyahu is in DC this week for meetings with Trump, and  an Israeli-Saudi normalization deal is something Trump wants to move fast on. As long as the Gaza ceasefire holds and hostage releases continue, Trump has enough sway with Bibi and MBS to force an inital agreement, though there are alot of devils in the details to get to a final agreement by August 4. 

https://www.atlanticcouncil.org/blogs/new-atlanticist/israeli-saudi-normalization-is-within-reach-heres-what-trump-and-netanyahu-need-to-do-next/

Files
Why might you be wrong?

So many thngs have to go right, and just one thing going wrong (which is a daily ocurrence in the Middle East) could tank this deal

Files
New Prediction
geoffodlum
made their 6th forecast (view all):
Probability
Answer
Forecast Window
20% (-5%)
Yes
Jan 29, 2025 to Jan 29, 2026
80% (+5%)
No
Jan 29, 2025 to Jan 29, 2026
Why do you think you're right?

I am still higher than the conventional wisdom, though lowering somewhat. With Trump in office, Iran's margin for error if it acts aggressively in the region or with regard to its nuclear program is much smaller than it was when Biden was in office. Since Netanyahu demonstrated the effectiveness of a surprise leadership decapitation strike (against Hezbollah), I can imagine Bibi egging Trump on to do the same with Khamanei, even though the Supreme Leader is an order of magnitude harder target to hit. 

Files
Why might you be wrong?

The status quo in Iran's leadership tends to drag on and on.  

Files
LogicCurve
made a comment:
I'm forecasting higher on this also, as change can happen suddenly sometimes and the Iranian regime is really pushing it right now.  
Files
New Badge
geoffodlum
earned a new badge:

Star Commenter - Jan 2025

Earned for making 5+ comments in a month (rationales not included).
New Prediction
geoffodlum
made their 2nd forecast (view all):
Probability
Answer
15% (-80%)
Chemical
3% (0%)
Biological
3% (0%)
Radiological
0% (0%)
Nuclear
Why do you think you're right?

Revising based on discussion today (1/31) with the RAND TASP team, who clarified that Russia's confirmed use of CS gas as a weapon of war on the Ukrainian battlefield -- though it is a violation of the CWC -- does not lead to a "yes" resolution, b/c CS gas is not explicitly listed on a CWC Schedule of controlled chemicals and precursors.

As a result I am lowering my earlier forecast of 95% likelihood of CW use down to 15%. Chemical weapons still seem to me to be the most likely CBRN weapon to used by a state actor in a conflict given the ease of access to Scheduled chemicals and the dificulty the OPCW faces in attribution.  

Files
Why might you be wrong?

A currently responsibly-acting state actor could face an unexpected revolution or change in leadership or governance system in the coming 5 years that results in a "rogue" government taking control which does not feel bound by NPT, BWC, or CWC obligations.  

Files
New Prediction
Why do you think you're right?

If chemical synthesis techniques advance in a linear slope over the next five years, I think it is unlikely that researchers will accumulate the breakthroughs necessary to use automated chemical synthesis to produce a protein of 400 aminos in length. 

However, I believe there is a significant likelihood (slightly under 50%) that AI models and agents advancing at an exponential rate and operating at greater than human intelligence in these specific domains over the next 2-4 years will discover breakthroughs in knowledge and techniques that will allow researchers in a lab setting to active this outcome.

Files
Why might you be wrong?

I may be overly optimistc or amitious in my forecast of AI advancing at an exponential rate over the next seeveral years and achieving greater than human intelligence in these domains. 

Files
New Prediction
Why do you think you're right?

As always I appreciate the comments and insightful anaysis from colleagues inc @DimaKlenchin @ctsats @DKC and others.  I fully accept the argument that this specific resolution criteria -- "The product must be designed to replace all major components of whole blood (red blood cells, platelets, and plasma) or provide equivalent functionality" -- currently seems impossible to achieve in 6 years, or even 66 years.

The one cautionary caveat I think should be applied here -- in addition to the general caveat that nothing is ever 100% or 0% likely -- is the possibility of currently unforeseeable and unimaginable breakthroughs in biotechnology, chemistry, genetics, nanotechnology, and medical science that will occur when artificial general intelligence is broadly achieved in the next 1-3 years, and especially when artificial superintelligence is achieved a year or two after that.  Such an occurrence might actually render this question moot, as we might not even have an FDA as we currently know it by 2030.  (although in the event there is no longer an FDA by 2030, I guess that would result in a negative outcome to the question, which hinges on FDA approval...)  The point is that radical scientific and medical breakthroughs will be occurring at exponential rates in the coming years, and what seems impossible now may be achievable in just a few years.  Its enough of a possibility for me to feel safe forecasting  this outcome at 5%. 

Files
Why might you be wrong?

I totally accept that AGI/ASI timelines are wildly speculative, AND that even if such breakthroughs occur, there is a decent chance that FDA clinical trials will continue to reflect pre-AGI safety requirements and testing protocols, which move along a sufficiently slow timeline to ensure this question resolves "no". 

Files
New Prediction
Why do you think you're right?

I endorse @DKC s and other colleagues' analysis that under normal circumstances, any BCI with indications for use in able-bodied individuals winning FDA approval would extremely unlikely by 12/21/2030.  However, as the past week has demonstrated, we are not living in normal times.

The possibility that President Trump could, at Elon Musk's request, simpy order the FDA to grant full approval to Neuralink whether or not Neuralink's BCI have passed all clinical trials is, unfortunately, not insignificant.  Even though such a Presidential  order would be unprecedented, unethical and likely harmful, it would be entirely consistent with how Trump has demonstrated so far that he intends to "govern."  As long as Musk has Trump's ear -- which I believe is the primary driver that will determine the outcome of this question --  I put this possibilty at between 15-20%.  

Files
Why might you be wrong?

Musk may realize that asking Trump to order FDA approval for Neuralink's use on healthy individuals w/o the necessary clinical trials, while achievable politically, would still leave Neuralink catastophically vulnerable to class action lawsuits for the harm it causes to patients.  Or, even if Musk pushes for an approval order, Trump may feel doing so is a step too far from a legal and reputational standpoint (though frankly so was pardoning all J6 insurrectionists, profiting off of personal crypto coins while serving as President, etc).  

Files
New Prediction
geoffodlum
made their 4th forecast (view all):
Probability
Answer
6% (0%)
Moldova
2% (0%)
Armenia
8% (0%)
Georgia
1% (0%)
Kazakhstan
Confirmed previous forecast
Files
New Prediction
Confirmed previous forecast
Files
Files
Tip: Mention someone by typing @username