Reviewing 2023 Predictions
Maintaining intellectual honesty
All good machine learning practice starts with checking the performance of your predictions. In some cases you might measure accuracy, in others you might use a different metric. Either way, it's important to understand how good your model is. Similarly, being able to make better personal predictions requires reviewing how well your predictions turned out to be. I made a bunch of predictions right before 2023 started. In this article, I'm going back to see how my predictions for the year stacked up. Here's what they were and how they turned out.
Increasing number of sanctions on China: I predict the US to come out with several anti-China policies and sanctions to put economic and technological pressure on China in the hopes of staving off a war.
Yep. The main sanction the US placed on China was preventing the export of GPU chips and other AI technologies. You can find other sanctions here.
Onshoring in Mexico: As tensions rise between the US and China and people realize the lack of robustness in some of the current supply chain setups, I anticipate more manufacturing moving to Mexico instead of China where labor costs are lower and there is less shipping risk.
Yea this happened. Mexico surpassed China as America's largest trade partner from nearshoring. You can see a breakdown of some of that here.
Major breakthroughs in biotech due to AI: As AI becomes embedded into everything, most people will be focused on chat assistants. However, I predict that under the radar will be major biotech breakthroughs occurring due to the use of AI. This may be new devices, molecules, delivery systems, or discovery of mechanisms of action. Given the current existence of AlphaFold, I would expect no fewer than 3 major breakthroughs.
Yes, and expect more breakthroughs to keep coming. AI has discovered two new antibiotics. It was also used to create proteins that have never existed and kicked off a wave of protein engineering. AI was also used to identify drugs that could be repurposed for rare diseases.
VC Generative AI Hype: Based on the current excitement around ChatGPT, I believe that generative AI startups will have the largest amount of VC funding in 2023. This will ultimately lead to a bubble in the space that will burst in 12-24 months because most startups will not be positioned correctly.
Yep. Not surprising either. About a quarter of all funding in 2023 went to generative AI per DealRoom.
Generative AI will cause a political disaster: Due to the widespread use of generative AI, I predict that there will be a political disaster caused by the use of generative AI. This is likely some form of false content that most people are led to believe is true.
Not really. I don't think we saw any true political disasters this year from generative AI. The GOP created the first image generative AI political ad. Putin spoke to a deepfake version of himself. However, no one launched missiles, initiated a trade war, created sanctions, rioted, or was arrested due to a generative AI message.
Regulation clampdowns on AI: There is a growing feeling of concern around the power of big tech and AI within government. I believe this fear will cause a regulatory crackdown on AI and tech in general. I’m not sure what that means for the ecosystem yet but it will result in the need for a zero trust framework for AI. WeChat has already banned the use of ChatGPT on its platform.
Definitely. The Biden administration issued an executive order on AI and the EU signed their AI act.
Increase in happiness due to AI: I think the happiness rating of many countries that have access to advanced generative AI systems will increase as people find ways to use AI to provide companionship, inspiration, and remove annoying aspects of work.
Mostly. Workers who used AI were happier and less likely to quit. I think this will take longer to really see after we get over the polarized content of AI fear vs AI salvation.
Extreme weather event: Due to the increasing climate volatility, I predict there will be an extreme heat event that results in the deaths of thousands of individuals. This will most likely occur in the Global South
No. While there were extreme heat waves all over the world this year, with July being the hottest month in 120,000 years, there was not an excessive number of deaths. However, the deadliest weather event this year, Storm Daniel, caused record flooding which led to over 4,000 deaths. Storm Daniel was the deadliest Mediterranean tropical-like cyclone in recorded history.
Shift from lowering emissions to focus on adaptation and resiliency: I predict that this year there will be a realization that without fast action, we will not be able to hit the 2 degree Celsius limit of the Paris Agreement. This inability to hit the target will shift focus from emission reduction technology to those that enable adaptation and resiliency. While I do not hope for this outcome, I believe it is more likely than not.
Not really. While one of the big issues at COP28 was adaptation, there was no real plan for how to fund it. COP28 resulted in an adaptation framework, but it focuses more on intention than action. The bigger focus was on shifting away from fossil fuels.
Rogue Geoengineering: With people becoming more desperate to try to stop climate change, I think there will be individuals not associated with a government that start live experimenting with geoengineering. Neal Stephenson illustrated a potential scenario for this in his book Termination Shock.
Surprisingly this happened. The startup 'Making Sunsets' launched sulfur into the air in Mexico in attempt to offset climate change and sell cooling credits. The test was unsanctioned and was quickly halted by the Mexican government. The US government also just released a federally mandated report on solar geoengineering, which is uses artificial methods to reflect sunlight away from the earth to cool the atmosphere.
Overall not bad at 7/10 predictions. It's important to note that the size of the impact can be more important than frequency of success when assessing predictions. Maybe some of these were surprising to you, maybe they weren't. Either way I hope they were helpful for how you thought about the year.
Here's the five themes I saw that influenced the predictions above. Since they aren't as binary you can make your own assessment on their role and efficacy.
Growing isolationism
Rising tensions between US and China
Increasing climate volatility and miscalculated risk
AI embedding into everything
Greater focus on energy
As I've said previously, reflecting and being held accountable for predictions improves your decision making over the long term. I hope this inspires others to publicly review predictions they've made.