Predictions for 2025
I like making predictions, it’s fun, opens interesting topics of discussion with others, and if done seriously can be a great way to learn new things. For these reasons alone I think we should try to make more predictions about stuff, not transforming them into bets by adding money, but just to push ourselves to do more research, improve our critical thinking skills and boots our intuition. Adding money to them creates a whole new dimension that can be both very positive (skin in the game, amplifying research, …) and very negative (bad sizing, gambling addiction, insiders, …). So let’s some have fun with some prediction for the current year!
Phone and Social Media detox
I believe this year will see many people reduce their social media consumption, accelerating the dead internet theory and the dark forest hypothesis (applied to the internet): many more accounts going private, communities and sub-communities preferring closed systems (private chats, discord servers, gated social media, etc) and requiring more identification. I believe it’s a trend that started a few years ago but will become very apparent in 2025.
The number of minutes spent on social media worldwide had already plateaued after covid and went down for the first time in 2024, which seems pretty crazy to me given that it was the year with the most elections worldwide, including many important and contentious ones, and in 8 of the top 10 most populous countries, counting the United States, the European Union (27 countries), India, Brazil, Venezuela, the United Kingdom, and even an unanticipated one with the French legislatives elections, and many others calling around 4 billion voters worldwide. When you also include the Paris Olympics, it’s incredible that social media consumption went down from 2023, and a clear sign it will drop further in 2025. All this political content, alongside lower consumption leads me to think that, while spending less time on social media, many people ended up with a higher amount of junk content that can only fuel a deeper desire to leave.
It really sounds like people are realizing how bad and useless their social media consumption is, and to the point of actually taking action to reduce it. Anecdotally I have had friends talk openly about being addicted to their phones and short-form video content and using their phone’s built-in parental control to fix it. This used to be in the niche of productivity hackers, but is getting more traction outside. “Brainrot” became a thing in 2024 and gained popularity , same for the concept of dumb phones that picked up interest that same year.
I believe all of that will lead to a noticeable decrease in social media usage that will be reflected in key metrics (minutes spent on it, daily active users, lower non-bot engagement, increased digital marketing cost of acquiring user, …), tutorials on how to make dumb phones and maybe a rise in offline relationships.
The Year of Vietnam
I believe Vietnam is in a great spot for 2025. It has been doing very well over the past 10 years with near constant minimum GDP growth of 6% and 7% and coming back to pre-covid level, while having maintained inflation at around 3% and unemployment at 2.3%. This growth puts it at the top of its region, and I believe it will remain at the top in 2025 as its Foreign Direct Investments and exports keep growing quickly.
The last 30 years has seen Vietnam liberalizing more, and in the last few years kept crafting pro-business fiscal policies while vastly improving infrastructure, giving it China 2000 vibes. The state has been able to reduce its debt to GDP from 46% to 37% in the last 10 years, showing a reassuring level of government responsibility.
However, Vietnam has potential risks in 2025 with its biggest trade partners, US on exports and China on imports, but I think that if managed correctly (which the government has shown being fairly good at with covid management, reducing debt, liberalizing the economy) can become opportunities that are the reasons I’m making this prediction. The recent trouble in the Chinese economy, and the US-China tariff wars (that will only grow bigger with new Trump tariffs) have pushed many companies, such as Google, Adidas, Hana Micron (chips) and Samsung to move their investments to Vietnam. Vietnam really seems to appear like a perfect replacement of China manufacturing, at least on some parts of the supply chain, and is much less targeted by tariffs than China.
It remains to be seen in 2025 but I don’t really have a reason to think this new US administration will change course with Vietnam given that it’s “only” the 7th trade partner of the US. I am worried though about China’s decline and its impacts on Vietnam, my main intuition is that Vietnam overall profits from it (like it’s been happening in the last 10 years), but the consensus revolves around Vietnam being bottlenecked by its legal framework. Vietnam is just in a great position to keep growing on most fronts and I think all economic indicators will show that in 2025.
Financial Crash & Credit Event
The stock market is at a massive all-time high (S&P, NDAQ & MSCI World). A third of the S&P500 market capitalization is held in the top 7 which have P/E ratios ranging from 25 for Alphabet to 108 for Tesla with a ratio of 48 for the Nasdaq as a whole. None of that (or anything) predicts a crash, but it begs the question: for what? What is pushing these valuations so high? It feels like a lot of the bullish feeling of the last few years is based on a combination of money printing and a sense that the economy is doing well. When the Fed started raising interests rates the market started going down, inflation slowed and it seemed like the economy was doing better. The latter started being reflected in the market in 2024 which was turbocharged by the AI boom and an easing by the Fed.
When we look into these factors, they look pretty feeble. It’s true that the US economy got stronger in the last two years, but really only in the things we measure, and we might not be measuring the right things: sure consumer spending is high but so is consumer credit, sure unemployment is low but the number of self-employed Uber drivers and gig workers is high (and not always counted as unemployed when they themselves would consider so) and inflation slowed down but not necessarily in places that most hurt the wallet. It makes for a great picture of the economy that comforts the market but is not that pretty when you look at the details. Another factor is the AI boom that started with ChatGPT and pushed every big public company (including the top 7) to make AI announcement that fueled expectations. However the only thing that came out of this hype for now - from an earnings perspective - is Nvidia selling more graphics cards. Products for which people are paying aren’t profitable and due to the nature of AI (both in research and resources required) it’s not sure that it’ll be profitable in the form that the market is chasing. As for the Fed easing rates, it’s not clear at all that this will happen (at least significantly) in 2025 despite the new administration wanting rate cuts, the Fed actually hints at the opposite.
That’s for the state of the market at the end of 2024, but how about 2025? We can’t really be sure of anything, but we know that Trump policies will have an impact of the economy, and quite a lot (but not all) of these policies will cause prices to go up, two of which are tariffs and deportation. There is almost no case in which tariffs don’t increase prices directly or indirectly; directly by charging American importers (which pass on that cost to their customers) or indirectly by artificially making some American companies more competitive which immediately lets them increase their prices (and this is without counting the effects of other countries counter-tariffing). Deportation though, probably has an even worse impact on prices. The vast majority of illegal immigrants (and immigrants in general but I’ll focus on the first deportation targets) are of prime working age and work in understaffed low-paying industries like construction and agriculture; removing them automatically increases labor and supply chain costs, which will cause higher prices for consumers. It also destroys many “legal” jobs on top of not filling the understaffed ones: less people means less consumer demand (food, housing, transportation, education, etc) and less entrepreneurship as immigrants are statistically more likely to start new businesses. I’m not sure you could come up with more inflationist policies than this. Not all of the expected Trump policies are like this though. Energy liberation can very well decrease gas prices (and do many good things in terms of new jobs) but energy costs are just one aspect of the hard inflation that people feel, and I don’t think it’ll be enough to offset the other inflationist policies, particularly for food (and as more cases of the bird flu keep being discovered).
So I’m predicting more inflation in 2025 that will push the Fed to limit its cuts more than anticipated in 2024. I also think that inflation and these various policies will lead to a reduction in consumer spending that will be felt hard by US companies, and it’s one of the things that could lead to the crash.
Another prediction that goes with the market crash one, and is a big part of it, is credit defaults. Household debt is at an all-time high on every type (housing, credit card, auto) with no signs of stopping (waiting on the New York Fed data for Q4 2024). Looking at the details you see that credit card (CC) debt is rising the fastest. This kind of debt is mostly used for emergencies, groceries, utilities, retail purchases and vacation, and it’s the most expensive to pay back. What’s worrying though is that delinquency rates are also going up fast for CC debt. On the other hand, the GDP-to-debt ratio is at an all-time low, but this could just confirm that far from everyone is benefitting from the strong economy. While the New York Fed data shows a slower transition to delinquency, I believe 2025 will see that rate go up quicker. I don't see how it gets better for people who are using CC for everyday expenses in 2025 without the economy booming.
Scientific Breakthrough
I think a great source of optimism in 2025 will be coming from science. I am sure we will see great advances with GLP-1 research outside of obesity, particularly in aging-related diseases, powered by the success of Ozempic.
Probably the most positive consequence of covid has been mRNA research that is now fueling so much research in vaccines, including against cancer. The first big trial for a cancer vaccine has started in 2024 and any new positive developments in 2025 would be such a win for humanity.
This prediction also includes some big leap in digital twin research, maybe a lot of new patients getting surgeries by “simulation-assisted” surgeons. I’m also eager to see other trials for transplants of genetically-modified animals’ organs to humans. Overall I think it will be an incredible year for longevity and healthspan.
Agentic AI, but still not profitable
I anticipate we’ll hear a lot about agentic AIs. Bots that can complete tasks for us. It’s the next logical step after LLMs: connecting them to APIs that have real world effect, not just returning a text answer. Although we might see some cool use-cases, I don’t think AI companies will become profitable in 2025. Sure, given the R&D expenses it’s not a very bold prediction, but even without counting research costs I still think they’ll remain unprofitable, as in running their core product offerings will still cost more than the revenue, and won’t get cheaper. This could also lead this year to be The Year of Open-Source AI, as I think that whatever interesting and cool use-cases emerge out of AI agents will most likely be runnable by open-source AI and frameworks.
The Year of Distributed Social Networks
This prediction seems like it clashes with the one about decreased social media usage in 2025, but on the contrary I think they go together and might even fuel each other. I believe one of the most important trends on the Web as a whole is the shift from the implicit (forced) compartmentalization of users to an explicit (opt-in) one. I mean that over time social media companies have perfected their bubble-building algorithms to maximize user engagement. These algorithms put users in implicit bubbles, meaning that their frontpage appears like an open window to the world, when in reality it’s just very highly curated content whose sole objective (excluding state-sponsored actors) is keeping the user engaged; whatever that engagement is: a good deal of reinforcing already held political ideas, a cup of opposing views tossed in there (usually by a political ally “exposing” enemies), a teaspoon of intra-bubble drama, etc. This was “forced” upon users (in quotes because people could always just opt-out a long time ago) because it works. People like to be in bubbles and like to be validated constantly, but combine that with an increasing number of bots (AI, useful idiots, state actors), an incredibly polarized political scene probably hiding a deep sense of fatigue for everyone (how long can you last being always on edge?) and the fallout of Twitter leading to many competitors and an interest for decentralization (ActivityPub, AT Protocol, Farcaster), I think you end up with the perfect recipe to push people towards their own bubbles by themselves.
These social media alternatives are still either too new to have developed an actual federated/distributed network that goes beyond the main central actor behind the protocol (AT Proto doesn’t carry on without Bluesky currently, and the same is true for Farcaster with Warpcast) or too complicated to navigate (Mastodon). However, my prediction for 2025 is that some of these services will see significantly more usage, and not just on the main client. They'll have new clients focusing on different social media interactions and gating access in various ways that target bot-reduction and/or fostering a certain community (political, hobbyist, work, ...).