F2F #17: Predictions for 2025

Happy New Year to you, my friend.

I have always been an admirer of the Nostradamus effect, whereby you can make a lot of predictions and crazy wild guesses because statistically speaking, you will always get one or two right. This is the same as "The Simpsons predicted it".

That is yet another manifestation of survivorship bias. But hey, I can't help it if people are fucking stupid.

We tend to believe we're good predictors due to a combination of psychological factors and cognitive biases that influence our perception of our own abilities. Mostly, Confirmation Bias and Dunning-Kruger Effect are at work, but there are other biases.

I struggle to separate what I want to happen from what I think that will happen but there's worse: most people don't know the difference between the two.

Every year, on January 1st I receive a birthday email I've written a year before (usually during the first days of the year), with a quick recap of the previous year, a general status update of how are things with me/family/business/friends and some predictions for the year ahead. While I couldn't write any scientific paper on them, I find it to be a good exercise to help me predict stuff I want to happen but also stuff I do not want to happen but I think might happen anyway.

Going back to the Nostradamus effect (or Diario Sport with Barça signings, if you're into football), take this twitter thread as an example:

No one will remember this thread one year from now. Not even his own author if he doesn't get anything right... but he threw in a wild amount of predictions - some wilder, some more reasonable - so statistically speaking, it's safe to assume that he'll get some right. If he somehow made the right call on a crazy bet, he will go back to this tweet and tell us all "I told ya".

Chances are, this won't happen.

Anyways, here are some of my predictions for 2025. Let's play this game.

#1: We finally understand that AI's bigger risk isn't Skynet killing us all. AGI is not coming anytime soon because we've got a HUGE problem to solve: hallucinations. AI's biggest problems are data security, ethics, regulation, cost of energy and intellectual property, mainly, but there are more.

#2: The verticalisation of these problems will create the next must-watch markets. This is a second-degree rippling effect of the AI boom. I am excited to invest in companies solving the aforementioned problems and I predict that one of the next big companies to knock it out of the park will be a services business, not a product one.

#3: The big ones (Apple, Amazon, Facebook, Google) will still be big. Nvidia will continue its good run because it's too early for any other company to catch up with everything they've done but I somehow suspect that their narrow focus on AI is forcing them to forgo other business opportunities that someone else will take.

#4: Last year, I predicted that Musk would get rid of Twitter. Recent happenings, like the interference in the US election, prove the point that it's an extremely valuable business, even if its architecture and product is in shambles: it has a massive global impact. Goddamn it: it got Trump re-elected and is the closest thing I've seen to the all-seeing eye of 1984. Twitter will start a hard pivot towards the end of this year to something else, raising a ton of money in the process and bouncing back in valuation. My guess: real-time AI analytics on what is happening right now.

#5: Social media will continue its stark decline. The proliferation of bots, spam, scam, and the general feeling that "the algorithm" is controlling us all is permeating into society. Distrust is at an all-time high, but it's not yet enough. The pendulum of history is swinging back to ourselves after the period of "live for others". The new trends will focus on meditation, self-reflection, journaling, improving your own health, wealth and such.

#6: Proof of humanity becomes more widely accepted. By now, we all know how to tell if a text has been written by an AI or a photo has been edited by an AI. In fact, today I was trying Apple's "Clean up" feature to remove photobombs from a few of my photos, but ended up discarding all the changes. Two years ago, this would've been the best thing ever since sliced bread. Now, I see the imperfections caused by AI's cheap inpainting. Screw that, give me the photobombs.

#7: No one will build HER. I have been fantasizing with the idea of having the perfect virtual assistant. In fact, I use a lot of AI combined with a personal assistant and somehow it all feels clumsy: every tool needs to be reminded of the context, updated, upgraded and constantly configured using more or less sophisticated text files with instructions and directives, but HER's AI is an omniscient all-seeing entity whose ethical and legal implications far exceed the technical limitations we could ever encounter.

#8: Boom in digital nomads & solopreneurs. Every year, the cost (in difficulty, not in price) of writing code keeps going down. Now, even non-technical people can work it out - just not everyone. But this virtually-everyone-tech-savvy-build-apps will enable and inspire even more people to try it. While I find hard to believe that we will have a one-person unicorn, I think that the big win is on most people finding that they can solve more niche problems and have very good businesses thanks to AI and automatisation and enrich the fabric of SMBs instead to taking the wild bet of VCs and building unicorns. Sort of a "think big, build small".

#9: A new batch of boutique agencies is born. The pandemic and the tech decline of '22 & '23 wiped away thousands of small agencies: those that overhired, the ones that had all the eggs in one basket and those crushed by the economic downturns & bad decisions. We need the MarsBaseds of today because they will thrive on the projects we discard, being able to be more flexible with pricing, delivery dates and - more importantly - usage of AI. AI-native boutiques, we could call them.

#10: Boom in M&A. After two years of all-time-low M&A activity, we will see a surge in quantity of deals, but not in quality. There are a lot of companies in distress being sold right now, and the trend will continue, as a lot of scale-ups will see that the only way to get investment is through partial sales to private equity firms or industrial players, as VC.

Now, I'd like to do a few more regarding my own projects, for social accountability:

• MarsBased will hit USD 3MM in annual revenue.
• I manage to grow the Life on Mars podcast by doing one of these two options: either I accept that we do only live podcasts in Barcelona with local people and we get a great production setup, or I shut down video and focus on audio only, so I can continue having great guests from all over the world.
• Foc a Terra becomes a video podcast and we find a great sponsor (besides Parlem) to take it to the next level. We appear on mainstream radio before the end of the year.
• I will have one small exit as an angel investor (2-3x), a bigger one (around 5-6x) and two write-offs.
• I will have close to 1000 subscribers of this newsletter before the end of the year with around 5 to 6% paid members.
• I will have launched a side project but it won't be monetisable. Or, at least, I will choose not to monetise it for the time being.
• I won't join the scientology cult:

Anyways, I wanted to wrap it up by thanking you all for reading me and keeping the great metrics. Hundreds of people follow this newsletter now with an average open rate of 72%, which is WILD and far exceeds my expectations.

Since this past week has been about family and celebrations, I haven't got much to share in terms of articles, podcast episodes and so on, so I will leave it for next week.

Bonus points if you can answer to this email enumerating the biases I've fallen prey to in each one of my predictions.

Have a lovely start of 2025 and let me know how can I help you back, anytime. You absolutely deserve a fantastic year.