At the start of each year, I make a few predictions about where I think AI and the broader tech industry could be headed.
The future is far from being logically foreseeable though — and any prediction, if it is to be accurate, must be an act of faith rather than logic.
So this year, I’m calling them “hopes”. Instead of reading them as serious predictions, see the below as a collection of interests — things I’ll be focusing on and thinking about over the next 12 months.
Enjoy — and feel free to share your hopes for 2025 with me too!
1. A human-first Renaissance
I don’t know what the future of work will look like, but I know for sure it won’t happen by competing against the machines. Social, artistic, and technical progress will come from doubling down on what humans can uniquely do — taste, craft, experiences, storytelling, connection, and ideas that fall outside the distribution of what’s already been done. Combine these inherently human skills with the raw speed and power of machines, and you get the ultimate recipe for human augmentation. There’s an interesting human Renaissance coming.
2. Rise of the communities
The return of craft brings renewed opportunities for social connection — maker spaces, galleries, markets, and niche experiences centred around specific interests. While it’s never been easier to “find your crowd” online, I’m seeing a growing struggle to bridge the gap between online and in-person pursuits, leaving many in an in-between state of following their interests by proxy while never managing to materialise them in real life. In response, I’m also noticing specialist groups emerge where highly-curated members assemble to manifest an interest together — supper clubs, startup studios, design meetups, and much more. I expect this trend to accelerate in 2025.
3. Simulations all the way
The biggest technological shifts since the 70s have been either medical technologies or information technologies — largely, technologies of simulation. Beyond audio, video, and special effects, software itself is simulation — a virtual representation of people, markets, and ideas, amplified by the Internet and made ubiquitous with smartphones. Generative AI is accelerating this trend at a rate we couldn’t have foreseen only 3 years ago. In 2025 I’m hoping to see the full power of simulation at play in fields like education and medicine. I’m particularly bullish on tools that help us practice skills that once required heavy supervision (medical procedures, maintenance of machinery, but also soft skills like communication, planning, and teamwork).
4. Advent of the agentic swarm
Two years spent building with LLMs have taught me that “divide and conquer” really is the best way to build agentic workflows. Rather than having single, general-purpose models doing everything at once, we’re about to see many more pipelines made of small, specialised agents operating in unison as swarms — orchestrating work, breaking down tasks, voting on the best course of action, and behaving like independent “reasoning units”, each imbued with its own specialised knowledge (I’m using the term “reasoning” loosely here). Agentic swarms have powerful applications, notably in anything relating to research, decision-making, and large-scale simulation (echoing point 3).
5. The golden age of personalisation
Last year, we saw some early attempts at personal AIs — therapists, teachers, coaches. Most of them fell short on a few fronts: shallow wrappers, wrong interface, latency issues. But by far the biggest bottleneck was on the data front — companies working towards better models instead of leveraging the treasure trove of data they already had (Apple was the exception — they understood this early). Here’s the catch: personalisation is not an AI play, it’s a data play. Whoever can best understand, structure, and transform their data into actionable insights and serve those to users through a compelling UI wins. This, combined with some of the previous points (storytelling, community, craft), provides fertile grounds for state-of-the-art personalised tools in 2025.
There are so many more things to look forward to this year, and endless opportunities to learn and build. What a time to be alive!
Here’s to a fun year ahead — I hope you’re as excited as I am.
Until next time,
Nico