2026-01-26 - The predictions game for 2026
Disclaimer
Prospective thinking has always been something I cared a lot about, I liked to read about it, to follow the analysis made by specialists, from renowned think-tank analyses to fancy consulting reports. As a hobbyist, I am more into predictions, which is a bit less serious, more like a gamble. These predictions are my bets for 2026, I am not sure to be original, and I have been inspired by many different sources. Not sure I will be able to mention them, maybe they won’t even be related to my poor analysis :) Anyway, thanks to them!
About technology
GenAI frameworks for the rest of the world
In 2025, I went from amused to impressed by GenAI, mostly after starting to use Claude Code with proper guidance. It was a game changer for me, for some other devs too, and mostly because it liberated us from the chatty interface and gave us many options to build and automate tasks (call it agentic if you want). In 2026, this should spread to many other industries (legal, medical, marketing, …), the bricks are here, we just need to design the framework (prompts, skills, MCP connections, MD files) related to specific businesses. For better and for worse.
Semantic Web, the revenge
The semantic web concepts have been defined for more than 30 years, it was a really promising approach for interoperability and decentralisation, but as you may have noticed, big actors have focused more on data locking and monopoly, still wondering why… Anyway, the emergence of GenAI and its thirst for structured data may bring the semantic web back into the spotlight. Also, interoperability can be easily managed by GenAI with clear API definitions, and semantic search can save tokens. If you are seeing the words ontology and triple, it should be it.
Cyberattacks, the industrialisation phase
Again related to AI, and from two sides. First, the automation of data analysis, especially codebases, gives a new way to detect code vulnerabilities. Logs analysis allows also to detect suspicious behaviors. Second, the power of GenAI on the attack side just allows to explore many more ideas than before. For example, you can also create agents that react directly once a vulnerability is found, and then try to attack any system not yet updated. You can also automate fuzzing at an unimaginable level, breaking software at high speed. Then, you can add to this a bit of Supply Chain nightmare from open source repositories, with teams getting flooded by AI-generated PRs. I don’t want to be pessimistic, but I think that the Blue teams will need a lot of support this year. Also, not to spoil the mood, I will not mention the new phishing techniques, involving for example voice messages generated from public content and based on AI profiling.
About society
Technofatigue
You want to lower your anxiety level easily, shut down your mobile phone. You want to sleep well at night, read a book instead of binging a series. You want to feel alive again, cancel your social network account and go talk to your friends. Technology is useful, but it cannot solve all our problems, it is also the source of new problems. A lot of people are choosing alternatives far from silicon, it does not have to be radical, but regaining that sense of choice is what helps them. Coming back to the Stone Age is not the point, but seeing how much space the digital has taken in our life, we could rebalance it a bit, especially regarding our passive/active usage. So this trend can grow this year, and could even become more and more a political question, I remember even an article talking about neo-Luddism, come on comrades!
Safe communities
Something is broken in the realm of information. Social networks are becoming mass-diffusion ad platforms, enshittification is polluting the user experience and people started getting bored of it, trust has been broken many times. They want safe exchanges, with a quality-over-quantity focus, and the possibility to tune the way they are exposed to the flow of content, to adjust the noise-versus-signal ratio. Technically, this can be achieved with algorithm control, mesh networks and non-violent communication, but the core issue is not really technical. The objective is to recreate trust amongst communities, and bottom-up approaches are the way. People want control and respect, it is not new, but I think it can become more structured and open in 2026. I am seeing a lot of people opening groups on Discord or Telegram. Historically, it was maybe something flagged with lack of openness or sectarism, it can still be, but we all saw what free speech without respect did to a social network like Twitter, it got consumed into spits and hatred. People want to feel at ease online, with quality exchange and not just comments or likes.
A financial crack related to insurance sector
Many people see the next financial bubble on AI, it could be, it should be, but it is really consensual, so for once in this paper, let’s not talk about AI and take a bit more risk. Talking about risk, insurance is a real concern here: More ecological disasters, more wars, more vibe coding crashing big tech services, … Someone will need to pay to hedge against this, but in the age of YOLO politics and growth-only economics, who will? And you can trust insurance companies for garbage financial products like subprimes.
About politics and geopolitics
The USA trust issue
USA has started the year off the rails… I have no idea where they are heading, but the Trump administration is clearly doing heavy damage to the image of the country as a trustworthy partner. The shift from soft power to hard power has been brutal and ugly. Whatever will happen in the midterm elections in October internally, and it could go from a Trump impeachment to a big step toward a new civil war, but many historical partners started to turn away from the USA and seek different alliances. This move has started even before Trump, the egemony cost is to high for the USA, and with poor results (eg: Afghanistan war). The old recipes cannot work indefinitively in a world evolving really fast, it is just history at work.
Energy policy
One of the big changes happening globally is the shift toward energy sources. USA keeps on oil more than ever, while China is now pushing hard on electrification. This is of course related to their resources and strategy: USA is the superpower of oil, from its partnership with Saudi Arabia to the fracking technique, and China takes advantage of its dominant position in solar panel and battery industries. They are both ecological nightmares, but at a geopolitical level, renewable energy is a good bet as it diversifies the energy sources, spreads the sources over a grid instead of few plants, and gives you more autonomy than oil partnerships. And many countries starts understanding it. Pakistan is an example of such a move, even if it can rise other issues.
Conclusion
Trust is the keyword for me in 2026, I am not taking many risks here, but it is really the connection between most my predictions. Maybe I should call my shrink about this :) For years, social relations have been rationalised and quantified, human beings are now interacting more with screens than with other living forms, loneliness is almost considerer as a pandemic, with, as usually, an awesome response from the market. Gaslighting, astroturfing and other neurohacking techniques are now industrialised by autoritarian regimes and profit-driven companies. Reactionary propaganda is everywhere and it won’t stop till the trust is back. It requires narratives oriented toward relationships and self-respect, not just on freedom. Easier to say, but many initiatives are active since years, so I hope 2026 could make their fights more visible.