When sitting down and actually having to put all the puzzle pieces together, I get a bit hesitant to mark down exactly what I think about how so many different things will change in the next five years. So, I think it will be better to sort out what kind of puzzle pieces we have and then later try to piece them together — all assuming that there is no World War III.
Now let’s start with the most obvious one that we have in mind: AI. More specifically, a transformer-like architecture that works as a context model that could, right now, earn a gold medal in a math olympiad. To me, I like to call this sort of model a “context model,” where it takes all the digitized information of humanity and creates this black-box model that could perform inference from that context. Note that this is a multimodal system, and such inferences can be made locally. In five years, we will likely have phones or handheld devices that can run models smarter than today’s GPT-5.2 Pro or Gemini 3.0 Pro locally. And the best models will be served from space. So that means all the smart devices are actually smart, and they will understand us better than the previous Netflix recommendation algorithm. These context models will soon reach their intellectual limits, as I don’t believe they will be able to solve complex problems like reconciling general relativity with quantum physics nor Millennium Prize problems. However, they will surpass the intelligence of any individual human in almost all domains we currently know about, even for experts. Despite this, they will still lack emotions, instincts, agency, and consciousness.
Even without some of our human abilities, that still means all of our hardware will be personalized. A lot of software will disappear (i.e., become irrelevant; of course you can still download them, but no one will seriously use them), such as workout planners. They will just exist in a form that is more personalized. However, we will still have Messenger, Instagram, and apps like that because we will still need to connect with other humans. Yet since these models still do not have agency and full emotional and directional alignment with human goals, we will still hire mostly humans, but the form or way that we work will change as people will more likely change their jobs and move between different domains.
On a similar note with AI, self-driving will be fully accomplished — by that, I mean self-driving cars will be better than human drivers in any situation. Robots will appear more in factory settings, though I don’t see a specific breakthrough that will lead to humanoid robots becoming the most prevalent thing in our lives.
With all the progress I am envisioning, there will be a lot of hype that I believe exists today which will not fulfill its promise in 2031. First and foremost, I will start with something that I have the most experience with: brain-computer interface, or you can call it brain implants. It will still only be functioning at a small scale, helping some people potentially control their limbs, but with no memory uploads and no consciousness in the cloud. We will not have nuclear fusion; there will still be limited total energy, but our ability to harvest clean energy will dramatically increase.
In terms of our day to day, I do not see dramatic change in our daily lives. We will spend about the same time using computers or phones, just interacting with them with a different spin.