In 2020 I wrote about my Post Covid-19 predictions, they were more right than wrong (accurate on technological shifts, and inaccurate in human psychology and geopolitics). Now I want to do the same exercise with what’s happening with AI to read it again in 5 or 6 years.
Before doing the exercise for the AI boom, here’s a quick analysis of how the Post-Covid forecasts turned out:
- Remote work adoption will contract post-vaccine and then resume growth: Highly accurate. The permanent widespread adoption of hybrid work models, white-collar jobs coming to LatAm (particularly software engineering jobs), and the multi-billion dollar explosion of workplace collaboration and cross-border hiring software played out exactly as I anticipated.
- Telemedicine will become standard and medical logistics will be disrupted by drones: Mixed accuracy. While telemedicine successfully cemented itself as a standard global healthcare practice, drone delivery and the elimination of complex medical bureaucracy remain largely unrealized for everyday consumers.
- The value of expensive universities will be heavily questioned while elite institutions survive on prestige: Highly accurate. Public skepticism toward non-elite four-year degrees has surged in favor of high-ROI alternatives, while brand-name institutions remain fully insulated by their credentialing power.
- Online dating will cross the threshold into full social acceptance: Inaccurate regarding the timeline. Online dating had already largely crossed into mainstream acceptance prior to 2020.
- Society will develop a permanent obsession with daily hygiene and antibacterial gel: Highly inaccurate. Though obsessive sanitization dominated the immediate pandemic years, human behavioral baselines eventually snapped back, and daily reliance on antibacterial gel reverted to pre-2020 norms.
- Vulnerable businesses will collapse and a new generation of category-defining startups will emerge: Highly accurate. The post-pandemic era saw the high-profile collapse of institutions like Silicon Valley Bank alongside the birth of the Generative AI boom led by generation-defining companies like OpenAI and Anthropic.
- The nation that invents the vaccine will achieve an ideological victory comparable to the Space Race: Inaccurate. Rapid vaccine development across multiple countries fractured into “vaccine diplomacy,” highlighting global supply chain inequities and fueling intense nationalism rather than an international ideological consensus.
I’m going to leave out geopolitics and human psychology. I’m gonna focus on technological and business shifts. So here are my 6 predictions for the post AI-boom world and how I think they’ll be in a 5 or 6 year timeline:
- There will be more white collar jobs than before, including software engineering jobs. Respect to 2026, not respect to 2021. This is being driven by Jevon’s paradox: We will be building so much software that we’re going to need more people to maintain it and stitch it together reliably. AI is just another high-level abstraction layer (as high-level programming languages and frameworks were back then). We will have less coders and more architects.
- The math of building custom software also completely changed with AI, therefore we will enter the era of fast-fashion for software. Purchase custom software, change it 2 years later. You get something more custom and cheaper than mass industrial scale software. Certain categories of SaaS will disappear. The SaaS that stays is the one where high availability, compliance, and security are critical.
- There will be a huge migration of software developers across companies, but this will be more due to interest rates than AI. The world that existed last decade fueled by low interest rates doesn’t exist anymore mathematically. We will have flatter and leaner companies. Higher interests rates are symbiotic with AI. Higher interest rates gave the mandate for leaner companies, AI gave the mechanism.
- AI will drive severe role compression. Non-technical middle management that only manages processes and people is a dying profession. Frontend/Backend development are also dying professions, it’s easier than ever to own the entire stack, where domain expertise and distributed systems knowledge becomes the differential. Roles will be compressed: Product Managers, Designers, and Software Engineers will be compressed into a single role. Engineering Manager and Staff Engineer (aka Tech Lead but with managerial responsibilities as well) is another role compression happening right now. Some larger companies might have the older separation of reponsabilities playbook in place, but there will be a clear before and after in terms of mindset.
- Certain types of software engineers will end up doing sales engineering. There will be a lot of software with a real market ready to be purchased, but a high-amount transaction requires the empathy and accountability that a human being brings to the table.
- We will see more physical innovation, that happens to have software as something that is table-stakes. Think about building hardware and programming GPUs, EVs, rocketships, and also think about a boom of IRL communities. That said AI can’t just magically compile a factory (yet?). The iteration cycles of the physical innovations of the early 2030s will be much slower than the software innovation cycles of the 2010s and 2020s.
See you in half a decade to see if my crystal ball still works