AI Predictions that are Falsifiable


I am spending a lot of time reading articles about AI. Many of them I find to be ridiculous. I especially love when there’s actually a falsifiable prediction, and I look forward to seeing how often these are wrong.

I will update this list as I encounter more:


Source: AI-2027.com

Summary: This article has a lot of juicy falsifiable predictions. The source has the appearance of being very serious and legit. According to this, basically by 2030 AI will take over every job, and 2027 is when things REALLY change drastically. I felt very depressed after reading it, and honestly that made me extra bothered.

  • Mid 2025:

    • Research agents spend half an hour scouring the Internet to answer your question
  • Late 2026:

    • The stock market has gone up 30% in 2026, led by OpenBrain, Nvidia, and whichever companies have most successfully integrated AI assistants
    • The job market for junior software engineers is in turmoil: the AIs can do everything taught by a CS degree, but people who know how to manage and quality-control teams of AIs are making a killing
  • July 2027:

    • Gamers get amazing dialogue with lifelike characters in polished video games that took only a month to make
    • OpenBrain announces that they’ve achieved AGI and releases Agent-3-mini to the public
    • Hiring new programmers has nearly stopped
  • August 2027:

    • The government placates the public with job training programs and unemployment insurance, and point to the stock market, which is in a historic boom
  • December 2027:

    • An AI system that is much better than the best human at every cognitive task. The new AI assistant is both extremely useful for any task—the equivalent of the best employee anyone has ever had working at 100x speed—and a much more engaging conversation partner than any human
  • Mid 2028:

    • GDP growth is stratospheric. Government tax revenues are growing equally quickly
    • Agent-5-advised politicians show an uncharacteristic generosity towards the economically dispossessed

Source: The Age of Abundance
  • My bet is that the AI coding agents quickly beat top 0.1% of human performance, at which point it wipes out the need for the vast majority software engineers.
  • But at some point in the next five years, I expect some (obviously not all) coding teams to be entirely self-directing.
  • In the near future, I can imagine a software team that is entirely composed of AI agents. You will have a long-running “product manager” or architectural agent that sets the direction of the product and breaks it down into individual steps. Individual coding agents will take tasks and execute them before passing the code over to a QA agent that tests it against a set of pre-written integration and user-acceptance tests. You may have specific agents that review everything for scalability and security and then pass suggestions to the product manager. Once the software is live with customers, any feedback will be sent into a customer service agent which will distill it and pass it back to the product manager.
  • The cost of accessing world-class medical or legal advice will come down to a $20/month subscription to OpenAI or Anthropic.

Source: The Curve is Bending
  • Without too much justification other than the pace of improvement on the models and the tooling, I expect code-writing agents to essentially match college grads in almost all software tasks by the end of 2026, but wouldn’t be surprised by sooner — as early as the end of 2025 depending on the exact release cycle of the big players.

Source: “I Don’t See Why Not”
  • Nobel-prize winning CEO of DeepMind, Demis Hassabis, was on 60 Minutes and floored me when he predicted: We can cure all diseases with the help of AI. [The end of disease] is within reach, maybe within the next decade or so. I don’t see why not.

Source: Mark Zuckerberg says in 18 months coding will be done by AI
  • Mark Zuckerberg says most Meta code will be written by AI in next 12-18 months

Source: Demis Hassabis - The Man Who ‘A.G.I.-Pilled’ Google
  • AGI in 2 to 10 years from May 23, 2025
  • “I think we are past the middle game, that’s for sure, but I’ve been working every hour there is for the last 20 years because I’ve felt how important and momentous this technology would be and we’ve thought it was possible for 20 years and I think it’s coming into view now. I agree with that. And whether it’s 5 years or 10 years or 2 years, they’re all actually quite short timelines when you’re discussing the enormity of the transformation that this technology is going to bring. None of those timelines are very long.”

Source: Retool - agents automating labor
  • Our next target: to automate 10% of U.S. labor by 2030. That’s approximately 15 million full-time jobs worth of work. Audacious? Yes. Achievable? Based on our growth rate over the last few months, we’re on track to get there by 2030.

Source: Behind the Curtain: A white-collar bloodbath
  • Anthropic CEO Dario Amodei predicted that unemployment could reach 10-20% within the next five years - by 2030
  • AI could wipe out half of all entry-level white-collar jobs
↩ More posts