Dr. Dongyang

Personal Predictions of AI, Tech and Investment 2026

Let’s see if they are true in a year by EOY 2026

AI

  1. Coding agents will weigh more on security and develop more guardrails into the process.
  2. Vertical integration of GenAI continues, but it will look like a more typical software building where deep insights about the problems win the customers
  3. LLM foundational capability will still improve but few and at a smaller magnitude.
  4. We will see more adoptions of smaller models for special jobs, still quite good but much cheaper.
  5. Local models or push Compute to the edge.
  6. Fine-tuning could be cheaper and better and allow more companies to run it.
  7. More applications in science, leading to breakthroughs.
  8. The next real AI native app that is NOT code related may show up.
    1. Technology ecosystem mature: Memory will become more important for application stickiness.
    2. Along with agentic systems with reasoning capabilities, it could create new applications that hide all the details and just solve users’ problems.
  9. Robotics GPT moment may happen, high uncertainty.
  10. Reinforcement learning will be the hot kid in the research, but wait and see if anyone has significant breakthrough

Study and Work

  1. Still the employer’s market, but they will start realizing that engineers are more than just coding and sometimes can be the determinant of the success depending on well execution.
  2. Education continues to be more accessible and AI presents the opportunity to provide better UX.
  3. The workforce will undergo more volatility because roles need new definitions and people need more upskills more frequently.
  4. New and more diverse jobs will appear beyond the corporate america (e.g. freelancing, contractors)

People and World

  1. More conflicts between political leanings and more fights internationally.
  2. There is a lot to do to find a new world order where most people reach the consensus that peace is important, that being good is smart. Unlikely this year is the year.
  3. Another tough year for Chinese people, economically and politically. AI boon may spark some hope, but only momentarily.

Investing

  1. VC AI Investment will raise the bar, less money chasing a few really good deals.
  2. Some high burn rate AI startups will die or merge, because they can not raise further at a similar level and they are not profitable yet.
  3. The stock market is unlikely to deliver a super high return, but the risk of the crash is also not high (30% at most). Unemployment is low risk, but watch out for inflation. The tariff is likely to be very loud noise, but TACO.
  4. The AI productivity gain will continue to integrate with every sector and release its power. The current adoption is still small, and the adoption is likely more linear than exponential.

Things that won’t change (or at least not quickly)

  1. People need real-world human connections and seek meanings in their existence. This is more important than AI and technology.
    1. Most people need to work, raise their family, get education and health care.
  2. Evil exists and wins when the good is not careful or not strong.
  3. Largely capitalism will still govern the society.
  4. The US and especially Silicon Valley will still dominate the world in wealth, power and innovation.
  5. Security and risks are always must-dos for sustainable business.