I think this is the most suitable investment approach for ordinary people.

If it were me, I would allocate like this:
50% index ETFs
30% SPY + 20% QM, as the core of long-term holdings, sharing the long-term returns brought by America’s economy and technological innovation.
30% AI-related high-quality companies’ common stock
Choose companies you truly understand and are willing to hold long term, and share the benefits of AI’s long-term development (memory/storage, optical interconnects, robotics, advanced packaging, CPUs, energy, etc.).
10% event-driven short-term trading
Only do some trades with relatively high certainty—for example, “buy the rumor, sell the news” opportunities around events like product launches or industry conferences. If you can make money, you make money; if not, it won’t affect overall returns. Manage risk, and it’s best to set take-profit and stop-loss.
10% cash
Keep it for adding to positions when the market adjusts, so you always have ammunition rather than having to just watch during every pullback.
Of course, everyone’s risk tolerance is different, and the proportions of stocks, cash, and short-term positions can be adjusted according to your own risk preferences—there is no standard answer. Personally, I tend to keep a larger share in cash.
My investment thinking has always been relatively simple: most of the funds go to long-term trends, and a small portion goes to short-term trading.
I still believe the AI bull market hasn’t ended yet. At least for now, I haven’t seen evidence sufficient to overturn this long-term logic. Personally, I see the time horizon as ending at the end of 2027, and at the latest, the first half of 2028. By the end of 2026, S&P is expected to reach 8,000–8,200; before the end of 2027, it could reach 9,000.
Many people always say that long-term investing means holding for ten years, but for an industry like AI that’s still growing at high speed, I think it’s enough to focus on the next 1–2 years. Holding patiently for 1–2 years shouldn’t be particularly difficult. As the stock price keeps rising, the market will gradually start to price in the growth expectations for the coming few years in advance. Then we decide whether to keep holding based on fundamentals (whether the rate of price appreciation is far exceeding the pace at which fundamentals deliver), valuation, and market sentiment (whether it’s consistently optimistic—there are doubts now, optimism, or a healthy state).
In many cases, holding medium- to long-term can get you better returns more easily than frequent trading. Of course, real short-term trading experts are the exception—though there are ultimately very few of them.
As mentioned before, take Bitcoin as an example: many people know it has a four-year cycle (which is also an economic and business cycle). If you had chosen to hold long term around the end of 2022, continuously capturing the later part of that cycle, the overall performance would already be quite impressive. But not many truly managed to do it, because along the way there were many 30% and even 40% pullbacks; many people would exit early during the volatility, and in the end miss the subsequent major upswing.
I think AI is similar. A long-term upward industry doesn’t mean the stock price will rise in a straight line. There will definitely be pullbacks, consolidation, and sentiment fluctuations along the way. If every time it adjusts you start to doubt that the bull market has ended, and every time it rises you rush to lock in profits, it’s hard to truly make money across the whole industry cycle.