# tony 🚀

Work. Tony. 💪
Work. 🔥

## Hot takes ⚡

- If you can't solve a problem, there are three possible reasons:
    - you are dumb. 🤦‍♂️
        - solution: get smarter. 🧠
    - you didn't try hard enough.
        - solution: try harder.
    - the problem phrasing is wrong. 🤔
        - solution: solve a different problem. 🎯
- On organisations:
    - A meeting should not have more than three people, because the average contribution time will be less than 1/3. 👥
    - Hackathons are not very efficient and rarely get to ship out in time. Success rate is ~10%. ⏰
- On interviews:
    - If there is no 4 on a interview panel, we should not hire the candidate. 📊
    - If there is a 1 on the interview panel, we should not hire the candidate. ❌
- On ML:
    - Data is the most important thing. 📈
        - Data quality > Data quantity. 🎯
        - Data > Grid > Genius. 🧠
    - 60% of SUNO's issue can be solved by super smart prompt engineering. Skill issue. 💬
        - As of the rest, 20% can be "solved" by Victor. The rest 20% can be solved by data. 🎲
    - 2x SFE makes 2x MLE. ⚡
    - If sth is better, it may not test better. If something is worse, it won't test better.