虎嗅

Things I Wish Someone Had Told Me a Long Time Ago

原文:《那些我希望有人早就告诉我的事》

Summary of Key Points

Recently, the experience with ChatGPT has significantly improved, while the momentum of Gemini has slowed down. Claude and ChatGPT have once again become the two most commonly used AI models by users. Sam Altman, CEO of OpenAI, is reducing his public appearances to focus on internal strategic initiatives and team management. He also shared a list of 17 insights he had previously compiled regarding entrepreneurship, management, and life, covering key areas such as team building, goal setting, resource allocation, execution pace, and talent selection.

Detailed Analysis

1. "Small Changes" in the Big Model Space: ChatGPT Improves, Gemini Lags Temporarily

Many users have noticed that ChatGPT is now more user-friendly—perhaps due to OpenAI's ongoing technical improvements, such as enhancing the accuracy and speed of responses or adding more practical features (like improved multimodal capabilities). Gemini, on the other hand, seems to have lost some momentum, possibly due to initial integration issues with different product lines or because its user experience hasn't been fully refined. Claude, with its strengths in handling long texts and logical reasoning, has made a comeback as one of the preferred models. In short, the competition in the AI industry continues, and those who can continuously improve the user experience will have a competitive advantage.

2. Why Is Sam Altman Less Visible? Probably Working on Something Big

Altman used to be frequently in the news and give speeches, but now he's less visible. This isn't because he's becoming more modest; rather, OpenAI might be in a critical phase of development. They could be working on advancing AGI (Artificial General Intelligence) or implementing important commercial projects, which require his focus on internal matters—coordinating teams, setting strategic directions, and solving core problems. After all, building a solid product and technology foundation is more crucial than external publicity in the long run.

3. Is It Easier to Succeed with Important, Difficult Tasks? Altman's Team Philosophy

Altman believes that working on truly important and difficult tasks is often easier than tackling trivial ones. This might seem counterintuitive, but it makes sense. For example, if a team's goal is to create an AI tool that can transform an industry, everyone will find it meaningful and motivated, willing to work extra hours without complaint. However, if the task is just a minor feature optimization, it may receive less attention. He also emphasizes the importance of a united team with a sense of determination to overcome difficulties (similar to how OpenAI persevered with GPT). Additionally, avoiding unnecessary bureaucracy and meetings that don't contribute to productivity is essential.

4. Ten-Year Planning + Weekly Execution: Balancing Long-Term Goals and Immediate Actions

Altman suggests setting ambitious long-term goals (e.g., achieving AGI within ten years) while taking concrete action on a weekly basis (such as training models or optimizing algorithms). Rapid iteration is crucial; even making mistakes is fine as long as they can be quickly corrected, allowing for continuous improvement. He mentions the "compound interest effect"—the better your product, the more users you attract, and the more data you gather, which in turn makes the product even better, creating a self-reinforcing cycle. This is how many large companies grow significantly.

5. Results Matter Most + Talent Is Key: Altman's Approach to Success

For Altman, results are the ultimate measure of success. No matter how perfect the process, if no results are achieved, it's all for nothing. When it comes to hiring, he advocates investing more time in evaluating candidates based on their ability to complete tasks and solve real problems, not just their intelligence. He also believes that "superstars" (highly skilled individuals) can be more valuable than a group of average performers, but their impact on the team must be considered (e.g., whether they are too independent and hinder collaboration).

Finally, Altman emphasizes that inaction poses a significant risk. If you have an idea, act on it immediately, as both inspiration and time are limited.

These insights are not only relevant to AI companies but also provide valuable advice for professionals and entrepreneurs in various fields. Choosing the right goals, leading a team effectively, balancing long-term and short-term priorities, and focusing on results and talent are all crucial factors for achieving success.