Best of blog.samaltman.com on Kiwi News
Most upvoted by Kiwi News readers
6 stories from blog.samaltman.com have been shared on Kiwi News since 7 June 2023, with 20 upvotes and 4 comments in total. The latest was shared on 11 June 2025. They are listed below, most upvoted first.
Text under a title is an AI summary of the linked article.
- What I wish someone had told me
- Reflections
Sam Altman reflects on OpenAI's first nine years and the year since his brief firing, arguing that the company's work is still early and that its vision will stay fixed while its tactics keep changing. He says ChatGPT, launched on November 30, 2022, kicked off a growth curve the company had never seen, with weekly active users rising from about 100 million to more than 300 million. He calls his firing a big failure of governance by well-meaning people, including himself, and credits Ron Conway and Brian Chesky with help that he says kept OpenAI from falling apart. He also says the company expects AI agents may join the workforce in 2025 and is beginning to aim at superintelligence.
- The days are long but the decades are short
The author, who turned 30, presents a numbered list of life advice gathered over the past decade, with the closing line "The days are long but the decades are short." The list covers priorities such as keeping close relationships, doing work you care about, and making time for family, and it stresses avoiding regret, keeping a low burn rate, and not wasting time. Advice on managing existential angst and on being kind to others also appears, and the post says that most of it comes from personal reflection rather than formal study.
- The Gentle Singularity
Sam Altman argues that humanity is past the event horizon and the takeoff toward digital superintelligence has started, though the changes so far feel less strange than expected. He writes that AI driving faster scientific progress and productivity could produce large gains in quality of life, and that in the 2030s intelligence and energy could become wildly abundant. He says scientists report being two or three times more productive with AI, and that the cost of intelligence should eventually converge toward the cost of electricity. He calls for solving the alignment problem and then making superintelligence cheap, widely available, and not too concentrated in any one person, company, or country.
- How to be successful
Outlier success, meaning either a very large amount of money or creating something important, is presented as a set of 13 practices that build on an existing baseline of success gained through privilege or effort. The practices include compounding, cultivating self-belief balanced by self-awareness, thinking independently, and taking frequent risks. Working hard is described as necessary to reach the 99th percentile, since working smart or hard alone gets to about the 90th. Elon Musk is cited as an example of conviction about sending rockets to Mars, and Airbnb as a benchmark for persistence.
- Researchers and Founders
The author, who has worked with founders and now works with researchers, argues that the best people in both groups share common traits worth learning from, even though they often differ from average people in ways like very high self-belief. The traits named include spending much time on the Hamming question of which important problems in one's field one should be working on, a laser focus on the next step combined with long-term vision, and extreme persistence. The article also cites a bias toward action, honest assessment of what is and isn't working, a steady supply of ideas, strong valuing of autonomy, and motivation driven often by genuine curiosity.