Is college useless? Altman dropped out of school and became a CEO, but students from elite universities support the entire OpenAI!
Sam Altman
Knowledge is power, and the talent structure determines the fate of an enterprise. From Altman's dropout myth to the ranking of the alma maters of OpenAI employees, the theory that academic qualifications are useless has been shattered by data! Graduates from Stanford, Berkeley, and MIT are gathered at OpenAI, and the aggregation effect of AI talents is emerging.
There has never been a shortage of geniuses in the world.
Altman, the CEO of OpenAI, dropped out of Stanford after two years of college to start his own business without a college diploma.
Alexandr Wang, the chief AI officer of Meta, dropped out of MIT to start his business.
Christopher Olah, a former research scientist at Google Brain, the head of interpretability at OpenAI, and a co-founder of Anthropic, didn't even attend college.
Altman even said that "for most people, college doesn't have much effect."
Sam Altman
He implied that his children "probably won't" go to college and packaged this statement as a rational choice.
However, reality quickly slapped down this "theory that academic qualifications are useless."
When the distribution map of the alma maters of OpenAI employees was made public, everything became extremely clear.
Distribution map of the alma maters of OpenAI employees
The employee data of OpenAI shows that several universities stand out:
🥇 Stanford University: 230 employees
🥈 University of California, Berkeley: 151 employees (As a public university, it performs quite well!)
🥉 Massachusetts Institute of Technology: 100 employees
Some other notable findings:
→ The total number of employees from just the top three schools exceeds 480, accounting for more than 13% of the total number of OpenAI employees in this statistics.
→ Universities with strong computer science/engineering programs (Carnegie Mellon University, Georgia Institute of Technology) are on par with traditional research universities.
→ International institutions perform outstandingly. The University of Waterloo, Tsinghua University, and Peking University are all among the top 20.
Deedy, a technology investor, accurately summarized the characteristics of this list: "In essence, this is the global ranking of universities in computer science."
Social media of technology investor Deedy
Deedy believes that academic qualifications are still important. But he also agrees that this list only shows that the best students from these prestigious universities are highly proactive and may not necessarily reflect the quality of their education.
Social media of technology investor Deedy
Graduates from prestigious universities are gathered at OpenAI
Do you want to work at OpenAI, the hottest AI startup in the world?
Then, having a degree from a few specific universities will definitely add points to your application.
According to Harmonic Data, the top 10 universities that supply talents for the development of large language models such as GPT - 5 include Stanford, the University of California system, MIT, Carnegie Mellon, Harvard, etc. - all are American institutions.
Harmonic Data
It should be noted that the employees from the top 20 schools only account for a part of the total sample, so other institutions form a long "tail of talents".
Even though the data is incomplete this time, for AI researchers, the alma mater is very important.
Social media comments
In addition, this statistics does not consider the influence of the number of students in the school, nor the influence of distance on employment.
Social media comments
Social media comments
Although it is not the key to determining one's fate, academic qualifications are still very important -
It still represents training, resources, vision, connections...
But undoubtedly, the projects you have done, the papers you have published, and the skills you have mastered - these tangible achievements far outweigh the halo of a prestigious university.
Social media comments
Social media comments
The network and community formed among students studying together in technical fields such as AI help new ideas move from research to real - world applications more quickly.
This is not a coincidence. Some founding teams and early members of companies such as OpenAI or Anthropic are all associated with these institutions;
In Europe, we can also see the origin of Hugging Face and Mistral with the École Polytechnique, as well as the research cooperation between OpenAI and the University of Oxford.
All in all, the institutional sources of talents are crucial: it affects research, the application of research, and technical standards, and thus shapes the development trajectory of technology.
Talent is the real moat in the AI industry
AI giants know that talent is their biggest moat.
In July last year, consultant Ram Srinivasan judged that the battle for AI talent had entered the second stage.
Battle for AI talent
Let's start with the facts from six months ago:
The top AI researchers at OpenAI earn an annual income of more than $10 million.
Google DeepMind is reported to have offered a compensation package of $20 million.
A retention bonus of $2 million (only required to be employed for 12 months) is now the norm.
Now, the internships at AI giants have gone beyond the traditional concept of "entry - level work experience" and evolved into a high - end talent war.
The latest report shows that the salaries of short - term entry - level AI positions have increased explosively, and the salary levels are enough to shock many senior full - time employees in other industries.
Now, to lock in the next generation of top research talents, major technology companies are constantly breaking through the salary ceiling:
OpenAI's six - month resident researcher program offers a monthly salary of up to $18,300 (about 129,000 yuan) and the opportunity for a formal position;
Anthropic's four - month research project provides researchers with a weekly allowance of $3,850, a monthly computing resource budget of $15,000, and the opportunity to conduct cutting - edge research under the guidance of a tutor;
Google offers a maximum annual salary of $150,000 (about 1.054 million yuan) for doctoral researchers, and Meta's research interns' monthly salary has risen to $12,000.
Behind this salary competition is the urgent need for top - notch algorithm talents in the era of large models. Interns have become an indispensable part of the AI strategies of large companies.
In 2024, when a Meta researcher rejected an offer from Perplexity, he only said: "Come to me when you have 10,000 H100s."
Battle for AI talent
Even with a compensation package worth $1.5 billion, Meta still failed to persuade some top AI engineers to join - they chose the more autonomous path of entrepreneurship.
Battle for AI talent
This points out the essence: This AI talent competition has already gone beyond the scope of salary, and the core lies in infrastructure, resource channels, and grand visions.
We are witnessing the rise of a new economic order - a new world built around intelligence, leverage, and amplification effects.
This ranking list of the alma maters of OpenAI employees once again proves that OpenAI's real moat lies in its attractiveness to AI talents, not its AI models.
Social media comments
Compared with the "self - recursive improvement" of AI, the key to the exponential advantage that OpenAI currently has:
It's not the prompts or GPUs, but the aggregation of top - notch talents, pulling top talents from all over the world into the same feedback loop, and finally forming a compound interest effect.
References:
https://www.ram - srinivasan.com/post/ai - talent - war - 2 - 0 - has - begun - here - s - what - most - are - missing
https://www.linkedin.com/posts/josh - angle - 816b3436_want - to - work - at - the - arguably - the - hottest - activity - 7389688792737210368 - 6DOS/