Palantir is back with another manifesto that clearly demonstrates the value of AI.
On Tuesday, the company posted nine posts on X about the importance of “AI sovereignty.” The post urged companies to keep data in-house rather than outsourcing it to institutions that Palantir says are untrustworthy.
The post also slams “tokenmaxxing” spending as much money as possible on AI. Palantir wrote that the spending has an “addictive sense of false progress.”
Alex Karp has long been a critic of the Frontier AI Institute. In early June, he told CNBC that AI companies “don’t realize how unlikely they are.” Karp said his company’s products “don’t really work as expected” by customers.
Palantir’s 9 posts:
1. AI sovereignty will determine the future of educational institutions. Sovereignty is a prerequisite for choice. Relinquishing sovereignty transfers your institution’s future choices to others, who can exploit it for their own benefit and your detriment.
2. Data retention is your treasure. Please forward at your own risk. Your ability to win depends on your ability to recognize and use your unique edges, and continue to win by compounding the underlying data to generate new insights. Transferring that data inherits access to existing winning plays and gives you the means to produce new plays.
3. Tokenmaxxing hijacks your values and reduces your organizational strength and intelligence. Pursuing the use of large amounts of tokens encourages disposable scripts over robust software, and comes with an addictive sense of false progress. There are reasons why token sellers refuse to charge based on value.
4. Controlling your weight is controlling your destiny. Weight is a distillation of hard-earned and accumulated organizational knowledge. When you let someone else control your weight, you are allowing the alpha part of your business to shift into theirs.
5. There is no contradiction between sovereignty and alpha.. An architecture that maintains maximum sovereignty is one that allows institutions to own tribal knowledge and compound it as alpha.
6. Politicizing technical issues of sovereignty is what our adversaries want. The politicization of technology is the root of false sovereignty. The politicization of technology, especially on Western battlefields, results in decisions that appear to reduce dependence but ultimately limit agency.
7. Real expertise exists. Allowing politics and favoritism to dictate technical decisions rewards those who are best at politics, not who is right. Listen to the person closest to the problem, not the person who speaks most convincingly about it.
8. Learn from institutions that are successful or consistently performing. Institutions facing existential threat cannot afford to make technical decisions based on political preferences.
9. Listen only to institutions, countries, and people that are proven to be correct. A track record of correctness is the best and only signal of future correctness. It’s very wrong to judge something as right or wrong based on who likes it.