On point A (OpenAI):
I respect how Anthropic has navigated recent events. When Secretary of War demanded they remove safeguards, Anthropic held its stance despite a $200M contract being on the line and threats of being designated a “supply chain risk”.
And with the Mythos announcement, they started with selective access, letting a small group of trusted partners find and patch vulnerabilities before attackers could exploit them. This is also where I’d push back on points B and C: profit and safety aren’t inherently opposed. A company can make being the trustworthy option part of its commercial strategy.
On points B and C:
Kat Woods and Marius Hobbhahn have far more experience than I do. They’re in the field with real skin in the game, while I’ve only recently started engaging with these questions. That said, I don’t see profit incentives and safety as necessarily opposed. Thoughtful AI policy and governance can create conditions that push orgs to allocate meaningful resources toward safety. But ultimately, how a company balances commercial pressure against its stated mission is on the company itself.
On point A (OpenAI): I respect how Anthropic has navigated recent events. When Secretary of War demanded they remove safeguards, Anthropic held its stance despite a $200M contract being on the line and threats of being designated a “supply chain risk”. And with the Mythos announcement, they started with selective access, letting a small group of trusted partners find and patch vulnerabilities before attackers could exploit them. This is also where I’d push back on points B and C: profit and safety aren’t inherently opposed. A company can make being the trustworthy option part of its commercial strategy. On points B and C: Kat Woods and Marius Hobbhahn have far more experience than I do. They’re in the field with real skin in the game, while I’ve only recently started engaging with these questions. That said, I don’t see profit incentives and safety as necessarily opposed. Thoughtful AI policy and governance can create conditions that push orgs to allocate meaningful resources toward safety. But ultimately, how a company balances commercial pressure against its stated mission is on the company itself.