If an AI model is powerful enough to be dangerous, should the company monitor how you use it?
@jason says yes. Watching every search is responsible self-regulation, the kind that keeps the government from stepping in.
What do you think?
Many users dismissed Anthropic's AI monitoring to prevent abuse as invasive scaremongering and ideologically captured overreach, while a few viewed it as justified for managing infrastructure costs.
No Digg Deeper questions have been answered for this story yet.
It’s completely normal for a company like @AnthropicAI to monitor use of a product in limited release to make sure it’s not being abused
This isn’t a privacy issue as some partisans claim 🙄 — it’s standard practice
Here’s the backstory: This administration HATES anthropic winning because a) it’s backed by @reidhoffman (who backed the lawfare against POTUS), and b) because @DarioAmodei didn’t go to all the President’s CEO events (which was a mistake on Dario’s park IMO).
Everyone on the inside understands this. Now you know. 🌈
If an AI model is powerful enough to be dangerous, should the company monitor how you use it?
@jason says yes. Watching every search is responsible self-regulation, the kind that keeps the government from stepping in.
What do you think?
Thank you to our partner @trustvanta! TWiST listeners automate your SOC2 and get $1,000 off at http://www.vanta.com/twist
https://youtu.be/ZoK-prfL0sw?si=8UUW0aFZjs-oQAen
https://open.spotify.com/episode/4VbMriA3jZbrZ2AFx5eBOk?si=ef5640181c714d80
@Jason @AnthropicAI Did you seriously just describe prosecuting Trump for his blatant criminal activity as “lawfare”?
@Jason @AnthropicAI I can write the instructions on how to make a bomb in a word document. Should Microsoft word be monitoring everyone's usage to ensure this doesn't happen?
@Jason @AnthropicAI Makes sense, limited releases always need tight telemetry. We run the same checks on our robot vision stacks to catch edge cases early.
@Jason @AnthropicAI I don’t think it’s as personal as you’re making it seem Jason. If this were personal, the administration would promote (and use) other models any chance they got. They wouldn’t partner with Anthropic for DoD, wouldn’t even be a thought.
@Jason @AnthropicAI No wonder @DavidSacks is always ripping on them. I miss the @DavidSacks that wasn’t compromised.
@Jason @AnthropicAI The administration hates Anthropic winning because they want SpaceX winning as they're all likely heavily invested.
Let's be honest.
Anthropic was on the verge of running away with recursive self-improvement and they can't allow that to happen.
@Jason @AnthropicAI 這種公告乾貨我不做任何存幻想或避談。
@Jason @AnthropicAI Maybe …there is still bromance between Trump and Elon
@Jason @AnthropicAI Context matters more than headlines
@Jason @AnthropicAI Anthropic is the most ideologically captured AI company. That’s the risk
@Jason @AnthropicAI And bc @reidhoffman should be put in jail in is a pedo
@Jason @AnthropicAI Yes it is
@Jason @AnthropicAI You mean it’s a pissing contest ?
No. 4A and 5A rights are being trampled. Also privacy rights, and companies can never accept this. We have trade secrets. patents and research in process. This is an insane step. Verify the users. And that is it. precognition based spying on well maybe someone might use this to do something bad is ridiculous. Just do clear user data collect. Banks do it. frontier companies can do it as well. But like a bank, I don't want them seeing how I use my money.
@Jason @AnthropicAI Anthropic is super woke I am also not a fan
@Jason @AnthropicAI It seems you bounce between defending and attacking both Anthropic and OpenAI depending on the situation. It’s a confusing read. I wish you could be more explicit in your nuanced take.