MatrixMantis

> But the reason it was not crazy for me was that I actually don’t have any system running PHP myself. Our PHP hero himself knows better than to actually use it. What a badass.

@doctorow.pluralistic.net has suggested that the AI Bubble will create good real estate opportunities for laser tag arenas, because what else do you do with that kind of space after the zeitgeist moves on? Dark thought of the day: Detention Facilities.

They say we need guardrails to prevent terrorism, really they want guardrails to prevent this.

The difference between the attestation done by a YubiKey and the attestation done by an OS, is that the OS likely accepts updates so that tomorrow it can attest differently. That's the hidden dial which makes most passkey implementations a trap.

This is the sort of thing that Palantir does.

The difference between the attestation done by a YubiKey and the attestation done by an OS, is that the OS likely accepts updates so that tomorrow it can attest differently. That's the hidden dial which makes most passkey implementations a trap.

But there's plenty of money on the table for those who decide to use it that way anyway. Platforms would be happy to pay a fee to the device vendor to know more about how best to sell things to you.

The difference between the attestation done by a YubiKey and the attestation done by an OS, is that the OS likely accepts updates so that tomorrow it can attest differently. That's the hidden dial which makes most passkey implementations a trap.

If authenticators are smart, then the door is open for the device vendor to use the attestation object field to strike up a conversation about your behavior with whatever party you're using that device to authenticate with.

The difference between the attestation done by a YubiKey and the attestation done by an OS, is that the OS likely accepts updates so that tomorrow it can attest differently. That's the hidden dial which makes most passkey implementations a trap.

Knowing whether a human generated something is not as important as knowing which human did, and in what dimensions they are trustworthy as determined by other humans that the user has designated trustworthy for making such assessments. We don't need a proof of human toil, we need a web of trust.

Signatures and annotations, not watermarks.

This sounds like the sort of target you'd chose if you wanted to expose the FBI for corrupt hiring practices. I suspect the hackers are the good guys in this story.

I doubt we'll see candidate agents harassed singly. Having a drug habit isn't that embarrassing for normal people. Its your body, do to it what you will. We might see the agency looking like a merry band of thugs hiring their buddies who are unfit for the job.

Gödel tells us that mathematics is necessarily inconsistent or incomplete, which means we get to make some choices about which mathematics we build.

Personally I hope this happens. The real numbers are a very spooky place, they're a square peg for almost every application. Nothing real about them. They're the tool of choice for so many because of historical accident.

Gödel tells us that mathematics is necessarily inconsistent or incomplete, which means we get to make some choices about which mathematics we build.

The training data has encapsulated a certain amount of taste from past humans, so LLMs know to recommended polar quardinates here, Cartesian there, sets here, categories there...

Gödel tells us that mathematics is necessarily inconsistent or incomplete, which means we get to make some choices about which mathematics we build.

I'm not super in the know about who's who in the Linux community, but I've heard of DHH. Is there anybody 1Password could've funded that would've caused more of a backlash? If not, then I assume that the backlash is the point, and DHH's particular objectionable view is just fuel.

In the US/Pinochet/Chile case it was about making off with Chile's copper to support, big surprise, the energy initiative of the day: electrification. In the 1Password/DHH/Linux case I expect it has something to do with passkeys and vendor lock-in.

I'm not super in the know about who's who in the Linux community, but I've heard of DHH. Is there anybody 1Password could've funded that would've caused more of a backlash? If not, then I assume that the backlash is the point, and DHH's particular objectionable view is just fuel.

The playbook was perfected via US foreign policy in the 70's: 1. pick a target that you want to delegitimize 2. find somebody who feuds with them 3. empower 2 until 1 is defeated 4. now 2 is in power, and they owe you a favor Example: The US-backed Chilean coup that put Pinochet in power

This is about more than which brand is used. I hope we see Axon and others in the news as well.