It sounds like AI is making doing that a lot easier, and probably hard to contest what it finds without running extra tests.

Like, obviously only a doctor can diagnose. But, they can use technology to help them look at problems, including AI (literally the most versatile pattern matching software we have, just in natural language and with lower success rates). And they can make their own judgement call on if what the tool output is saying is correct.

However, if a Doctor were to see a bogus secondary diagnosis on their screen, and then dismiss it, they're opening themself up to medical malpractice liability, no? Like, it seems like the safest option would be to run (and bill for) the extra tests to confirm if the AI's story is true or untrue.

At what point does it become a 5th amendment issue? Maybe when we get AI in our corneas a la Black Mirror's The Entire History of You?

The problem with the tutor's question is that they're going to AI to answer this question and not asking underlying questions like "what makes a square a square" and "what makes a rectangle a rectangle", then doing their own thinking to determine if a square's qualifications also fit inside the qualifications to be a rectangle.

Heck, even if you just feed the underlying into AI it'll reason the answer in a logical manner:

> In standard geometry, A square is a quadrilateral with 4 equal sides and 4 right angles (90° each). A rectangle is a quadrilateral with 4 right angles. Its opposite sides are equal and parallel, but all four sides do not have to be equal. So the key distinction is: Rectangle: four right angles. Square: four right angles plus four equal sides. > > That means every square is a rectangle, but not every rectangle is a square.

Asking LLMs to just give you answers is the real antipattern. But if you remember that the only thing it's good at is finding patterns given context/constraints, it can be an extremely good tool.

Also, I have a feeling the vast majority of the populus is using free versions of Claude or ChatGPT. Thankfully recently ChatGPT Free is GPT-5.6 Luna, but before the response qualities and capabilities of the free chat output were less than stellar.

IDK, governments tend to align with what results in the most economic activity being created / moved into their country ("growth"). If you were able to convince that overall economic growth would be multiple times higher over a few decades with the system, it might just become appealing.

A problem I see with implementation in the US, though, is that local municipalities are who tax land and property value, so entire rural counties and cities would have their funding stunted. And asking the government to buy into that while subsidizing low-growth areas for a long time (decades / forever, if a rural area never develops) is a really hard sell.

Most of what I've heard is that raw reasoning traces are really good for distillation, although no idea how much the summarization actually hurts distillation.

Eventually this part will end. Most are in a "move fast and raise our stock value as much as possible" mode, so are fine with infinite auto-scaling to handle the surges MCP traffic is causing right now. But eventually, we'll probably see more per-auth rate limiting and/or heavily restricted MCP usage for non-frontier labs agents (especially if they only want e.g. Chat users, not coding harness users).

I do wonder, was Polymarket given a chance to prevent itself from being accessed/used from Spain? They seem to be a-okay with blocking usage from US states that have asked them to cease operation, so I don't see why they would try to actively defy countries that have (legally) asked them to not do business there?

The internet is only more concentrated because most people choose price and convenience over a foolproof level of security. The cost/effort required to surveil traffic in coordination with TLS-terminating CDNs is multiple orders of magnitude higher than the effort needed to throw some high-Tbps middlebox in an IX rack to siphon every packet passing through it.

Also, the _main_ use case of `cloudflared` tunneling is using it as a long-term way to host production websites on your own hostname. the ability to create ad-hoc tunnels is more of a gimmick / advertising opportunity.

0: https://github.com/judge2020/cloudflare-connectivity-test/wi...

This isn't a thing. Discord's QR Code scanning is entirely a feature they made unrelated to passkeys or bluetooth. Passkey auth using a QR code has a step that verifies the proximity of both devices using BLE.

Passkey QR codes are only WebAuthn via CTAP hybrid transport (with BLE verified proximity)

However, passkeys can and are available to be shared via password managers. They’re not locked to the secure chip on the device where they live usually. iOS’ Passwords app has a share button and 1Password lets you share passkey-containing items.

In fact, the QR code login feature makes it even easier to do a one-time sign in to your account for a friend, if you don’t want them to be able to login to your account indefinitely.

0: Netflix doesn’t support passkeys because their main audience is people signing in via smart TVs and whatnot, which largely don’t support CTAP or Webauthn in general)

This is solved by passkey-implementing software and devices (with Bluetooth) allowing you to log in with a QR code (Webauthn via CTAP hybrid transport). iOS and Android support this, and it’s generally not a locked-down thing if other devices wanted to do it too.

The only use case left is in “how do I login if all my devices are stolen/fall into a body of water” in which there really isn’t an answer beyond “get (a|your) device back, sign back into your password manager, use that to get back into critical accounts”.