> “This was an email address that looked like the real thing,” says Exempt, explaining the mechanics of how he tricked Charter Communications. “The real domain of the Jacksonville Sheriff’s Office in Florida is jaxsheriff.org. We purchased jaxsheriff.us and then spoofed our number as the department’s, so that when we called them to verify receipt of the legal process, when they searched the number, it would come back to the sheriff’s office, giving them no reason to doubt it. We use real badge numbers and officer names as well.”
I'm honestly impressed. It's an interesting situation where the companies can only verify the same information that the hackers have access to
Name Space Within States:
------------------------
"locality" - cities, counties, parishes, and townships. Subdomains
under the "locality" would be like CI.<city>.<state>.US,
CO.<county>.<state>.US, or businesses. For example:
Petville.Marvista.CA.US.
"CI" - This branch is used for city government agencies and is a
subdomain under the "locality" name (like Los Angeles). For example:
Fire-Dept.CI.Los-Angeles.CA.US.
So you'd be counting on the sub-registrar of jacksonville.fl.us not to allow a registration for the fraudulent "business" of Sheriff, Inc. -- multiplied by every municipality across the country.
Many top-level TLDs have requirements you need to fulfill, .edu is a good example. Similarly you need to prove you're a local entity for many country-specific TLDs. At the end of the day though, this attack vector will always be there, no matter how diligent you are about it. Phishing is all about numbers and one in is often all you need.
gpt-5* reasoning models do not have an adjustable temperature parameter. It seems like we may have a different understanding of these models.
And, like the other commenter said, the temperature may change the distribution of the next token, but the reasoning tends to reel those things in, which is why reasoning models are notoriously poor at creative writing.
You are free to run these experiments for yourself. Perhaps, with your deeper understanding, you'll shed new light on this behavior.
It surely is different. If you set the temp to 0 and do the test with slightly different wording, there is no guarantee at all the scores would be consistent.
And if an LLM is consistent, even with a high temp, it could give the same PR the same grade while choosing different words to say.
The tokens are still chosen from the distribution, so a higher probability of the same grade will result in the same grade being chosen regardless of the temp set.
I think you're restating (in a longer and more accurate way) what I understood the original criticism to be, that this grading test isn't testing what's it's supposed to, partly because a grade is too few tokens.
The model could "assess" the code qualitatively the same and still give slightly different letter grades.
It's inaccurate that GrapheneOS fully endorses Signal and Tor. The GrapheneOS founder was blocked by Moxie (when they were still leading the project) for criticising their approach. They have also warned countless times about the limitations and weaknesses of Tor.
> Arch Linux is an independently developed, x86-64 general-purpose GNU/Linux distribution that strives to provide the latest stable versions of most software by following a rolling release model.
> This page complements the Installation guide with instructions specific to Apple Macs. The Arch installation image supports Apple Macs with Intel processors, but neither PowerPC nor Apple Silicon processors.
(FWIW, I understand that there is benefit to good coverage of a narrower scope, but I do wish Arch would fold https://archlinuxarm.org/ into the main project and be officially multi-arch, but that is not the world we live in.)
Arch package manager here, there is ongoing work behind the scenes to support multiple architectures (aarch64, riscv, etc), but as our volunteers (myself included) are doing this in our free time, progress is up in the air.
Linux is fine now, and has been for at least the past 5 years if not more. Even HiDPI works just fine now which has been a pain point for a while (at least, it works great on KDE).
That being said, my daily driver is macOS ever since apple silicon released, purely due to the laptop hardware. I keep a reasonably powerful Beelink mini PC mounted under my desk running ubuntu server and most of my work happens there over SSH with Tailscale. If you're primarily a laptop user, I'd definitely recommend this set up (or something similar), you get the best of both worlds.
I switched a month or two on my desktop. Then when that turned out good, I switched my laptop to Linux, too. No hardware issues on either one, and the WiFi on the laptop works just fine. (My desktop is connected by Ethernet.)
Sleep is really most useful for laptops and I'm not sure fast boot really solves that use case as well as it does on a desktop (where you really never got as much out of sleep anyways since you're always plugged in).
Just like Windows 11 isnt Windows 95, Linux today isnt Linux from 1995. Even if you use Arch, with nothing configured, still in the install CLI its pretty much just:
'station wlan0 connect "SSID"'
'enter Password:*** '
Done
and this is the worst case scenario, with arch, a minimalist distro, where most things arent there or configured by default.
raises hand As of this month, my Windows-only desktop gaming computer is now dual-boot, and I only boot back to Windows for a particular game.
The main pain-point was that the remote backup service had no Linux client. I ended up solving it with restic, but I acknowledge that isn't a turnkey solution for archetypal Aunt Tillie.
I built a new desktop PC last fall and every Linux distro I have tried this year has WiFi working out of the box. Contrast that with Windows where I need to keep the drivers on a USB stick so I can bootstrap myself on a fresh install
The MacBook I use for work sucks and has weird issues when it wakes up from sleep. I've started having to restart my computer to fix them. I can't remember the last time I've had to do that.
I'm honestly impressed. It's an interesting situation where the companies can only verify the same information that the hackers have access to
reply