The split is on who decides when the account should be terminated as criminal for legal reasons, not whether we should support criminals regardless.

It isn't the first time a new technology has been pitched to replace many worker's jobs, both successful and unsuccessful versions of the promise have come to pass several times.

I think what they are saying is "that something can replace a job does not inherently imply the next step is poverty". From that perspective, you can absolutely have it both (and many other combinations of) ways.

The author considered the same w.r.t. remaining bugs:

> More to find

> These were absolutely not the last bugs to find or report. Just while I was writing the drafts for this blog post we have received more reports from security researchers about suspected problems. The AI tools will improve further and the researchers can find new and different ways to prompt the existing AIs to make them find more.

> We have not reached the end of this yet.

> I hope we can keep getting more curl scans done with Mythos and other AIs, over and over until they truly stop finding new problems.

And that makes sense, it'd be quite the argument of coincidence to say there was just 1 proper find remaining & it was only Mythos that managed to find it just at the point in time it released while the other projects have been hoovering up every other find quickly until that point. Possible, but not the safest assumption to start questioning with.

There will usually be more than one issue, Sockbot's solution just happens to deal with more than one at a time.

https://archive.is/D0FOH "Microsoft did admit to darkening the green hill" in the caption for image 1 (Bliss).

Based on the borders of the image shown being extended from the actual wallpaper file (take a close look at the top and left) it was probably cropped as well.

It's entirely possible the color was edited by mistake (i.e. converted poorly) - IIRC the color profile on the tiff was not sRGB.

It is interesting to consider what we talk about when we talk about whether an image was photoshopped because I do actually think that is a fuzzy line and different people may think of different things.

I always assumed this discussion was about exceedingly crass color shifts, the removal or creation of elements not in the original image, not some dodging and cropping.

The Wikipedia article seems to say the same as above:

> O'Rear made it available as a stock photo through Westlight, which was bought by Bill Gates' Corbis in May 1998.[36][43] The photograph was initially titled Bucolic Green Hills.[42][44] By the time of its acquisition, Westlight was estimated to have been one of the largest stock photo agencies in the United States. Corbis had previously hired O'Rear to photograph wine auctions in Burgundy in 1995,[45] and after the acquisition, they digitized Westlight's images.[36] Microsoft contacted O'Rear through Corbis in 2000, wanting to buy full rights to the photograph.[40]: 3:37, 3:50 [6] O'Rear had to personally deliver the film to Microsoft in Seattle due to delivery services declining because of its high value. The Napa Valley Register reported that O'Rear was paid "in the low six figures".[6][40]: 3:57 He had signed a confidentiality agreement and cannot disclose the exact amount.[2][46] Microsoft renamed the photograph to Bliss and chose it as the default wallpaper of Windows XP.[6][37]

> You would have needed the round-trip network request anyway to get the new images, no?

It'd be a shared round trip request for all images (so long as you aren't still using HTTP/1.1) in the 1st example vs a request for the immediate images and then a separate page load.

Both have their upsides/downsides depending on the rest of the page and how users usually use it.

I feel like the option for simplicity lies between "web component" and "make 4 pages". Something near "the button changes the CSS variable controlling the size".

You lose out on pre-downscaled images but gain that the images look sharper for high DPI users and don't have to maintain the image sets or deliver multiple copies when the size changes.

While I hold the same conclusion as you, individuals chiming in to concur based on their own experience is nothing more than a way to validate what certain people of the time & place commonly believe to be true.

E.g. if people were apt to believe girls preferred green peppers more often than boys, there will always be plenty who say "Well, having both girls and boys, I can concur". It could be true, it could be false, or the cause could be something else. E.g., because people think there are certain differences it shapes differences in development which lead to some of them actually being more common for nothing more than the sum of environmental factors - even if those biases only started as misconceptions.

Whichever it actually is, there will usually be large segments of the populations who would observe it to be conflicting things from an individual at-home view and it takes a lot of work & really good data to be able to make a meaningful claim about what and why differences exist.

There a line after which too much science fanboyism can get a person to tie themself in knots.

Science fanboyism is just in understanding what you do and don't know, not eschewing the possession of beliefs or actions made from them.

Theory vs practice.

One can practice the theory of practicality without needing to be ignorant of how uncertain they are as they do so.

The "as many as 1,111" number was:

> The number of our intern goal, a nod to our 1.1.1.1 public DNS resolver, is intentional.

But like the sibling comment says, "over 1,100" does not reference any of their resolver IPs anyways. In all likelihood, they hired fewer than the maximum of 1,111 interns and they are probably chopping slightly more than that here (max vs min).

Also it's funnier to make jokes about hiring people, than it is about firing them.

Wanna bet they are firing in groups sizes, that match exactly CIDR subnets?

That doesn't make much sense as a distinction, though it's a fun brain teaser to see why.

CIDR subnets can be as small as /32 (an individual IP). Even in the case we take the strictest IPv4 formatting requirements (no assumed 0s and no extra leading 0s), that'd be any layoff of 1000-259999 employees. Exceptions start to appear after that, e.g. 260.0.0.0 would be invalid and there is no other valid way to group the last three 0s per the strict rules.

Say we modify the question to just /24 subnets (i.e. similar the classic class C sized subnet) while keeping the rest of the question and rules the same. Through similar logic, any press releases which round to the nearest 10 give essentially the same range, now 995-259994. Since press releases like this tend to use rounded numbers ("over 1,100" seen here), essentially any large layoff could be read as a /24.

One thing I would bet on is "If you try hard enough, you'll eventually manage to find patterns where there aren't any".