Running a business on that basis would look like the white(ish)-hat version of ransomware. "Oh, your data is gone? How much is it really worth to you?"

map() and reduce() are equivalent in terms of jargon, IMO. Map also suffers from name collision with dictionaries/objects/whatever your language wants to call a key-value pairing.

CJD develops in about 1/1,000,000 people per year globally, and the Fore people have a population of ~20,000. In theory, that suggests there should be an outbreak within 50 years of the society regularly consuming brain matter. Perhaps that number could be reduced somewhat by the infection rate, but the infection rate seems pretty high, to reach the level it did as fast as it did each person with Kuru needed to infect 1.6 people. Perhaps sporadic CJD isn't as sporadic as we assume, and it's tied to being in environments where others have had CJD? Another possibility is tied to the seeming link between genetics and the infection rate of Mad Cow Disease. Perhaps the population regularly went through boom-bust cycles, where everybody with 2 copies of the recessive gene that makes one more susceptible to transmission gets killed off, then the disease dies off for a long enough time for the carriers to re-introduce their recessive trait back into the gene pool.

That actually supports the point that people aren't acting under purely financial motivations. If the guy was purely following financial motivations, surely he would have chosen to sell the vulnerabilities to the shadier side of things. Instead, he dumped them publicly, burning their value while amplifying the "fuck you" factor to Microsoft.

Ignoring reports, or just fixing the vulnerability without acknowledging the work put in by a researcher, is rude and invites rudeness in return.

There's a pretty heavy survivorship bias on that, so you probably shouldn't extrapolate to all pre-civilization humans. For a group to remain largely uncontacted, it means they have to be remarkably insular, living somewhere other people don't want to be particularly bad (implying resource scarcity), and it definitely helps if they respond to attempts at contact with extreme violence.

Hunter-gatherers who survived to adulthood (and childbirth!) could expect to live to their mid 60's to early 70's. The big wins that modern medicine have gotten us are drastically decreasing infant/child mortality and deaths during childbirth. Compared to our ancestors, we've only really added 5-10 years to the life expectancy of elderly life. Largely unrelated to modern medicine, deaths by misadventure are reduced living in civilization, and then even moreso when you've got a large portion of the population not working in dangerous manual labor roles.

Level 5 is a pretty massive difficulty spike. Not sure if there's tech that I'm missing (it seems like 3 and 4 pretty strongly encourage double hits), or if I'm just bad.

You can do E2E encryption for organizations, you just establish all conversations as at least 3-way conversations: the two members of the org, and the org itself. You let the server refuse to host any conversation it isn't privy to, and then apps that support organization mode automatically add the org listener, and users using third-party clients can add it manually.

E2EE has well known issues for scaling, which is why it's found primarily in small group / personal chat

The org can just host non-E2EE XMPP, and it'd be the same as this wrt trust, easier to implement too.

Definitely my favorite of Shakespeare's tragedies. I've got a special place for the wild tone shift between the murder and the comic relief scene that follows in particular, as the rather temporary release of the tension that had been building for the prior act is incredible. In my AP Literature class, we had to perform a scene or so from a Shakespeare play of our choosing, but thankfully the only injury during our performance was somewhat intentional. After two weeks of trying and failing to get the girl playing Lady Macbeth to pull off a convincing stage slap, I told her to just give me her best shot the day of. Turns out this 4'10", 90 lb soaking wet fury had a heck of an arm, and thus I ended up finishing up my lines scrubbing my hands while seeing stars.

Surely at any point after "Make the "Add to Cart" button blue", it'd be faster to just do it yourself: "grep -r "Add to Cart", insert "addToCart" as a class to the html element, then crack open style.css to stick "button.addToCart {background-color: blue}"

I'm counting maybe 80 keystrokes? That's shorter than your second prompt.

This idea generalizes. Large Language Models are poorly suited for tasks that we have already purpose-built systems to be easy for humans to use. The easiest way to tell your website that you want a button to look a certain way is to update the code. If you know exactly how you want something done, we have developed an incredibly efficient way to tell computers how something should be done: it's called source code.

LLMs work best when they're handed tasks that you don't want to figure out how to do.