Thanks, I read your article and my main reaction is that I'm saddened by the loss of data on those few unreadable discs. I hope it wasn't something you'll need to dig up in a few years.

I love this and I love seeing that it's from 2026 and someone still took the time to do all this testing- it must have been seriously involved because even at 6x it takes a while to fill up a DVD, and then to repeat that hundreds of times on several discs would be an eternity.

I haven't used a DVD+-RW in several years, as wireless file transfer over networks and flash drives handle pretty much all of my needs now, but I sure used the heck out of my DVD writer when I had it. I had no idea these discs could go hundreds of writes before failure, I always got paranoid about reliability and probably never went above 20 writes on a disc.

Edit: at the end of the post the author says, "that’s about 4020 hours across two drives, 5248 burns and both drives are still seemingly operating just fine." What a colossal amount of time.

This article is a total overstatement designed to boost stock prices and none of the actual users can counter the claim because it would require revealing classified information.

This is the same kind of claim you’ve all seen before about AI systems doing something amazing and it’s really just a bunch of people sitting in a call center in a third world country controlling the system remotely.

Only in this case it’s a bunch of senior airmen and staff sergeants sitting in an intel shop doing all the work. Sure, Palantir made a UI but it just plain sucks. And Claude probably fixed some typos in the targeting packages. But let’s not believe that either system was influential to target selection. CENTCOM created a similar number of targets at the beginning of the Syrian civil war before any of these LLMs existed and it took a similar amount of time. We ended up not striking them, but the plans were made after Assad used chemical weapons. All the fixed locations in Iran had packages written and sitting on the shelf before Trump was even elected the first time. The AI in this war added basically no value.

Any claim that Palantir did something useful for the government should immediately be viewed as suspect. I’ve used their software, and it sucks. I cannot understand how they got such big contracts to make a shitty UI that poorly integrates other systems’ data.

North Dakota voted 67% overall for Trump, this is not too far from being representative of the general population. Considering that anyone who is openly hostile against energy companies is going to be removed during selection I don’t see the jury as the issue.

Edit: and considering this was the Southwest district, looking at results by county, 75% seems about right. This isn’t necessarily a biased jury in the sense that selection was unfair, this is probably the makeup you’d get with a fair selection. https://apnews.com/projects/election-results-2024/north-dako...

Both of the examples you gave both fall under "a special keystroke combination," which I did list. Typing "--" is two keystrokes compared to one for an en-dash.

The iOS example isn't just "long press the hyphen" it's "press the [123] button, long press the hyphen, and slide your finger over the em-dash" compared to "press the [123] button, long press the hyphen" for the en-dash.

If you're going to argue at least be genuine. I didn't say it was hard to type an em-dash, I showed that every way to get an em-dash into your writing takes an extra step. Taking an extra step compared to other characters means it isn't trivial.

For someone writing publication quality work, em-dashes appear and if I see an em-dash in a book I don't assume AI writing. But for comments on the internet or a blog posts that aren't meticulous everywhere else, an en-dash is a pretty good signal that the work is AI generated. When people are writing, needing an extra step to insert an em-dash is disruptive to most people's train of thought.