I think your issue should be with Reddit moderators not Ars.

ProPublica happily allows Ars to republish their stories by disclosing the source:

https://www.propublica.org/nerds/happy-birthday-creative-commons

Ars discloses the source compliantly:

https://arstechnica.com/tech-policy/2025/06/doge-used-flawed-ai-tool-to-munch-veterans-affairs-contracts/

https://www.propublica.org/article/trump-doge-veterans-affairs-ai-contracts-health-care

ProPublica literally thanks Ars above for spreading their stories.

The Ars headline (with its subheading) also says a lot more. So it’s just … going to get more upvotes, it’s better bait.

As for the second article getting squashed… Again Reddit moderators.

This looks like a nonsensical conspiracy theory … ProPublica was not suppressed or harmed. Ars literally spread their exact story. Hell, I wouldn’t be surprised if more people read the story because they see “ArsTechnica” and assume good content (because Ars has a super nice website and regularly has really good journalism on it).

Yeah, I think I’m going to stick to my energizers … pretty much everything I use AA batteries for these days is a low draw device (door lock, smoke detector) or something like an emergency flash light, rechargeable batteries don’t make sense for either use case.

I have some old school nice rechargeable batteries as well with an external charger. Those were nice back in 2016 when my bose headphones I was using at the time used AAA batteries, but it’s been a long time since I’ve used those with any frequency.

“Unused road” is ridiculous except in extremes. Unless people merge well over a mile back, 1 lane of traffic will make no difference. The only way “unused road” matters is for the people that haven’t entered the traffic jam yet who are getting off before they reach it.

Very few people (from what I’ve seen) merge more than 30 car lengths out. 30 cars is not going to make a difference.

What does make a difference is the fact that we can’t do a merge at speed because some people want to “zipper late.” It’s the zipper behavior that matters, the “at the very end” part never should’ve been added to that recommendation.

Looking at an actual research paper about this, the zipper merge demonstrated is not at the last possible point. A merge point forms ahead of that point and that’s what should be used. The pictures from their study show the zipper occurring over a wide area with many of the zipped cars driving in the middle.

https://rosap.ntl.bts.gov/view/dot/35694

I don’t know how studies like this have become the recommendations we have. They seem to me to miss critical bits.

Edit: based on my quick read, it’s worth noting the study finds only minimal support for the zipper merge and only in contexts not involving trucks largely based on visual analysis from their video feed as the quantitative data was not statistically significant. We need better transparency on recommendations like this frankly and the research supporting them. We should be able to have an honest debate on the merits of the papers.

The problem with zipper merges as this person describes them is a zipper merge is SUPPOSED TO get traffic back up to speed. However, when your take on the zipper merge is “up there where the wreck is at the last possible spot I can merge” there’s no time for a human to safely merge at speed. So everything has to continue at a crawl.

So the people jumping out of their lane and “zipper merging” at the last second instead of 50 feet out or so end up making things worse for everyone.

The zipper does not and should not be at the point of the physical problem on the road. Just like you should not just drive to the end of the on ramp and at the last possible second merge into the lane on your left without paying attention.

I really can’t more strongly disagree with this take.

Zipper merging is to interleave two lanes of traffic when there’s one lane of traffic available ahead.

It DOES NOT matter if it’s done with 3 feet to merge or 300 feet to merge. There’s no efficiency gain.

What does matter is some assholes trying to merge at speed at the last possible second.

The zipper point should not be the point where there’s NO ROOM to merge SAFELY without EVERYONE going 3 miles per hour.

The handful of times I’ve seen a zipper merge actually start to work, someone rushes down to the end of the line where the problem is, nearly causes a second accident trying to get over, and then everything starts moving at a crawl again.

You don’t need to zipper merge at the “physical barrier” causing the zipper merge to be necessary.

Why would a car alarm be a problem…?

Every place I’ve ever been, they take the keys and drive it into the garage to do any work they’re doing.

Car alarm should only be relevant if the mechanic locks the car, no?

I think it’s fine to have an opinion, just qualify it with “I’ve not been in that situation before, but … I think bla … because bla.”

It’s just about being honest.

I had a friend who’s latest and greatest dating advice was to go back and hangout at the college I graduated from (at the time already) several years ago.

I thought it was an incredibly disingenuous and creepy suggestion.

Him and his partner were like “it’s totally fine…”

Not a single female friend disagreed with me that, that would be very creepy and I absolutely should not do that.

He got mad that I would never listen to his (terrible) dating advice.

You nailed it.

I use containers for work stuff so I don’t have to deal with edge cases of mixing accounts.

I also put Proton in a container to hopefully reduce the odds of any kind of cross site scripting exploits from succeeding… But that base should already be well covered by proton servers and the normal protection mechanisms.

Yeah, that’s not terrible advice.

I think I’m just a bit … demoralized about that … since I do know a lot of older folks (I literally shoveled the driveways of 4 older neighbors this winter and have various 50+ friends) and no help has come from that direction…

I might give it a shot though, thanks

That is how inflation works… when costs go up prices go up.

Yeah, your computer probably should cost a lot more in “today dollars” but because performance of components gets more efficient over time, you can likely get a better computer for less money.

It’s the same reason you have a computer more powerful than multiple thousands of dollar super computers. The technology has improved enough you don’t have to pay as much.

Do you think prices should just be locked in place for eternity at $60?

I haven’t seen any good hiking groups in my area. The ones that exist on meetup.com are mostly “people with grey hairs” … and there’s nothing wrong with that other than I just turned 30 last month.

It’s been that way throughout my twenties.

I’d joined a “young professionals” group and was starting to meet some people that way but … it ended up dying out over the pandemic.

Sure, but that doesn’t mean the game developers don’t need to be paid. It’s still a bargain for the work that’s being done.

If you look at inflation adjusted pricing, it really is a deal. IIRC we should be at like 90 or 100+ dollar games at this point.

Yeah, it looks like basic reasoning but it isn’t. These things are based on pattern recognition. “Assume all x are y, all z are y, are all z x?” is a known formulation … I’ve seen it a fair number of times in my life.

Recent development has added this whole “make it prompt itself about the question” phase to try and make things more accurate … but that also only works sometimes.

AI in LLM form is just a sick joke. It’s like watching a magic trick where half of people expect the magician to ACTUALLY levitate next year because … “they’re almost there!!”

Maybe I’ll be proven wrong, but I don’t see it…

I don’t buy the “it’s a neural network” argument. We don’t really understand consciousness or thinking … and consciousness is possibly a requirement for actual thinking.

Frankly, I don’t think thinking in humans is based anywhere near statical probabilities.

You can of course apply statistics and observe patterns and mimic them, but coorilation is not causation (and generally speaking, society is far too willing to accept coorilation).

Maybe everything reduces to “neural networks” in the same way LLM AI models them … but that seems like an exceptionally bold claim for humanity to make.

I actually did try rock climbing, my ex and I did it.

I proposed going to first time we dated, she got into it while we were split, I would go a few times a month with her, and then basically haven’t been in a few months since we split.

The climbing gym in my area is on the other side of town and just constantly PACKED and I was also never very good at it … plus I’m colorblind and the routes are color coded which makes it even trickier … so it’s not really my cup of tea.

But yeah, I’ve done it probably 15 times or so and have my own shoes.

I’ll probably give it another shot (this time with friends) once it gets cold again (I’m in NE Ohio).

Yeah … I appreciate the effort but I don’t think that’s a real answer.

I don’t want to date half way across the country or world … or slide into other people’s DMs (which is something girls ALSO complain about dudes doing).

At that point if the answer is “someone’s going to be possibly uncomfortable anyways” I’m better off just doing the cold approach where I can at least read body language.

Not to mention there are a ton of elements of attraction that just don’t work on a forum.

We can offload some basic reasoning tasks to an LLM Agent

No, you can’t. It cannot reason. It’s just been fed so much existing text that it appears like it can in some cases. That’s an extremely dangerous foundation on which to build anything.