Jered Heeschen
banner
jeredh.bsky.social
Jered Heeschen
@jeredh.bsky.social
Gamer, tech writer, curmudgeon, information-obsessed.
Pretentious indeed. He starts by listing sides of the national debate that no one outside of his group of friends would recognize, then pretends not to favor one or the other, then goes on to double down on one side's assertion that true US citizens celebrate Christian holidays. Some law professor.
December 26, 2025 at 10:43 PM
My favorite (oversimplified) definition of knowledge is "contextualized information". LLMs do okay with common contexts that would be in training materials (like comparing measurements), but they fall flat when it comes to the unfamiliar and esoteric contexts historians can end up working with.
December 21, 2025 at 11:01 PM
Don't we already have Wikipedia as a starting point for new topics? It's not perfect, but it's reviewed by people with sources. And if it's something obscure enough that Wikipedia doesn't cover it, AI would be a bad choice because a lack of coverage in training materials would cause hallucinations.
December 21, 2025 at 9:03 PM
Hey now, this is a great approach if your goal is to confirm your biases. If you only "research" using familiar sources, you'll only get back what you already know.

It's the ideal method for the "do your own research" crowd.
December 21, 2025 at 8:41 PM
It really does feel like a missed opportunity if the "after mass deportations" picture doesn't show them comfortable in an inn room.

Bonus if they'd drawn the Magi being dragged away in cuffs in the background.
December 21, 2025 at 2:40 AM
Setting the AI thing aside, having the browser bundle a bunch of other functionality was what Mozilla _used to be_. We got Firefox from Mozilla because they split functionality into separate programs. This CEO is talking about rolling back the clock to what already hasn't worked for them.
December 17, 2025 at 6:47 AM
Nah, that's just Neil Diamond.
December 13, 2025 at 9:34 PM
Maybe it could use some comical question marks over Bruno's head.
December 6, 2025 at 8:28 PM
A major difficulty with the AI tools is that they're more user-friendly than a web search (just ask a question), but using the output reliably requires more information literacy than regular web search results do. The output's resemblance to human writing makes us less likely to scrutinize it.
October 30, 2025 at 7:48 PM
Fair. I saw "lazy apathy" as meant for people who think that not voting for candidates will totally show them, and no further action is needed to inspire change. But then, the number of people that would apply to seems like it would be pretty small. So I could just be steelmanning a terrible take.
October 29, 2025 at 10:08 PM
I mean, I could be wrong, but it doesn't read like a demand that someone vote for an ass. It seems more like pointing out that not voting, by itself, does nothing to change a politician's mind. Which is fine, as far as analysis goes, it's just easy to misinterpret it as a call to action.
October 29, 2025 at 9:53 PM
A literal reading of what he posted suggests that he's evaluating the influence of not voting. And I think he's right, in that strict sense, that a politician will not be influenced by someone not voting at all. There's no way for them to see that the inaction was because of a specific policy.
October 29, 2025 at 9:50 PM
It says something about me that at first glance I thought it was Casanova Frankenstein from Mystery Men.
October 29, 2025 at 9:27 PM
To steal a line from Mary Poppins: Kindly do not cloud the issue with facts.
October 24, 2025 at 5:02 AM
Now I'm wondering who would come out on top of a Bugs Bunny vs. Squirrel Girl matchup.

...Thining about it, they'd wind up teaming up against Elmer Fudd and/or Thanos.
October 1, 2025 at 2:54 AM
I have to admit, "wokey" is growing on me.
September 5, 2025 at 8:54 PM
The reason there were more VAERS reports back then was that we were prioritizing the elderly and people with pre-existing conditions, groups with higher rates of health problems with or without vaccines. Of course there were more incidents. Chart the ages from reports over time and it's obvious.
September 4, 2025 at 7:34 PM
Uh, hey, Best of Bluesky candidates in the replies, I think Kat is joking. Just throwing that out here.
August 18, 2025 at 6:54 PM
I mean, Pusch certainly pushed his putsch.
August 6, 2025 at 4:19 PM
Trump never sent federal agents to the basement of Comet Ping Pong.
July 13, 2025 at 12:57 PM
And of course, another part of the problem would be the training material they use for the LLM. They've already announced that they'll take some of the racism out of the training material, leaving one to wonder why they had it in to begin with.
July 9, 2025 at 7:53 PM
Part of the problem was an instruction saying that it shouldn't shy away from unpopular opinions if they are "substantiated". The trouble is that you can't determine substantiation without either understanding of meaning or coded conditions. LLMs boil down to word association, not meaning.
July 9, 2025 at 7:52 PM
Can be found here, and makes the context even worse, really: www.cbsnews.com/news/grok-el...
Grok, Elon Musk's AI chatbot on X, posts antisemitic comments, later deleted
The antisemitic posts from Grok came days after Elon Musk wrote on X that the chatbot had been improved "significantly."
www.cbsnews.com
July 9, 2025 at 10:25 AM
With DOGE's track record, it might be less of a police state than a Keystone Kops state.
June 29, 2025 at 11:55 PM
Poking around, I found that in Spanish, "tia" and "tio" are aunt and uncle, and as a result the word "titi" became a gender neutral reference to a parental sibling. That is a fun word.
June 28, 2025 at 8:46 PM