Ha ha ha. I did settle into using a Mac for development work, but not without a moderate degree of customisation. Reducing keyboard latency was, like you said, one of the changes that made the most difference.
I'm curious what other kinks you have had to iron.
I suppose that, for fans of information management, the need for external utilities such as org-roam-ui will be obviated once native equivalents to Obsidian's Graph View appear. The sooner this supersedes GraphViz-based stopgaps such as the one in grove.el, the better (I'm guessing that graph-fa2 renders to SVG and then rapidly swaps the image in an Emacs buffer).
It seems that, as soon as you transact with a sanctioned entity, you are globally in breach of the agreement and risking the revocation of all your certificates — also the ones for non-sanctioned countries.
Front matter:
- it is called a "Subscriber Agreement" and not anything that suggests that its scope is a single certificate
- it's a "contract [...] regarding Your [...] rights and duties relating to [...] Certificates" - plural
2.1 "Term":
- "[the agreement] will remain in force during the entire period during which *any* of Your Certificates are valid" - plural
3.1 "Warranties":
- "[by] requesting, accepting, or using *a* Let’s Encrypt Certificate" - plural
Weirdly I find myself writing more code lately, not less. A lot of it is AI assisted but a lot of it also isn't. I find that where previously the management chain wanted things meticulously planned out, and we had to have 17 meetings with different stakeholders and principal engineers to approve the system design.
That account is setup pretty well compared to many I see on here, no giant red flags in any individual comment and they’ve prompted/filtered out the repetitive structure decently so they don’t all look exactly the same. But exhibits number of common attributes/failure modes you notice when you spend a while reading through the history of obvious bot comments.
Some of the signs in comment history:
1. Multiple uses of the notorious HN bot phrase “X is real” e.g “The tradeoff is real”
Spam bot replies love that, at this point if I read that phrase on HN I immediately become suspicious. Also, ending a comment with an open ended question or random caveat/concern/qualifier too frequently.
2. Rapid fire commenting. Most new HN accounts run by actual people usually ease into commenting regularly after their first comment, spam bots go from 0-100 suddenly replying with detailed, uncannily on-topic comments 1-3x per day
3. Unnaturally on-topic but uncanny valley comments. Humans typically respond to an idiosyncratic idea of what the topic of a discussion is, but bots typically respond with a perfectly responsive summary and/or personal experience.
As for uncanny valley, this is the big one that caught my eye:
> Love this approach. SQLite's WAL mode + litestream backup handles most durability needs beautifully. The simplicity of keeping state in a single file you can query with SQL is underrated. Going to try this for our next side project.
It’s just … bizarre haha. It’s replying with what is pretty much the most basic, fundamental description of what SQLite is, with zero actual opinion/experience/commentary.
It’s like responding to an article about PepsiCo Q2 Earnings with “Great read - my family of four enjoys drinking Pepsi products, the carbonation and various flavor choices are a big plus compared to other brands”. If it wasn’t a bot it would have to be someone ineptly karma farming, and either is trash deserving of the flags it got.
Plus, the initial comment in this thread while not enough evidence on its own is already suspicious because it’s confidently stating an anecdote about billing that just doesn’t really make sense as if it’s a super common way tokens are billed.
It could be true in their unique case but it’s just weird and presented like it’s common knowledge so comes across either as a human pretending to have experience or a bot slightly off target.
But it’s not enough without the full comment history, and even without a giant smoking gun having seen a lot of bot comment histories it fits the pattern closely enough in most ways to agree with the other commenter (and the flags/HN shadowbanning) that they are a LLM bot.
In case you are interested in hearing a reader's opinion: the AI writing was noticeable and detracting. The most significant cues for me were the sensational tone and the prolific "clever" one-liners. I think that while it is difficult to make the judgment with any degree of accuracy based on a single suspicious sentence, but given the length of the article, all the individual cues eventually add up to a near certainty.
I think that the sorry thing about the article is that, even though I've read through an article of yours, I have learned nothing about what kind of person you are. I think that there's more to blogging than just showing the work. It's also a stage for you. The displays of character in the article ("I had a feeling", "I sat with that for a minute") written in first person are not actually yours, and are instead, in a way, a performance of the LLM that you used. So in my opinion (and you're free to disagree) you've robbed yourself of the attention you deserved.
However, I assume that the contents of the investigation were true, and if so, they are quite damning (in fact, my SO has just surprised me with a cheap Chinese projector. Nice timing!). It was also great that you've shared the prompts and results at each stage.
serious_angel is not contending with you that the design is bad, or that it is bad because it is unoriginal. In fact, they are not even specifically calling out the design.
They have noticed the design, recognized it as the output of an LLM, then proceeded to discover that an LLM was involved in much of the creation of the project. This is an academic project. Whatever the pedigree of the researcher is, this implies to the grandparent that the final result of the work may be amateurish or worse, to an extent generated. Therefore, he's concerned that it puts the legitimacy of the research outcomes (e.g. completeness, contents of letters, classification, maybe even hallucinations in the thesis proper).
Preemptive arguments:
1. "The author's a researcher, not a programmer; therefore it's fine to use an LLM. It is preposterous to ask each researcher to learn web development to publish their research." You are right, but given the amount of vibe-coded websites we see, and them all having the default (Astro?) style, the grandparent all the same has the right to associate that style with untrustworthy crap. I'm not saying that this academic website is necessarily crap. However, I think it's useful for the grandparent to share their sentiment, because the researcher might not know.
2. "A lot of pages have links to sources; you could verify the legitimacy yourself". perhaps, but doubting the veracity of research is a bad first impression, isn't it?
It's a bit sad, because the website is non-trivial, and would have taken quite a bit of effort without an LLM. But it is difficult to separate webdev enablement with the rest of the LLM baggage.
Thanks for the feedback on this. I'll note that I am not an academic nor a researcher. I'm a hobbyist who wanted to bring together data from different letters and read them for myself. I don't aspire to be a researcher nor publish academic work.
The primary goal is to provide access to these letters to non-academics who would like to read what Romans were writing in language they can understand.
I take all the points about research outcomes and the quality of the data itself. That is going to be an ongoing process to continue to improve it alongside LLMs. I have a day job, and this is just a side project, but where it can provide value, I want to lean into that side of things.
I was checking the submission on the phone and only peeked at the comments section. While it's not always easy to judge if something is AI-generated or edited, here it was obvious at first glance from the quotes. Assuming that all of the comments were done in good faith, I think that the low AI literacy even here is really concerning.
Flagged for AI content: I hope this submission dies and the user is penalized (look at their submission and comment history!), because IMO the article does not belong on the front page. Quick polemic:
>The US is winning the AI race where it matters most: commercialization
If you ask me, one could name different criteria for winning, and commercialization would not be the first thing to come to my mind:
> It also owns platforms that generate and organize the data of the AI age. YouTube is a video corpus. Google Drive and Microsoft 365 sit inside daily office work. GitHub sits inside software development.
Yeah, okay. China does not have any platforms nor data.
reply