PokerFacowaty

> Just use the LLM to make a thing that's actually good at what you need it to do.

Thinking about this every time someone shares their setup and there's LLMs doing simple periodical things directly instead of having the LLM write a deterministic Python script that will never try to hack your sandbox lol

You're right that it's not totally true, because it's not a universal Markdown flavor, but at the same time their additions are well-documented in their docs (they have to be for people to use them), so migration tools can just keep up with that.

They are not the same "cheating kids" though. How easy it is to get it to spit out your homework and how widespread it is encourages more people to give up, think "everyone is doing is, so why shouldn't I, I'm a loser if I work honestly", and do the same.

I'd say oh-but-what-about-calculators-i-am-very-smart comments call for a drinking game, but no one should be doing this, as they'd die

The idea always has been that it's writing on paper with all the convenience of a tablet like syncing your notes. A notebook will feel like paper, but won't sync, an iPad will not feel even close to paper.

I was considering reMarkable when upgrading my ebook reader with note taking capabilities, but it's mostly a notepad with an option for ebooks (and not the other way around) and the subscription model on top of $500 hardware was ridiculous.

> In these situations a good manager is what makes the difference; some will fold to the pressure in the hopes that being a "team player" will get them brownie points from leadership (it wont), and some will stick up for their team out of decency and hoping it will get them brownie points from their colleagues (it will).

One of the dimensions I judge my managers on is the "I screwed up, what now?" factor. A good manager will be the shield between me and the higher-ups and will meet with me to discuss the issue to understand it better. A bad one will tell me everyone including them is angry at me. So far I've been lucky to mostly have seen the former.

"helps couples rediscover the joy of swiping" never for a second have I felt joy while swiping

Off the top of my head there's two things that rub me the wrong way about Goodreads.

One is there isn't a separate section for professional reviews (Polish movie/TV site Filmweb has that), so that right off the bat the first comment might be that someone doesn't like what the book is even about, it's a 1-star, liked by 15 people.

Two is they closed their API completely, so there's no way you can get any book info from their DB, not with limits and/or authorization, not if you pay, just not at all.

This is a great article and I think I agree with most of it, not being an agent user myself.

One point the author might be missing is:

> If you were trying and failing to use an LLM for code 6 months ago †, you’re not doing what most serious LLM-assisted coders are doing

The tech giants and AI bros alike have been bombarding us with "but look what it can do NOW", "the new shiny model is revolutionary, now buy tokens", "look what's around the corner" and whatever else for over 2 years now. Can you really blame a typical "skeptic" for not being interested in putting their time to veryfing these claims themselves after seeing mediocre results so many times? You start to develop a filter for these things.

Seconded, I've been using plappa for a couple months after Apple messed up PWAs after bringing them back (Audiobookshelf didn't want to change to next chapter when the app was in the background) and I'm generally very happy