amoe_ · HackerTrans

2 か月前議論

The difference is that these rhetorical techniques need to be used with taste. LLMs just sprinkle them everywhere to try to make their copy sound good, even when it's completely inappropriate tone-wise. They don't make higher level judgements about when to employ specific features.

amoe_

3 か月前議論

> - Philosophy, really good and useful

I have had the opposite experience, philosophy questions tending to show the various chatbots at their most irrational and confabulatory.

amoe_

4 か月前議論

Completely agree. It's absolutely awful having software projects squatting on the names of great philosophers and artists. I appreciate that perhaps the author wanted to show their appreciation, but there are plenty of other equally communicative options.

amoe_

5 か月前議論

The problem for me is not services where the content is online, you can just avoid those, but cases where access to scarce real resources is controlled through online verification. E.g. renting recording studios, background checks for job applications, things like this. Often there is no route that does not go through a third-party verification service.

amoe_

7 か月前議論

Which particular LLM was that?

amoe_

9 か月前議論

If I'm already using Poetry and not facing any issues with it, is there any advantage to using uv?

amoe_

10 か月前議論

I actually like this and don't want them to be joined up.

amoe_

11 か月前議論

> Wall: Players group 16 elements into four categories (similar to the NYT Connections game)

I have to be the designated pedant here and point out that Only Connect was first.

amoe_

12 か月前議論

The knife crime analogy is a bit off, as we already have age restrictions for buying knives in the UK.

amoe_

昨年議論

I was pro-systemd at the time of the controversial discussions. I still think it's a net positive relative to what was there before. But personally speaking, it's only the core of the software (service management) that improved things for me. The other things (timers, journald) I either ignore or don't like, but perhaps they're useful for distributions.

amoe_

2 年前議論

This legislation was created by the Conservative party, but wasn't opposed by any major political party in the UK.

amoe_

2 年前議論

> You're going to pay for that one later if you do the simple thing and add an "email" column

You'll only pay if the project survives long enough for that new requirement to actually surface, which often it won't.

amoe_

3 年前議論

I'm going to echo what everyone else is saying in this thread -- the inverse problem (YX) is by far the bigger issue with tech questions on the internet in general (mainly SO and IRC). It's common enough that it puts me off asking anything that's not entirely trivial. And yes, ChatGPT mitigates this to a large degree.

However there is definitely a place for pointing out XY problems in a work context, especially if you have to directly deal with the resulting code. You need a lot of shared context to have confidence that you've identified an actual XY problem.

amoe_

4 年前議論

Yep. Pre-generics, Java was considered a good teaching language because of its simplicity and the ability to hold the entire language in your head. I think it's much more difficult to make that argument now. The same for Python. Moreover, while Python 3's unicode/byte string separation was a crucial step, IMV other features from Python 3 haven't made a big impact on how useful the language is.

amoe_

4 年前議論

You have low expectations; I have Racket programs that have worked for over 10 years essentially untouched. More or less the same with Perl. We should be aspiring to Plato's cave (while accepting that we won't reach that).

amoe_

4 年前議論

The OSB has become a big bucket which UK politicians can point to and say "see, we're doing something". When faced with a question about the internet, the stock response of Tory MPs is to say "wait for the Online Safety Bill". In that way, it's quite similar to the situation before Brexit, where representatives would evade responsibility for policy by claiming that their hands were tied by Brussels.

amoe_

4 年前議論

> If what you want is... automated, predictable deployments... then Docker Compose and Terraform is all you need.

Can you elaborate on the deployment story here a bit?

amoe_

4 年前議論

f-strings are quite perlish IMV.

amoe_

4 年前議論

It's been a while since I last was experimenting with them. I did try Guice briefly, but I stopped for some reason that I don't recall now. Perhaps I should have given it more time. Dagger I tried in a bit more detail, and I found the code-generation aspect of it to be unpleasant and not sufficiently abstracted. I don't know if that could have been addressed by IDE configuration, plugins, etc, but I had to stop working on that project.

amoe_

4 年前議論

I like Spring, Spring Boot is good, and the libraries do stuff that you'll need (while maybe you don't realize whether you'll need it yet).

I'm a DI fan and I've found alternative DI frameworks to not work as well as Spring's DI. I'm only talking about Spring's annotation-based configuration.

There are also several good reasons not to like it, though: the errors can be unbelievably confusing, and it's quite hard to get a good mental model of how and why things are done in it.

コメント