This weekly roundup thread is intended for all culture war posts. 'Culture war' is vaguely defined, but it basically means controversial issues that fall along set tribal lines. Arguments over culture war issues generate a lot of heat and little light, and few deeply entrenched people ever change their minds. This thread is for voicing opinions and analyzing the state of the discussion while trying to optimize for light over heat.
Optimistically, we think that engaging with people you disagree with is worth your time, and so is being nice! Pessimistically, there are many dynamics that can lead discussions on Culture War topics to become unproductive. There's a human tendency to divide along tribal lines, praising your ingroup and vilifying your outgroup - and if you think you find it easy to criticize your ingroup, then it may be that your outgroup is not who you think it is. Extremists with opposing positions can feed off each other, highlighting each other's worst points to justify their own angry rhetoric, which becomes in turn a new example of bad behavior for the other side to highlight.
We would like to avoid these negative dynamics. Accordingly, we ask that you do not use this thread for waging the Culture War. Examples of waging the Culture War:
-
Shaming.
-
Attempting to 'build consensus' or enforce ideological conformity.
-
Making sweeping generalizations to vilify a group you dislike.
-
Recruiting for a cause.
-
Posting links that could be summarized as 'Boo outgroup!' Basically, if your content is 'Can you believe what Those People did this week?' then you should either refrain from posting, or do some very patient work to contextualize and/or steel-man the relevant viewpoint.
In general, you should argue to understand, not to win. This thread is not territory to be claimed by one group or another; indeed, the aim is to have many different viewpoints represented here. Thus, we also ask that you follow some guidelines:
-
Speak plainly. Avoid sarcasm and mockery. When disagreeing with someone, state your objections explicitly.
-
Be as precise and charitable as you can. Don't paraphrase unflatteringly.
-
Don't imply that someone said something they did not say, even if you think it follows from what they said.
-
Write like everyone is reading and you want them to be included in the discussion.
On an ad hoc basis, the mods will try to compile a list of the best posts/comments from the previous week, posted in Quality Contribution threads and archived at /r/TheThread. You may nominate a comment for this list by clicking on 'report' at the bottom of the post and typing 'Actually a quality contribution' as the report reason.
Jump in the discussion.
No email address required.
Notes -
What code is it? It certainly sucks at the stuff that I work on, to my dismay, and that's weird because my area of expertise is all very much solved and there's a bunch of reference books on it.
It's also very convincing but makes very bad mistakes when talking about philosophy (such as inverting Kant's position on a particular issue, doing the common misreading of Popper's paradox of intolerance or plain hallucinating sources for connections that don't exist).
In my experience it's a very advanced rubber duck and a crutch for boilerplate, but anything beyond that it's just plain bad at.
All you've done is start to believe that sufficiently advanced parrots are human. But they're not on close inspection, and that's only okay in some circumstances.
Here, since you asked for specifics, let me recount one of the most impressive conversations I had with Bing AI. (Unfortunately it doesn't seem to save chat history, so this is just paraphrasing from memory. I know that's a little less impressive, sorry.)
Me: In C++ I want to write a memoized function in a concise way; I want to check and declare a reference to the value in a map in one single call so I can return it. Is this possible?
Bing: Yes, you can do this. (Writes out some template code for a memoized function with several map calls, i.e. an imperfect solution).
Me: I'd like to avoid the multiple map calls, maybe using map::insert somehow. Can I do this?
Bing: Sure! (Fixes the code so it uses map::insert, then binds a reference to it->second, so there's only one call).
Me: Hmm, that matches what I've been trying, but it hasn't been compiling. It's complaining about binding the reference to an RValue.
Bing: (explanation of what binding the reference to an RValue means, which I already knew.)
Me: Yes, but shouldn't it->second be an LValue here? (I give my snippet of code.)
Bing: Hmm, yes, it should be. Can you tell me your compile error?
Me: (Posts compile error.)
Bing: You are right that this is an RValue compile error, which is strange because as you said it->second should be an LValue. Can you show me the declaration of your map?
(Now, checking, I realize that I declared the map with an incorrect value type and this was just C++ giving a typically unhelpful compile error.)
I want to emphasize that it wasn't an all-knowing oracle, and back-and-forth was required. But this conversation is very close to what I'd get if I'd asked a coworker for help. (Well, except that Bing is happy to constantly write out full code snippets and we humans are too lazy!)
I see, that is legitimately impressive and I get why Microsoft is rushing to integrate it into all their tooling. I am struggling to find a mental model of what it is and integrate it into a workflow.
Sounds like some sort of insanely well read but very dim intern that you can always ask to do anything through a computer or something. Very weird but probably very useful in a Jarvis-from-Iron-Man sort of way.
I'm concerned that this tech is still very much on lock in from giant corporations. Microsoft's Office integrations all seem to rely on spying on everything you do and those training costs are still too prohibitive for FOSS to remain competitive. I sure hope that changes.
Yeah, that's a pretty good description of it! I'm definitely still the brains of the outfit. But it's getting closer to the "Hollywood UI" ideal where you use your computer by talking to it rather than by remembering the correct syntax of a Unix command.
No argument here. I personally trust Microsoft a little more than Google, but still, I'm really hoping this tech gets democratized sooner rather than later. (I've heard Alpaca, which is small enough to run on a PC, is pretty good, but "pretty good" might not cut it.)
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
I don't ask it to write code then plunk it into my projects - I agree that it sometimes gets things wrong there (although you can point out errors and it'll acknowledge and often fix them). What I use it for is to talk through my problems (it's not a rubber duck, because it's replying with knowledge I didn't have before). It uses its vast breadth of knowledge to help me with things like syntax, library functions, simplifying code, debugging a compile error, etc. ChatGPT is bit rougher, but Bing AI has even been smart enough to challenge me when I'm giving it mistaken information, asking follow-up questions that get me to the root of my problem (like a coworker would).
So, I don't want really want to argue the Chinese Room philosophy of when language understanding starts to "count". All I know is what my lying eyes are telling me: I'm now conversing with my computer in completely natural language, and it hasn't once failed to understand me. (Its reply hasn't always been helpful or right, but it's always made sense.) It's important to resist the cynicism of finding ways to break the LLM and going "oh, it's lame after all". Even if LLMs somehow never get any smarter, even if they're not on the critical path to AGI, just the capabilities we've already seen are enough for them to change the world.
More options
Context Copy link
More options
Context Copy link