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 -
Prediction: We are heading for an AI-middle ground boom.
Predictions regarding AI tend to cluster in two extremes, those who believe we are on a parabolic arc towards super-intelligence and those who believe AI just produces slop. The pessimism regarding AI clusters around to conflicting narratives, it is so good it will cause mass unemployment and AI is so useless that the AI bubble will pop. My take is that AI is mid and that is a good thing. Gemini 3 and Claude 4.5 are useful. However, since LLMs are limited by context windows, social skills, an inability to learn after training and don't have human judgment they can't replace us. Both doomer narratives are false, we aren't going to replace all the software developers with claude, and claude is not so useless that users will abandon it, causing the AI bubble to pop.
The AI speedup is more than worker speed being improved by LLMs. Many projects are stalled waiting for someone else to complete a task. A typical corporate scenario is that someone works an hour on something, emails it to someone who waits a week before working an hour on it, and then sends it off to the next person. With LLMs enormous speed ups can be achieved by not having to wait for answers.
AI is much more than large language models. It has long been used in areas like weather prediction, and over the past decade its capabilities have advanced dramatically. Twelve years ago, AI systems struggled with basic image recognition tasks such as distinguishing cats from dogs; today, they can reliably detect subtle anomalies on factory floors. AI is now widely applied in biotech, scientific research, mining, oil extraction, fraud detection, and many other fields.
What once required a machine-learning PhD can often be accomplished in a matter of days by a technically competent practitioner using cloud platforms such as Google Cloud. While humanoid robots have captured public attention and robot butlers remain unrealistic, AI is already accelerating the deployment of industrial robots and other forms of automation. Advanced driver-assistance systems are reducing the risk of traffic accidents, and AI is speeding up academic work and scientific discovery. More broadly, AI excels at uncovering patterns in massive datasets and surfacing insights and information that would otherwise remain hidden.
Scientific work is iterative. Progress is built upon earlier progress, and one bottleneck in a chain of discoveries prevents the subsequent discoveries from happening.
If AI can unlock a few bottlenecks, that could unlock subsequent discoveries that depend upon them. We could see a small jump in scientific discoveries.
Predictions for the culture war:
We are not going to see mass unemployment, even if a few sectors end up being impacted. Smaller organizations are more nimble and able to react to changes while having similar access to AI as large corporations, this benefits small players. AI deflationary as the cost of production go down. AI in ecommerce is making the field even more cut throat driving prices down. Low inflation will cause low interest rates and high asset price inflation. The economy is going to have wind in its back over the next decade as productivity rises. Government is going to be worse at utilizing AI than the private sector leading to an increasing view of the government as incompetent and falling behind.
Usually, the one guy on the team not using an LLM.
I swear, people who are not me have to have been using entirely different LLMs than I have. Every time I've used them for anything beyond the super trivial, I get results that are missing major components, or don't solve the business requirements, or contradict themselves. Like, I just opened up cursor and asked it for the CSS to render rounded corners in outlook, and it got it entirely wrong (it produced non-rounded objects because it used border-radius, which anyone who has coded anything for outlook knows doesn't work). When I told it that, it produced different code that (and I can't stress this enough) still fucking used border-radius.
Seriously, at this point I'm more likely to say that the person using the LLM is going to stall the project; they're going to produce verbose but contradictory requirements, they'll produce code that is written fast, but doesn't actually work, or they'll use it to answer emails in such a way that it doesn't actually answer the question that is stalling the damn project.
All I can say is that my company pays for the service. I'm a weird use-case because I use it a lot for code context, light scripting and other ops/dev/ops tasks. I"m not a dev. Anyway, I asked Claude-Code to describe itself:
More options
Context Copy link
It also really depends how you use them - a lot of people open a chat and then ten messages and a bunch of code lookups later they ask for something, then they don't like it so they ask for a correction, the correction is bad and they complain and ask for another correction, etc. So you have 30,000 tokens or more, containing a bunch of broken code that you don't want. Some people use the same chats for days or weeks.
Important points:
But yeah, sometimes the LLM just derps.
More options
Context Copy link
I think somebody here said a while ago that the paid LLMs are leagues more competent than the free ones, so if you're using a free one you may not be getting the full picture.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link