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 -
I really don’t think this is necessarily about the big frontier labs, there are often a number of layers between them and the creditors for these huge data center projects (in fact a lot of smart treasury and finance people at Meta, Google, Amazon, OpenAI etc have taken huge advantage of the private credit bubble and general syndicated debt market hype for AI and set up the funding such that investors will have essentially zero recourse to them if they decide they don’t need the compute; coreweave might go out of business but they won’t).
It’s about the fact that a lot of inference is essentially more about the layer of computed-human or AI-human or human-AI-human interaction than it is about the kind of work that a fully automated system does. I don’t think it’s as easy as the comparisons you draw. If you want a kind of dumb/funny example imagine if we’re in some kind of premodern agricultural scenario with LLMs (and literacy). We might actually use a lot of inference, send a lot of emails, we need a summary of the meeting about worker morale on the strawberry field, barley yields have been low this year due to slacking, Martin needs to stop spreading his weird disease, you two need to read up on crop rotation. This is all kind of slopwork. Now we replace fifty workers with one guy and some modern farm machinery, objectively the inference done is much lower. That’s true even if we replace that one guy with a multimodal combine harvester robot etc etc. Commoditization is more of a problem for compute than it is for the model providers. I used to agree with you and argued that view here extensively, but I think Mythos shows you that if you have even the hope of a true frontier model that has capability that no other model does you’re going to be able to extort entire sectors that rely on security especially (banks, defense, governments) at insane margins until everyone catches up. Most LLM work will be commoditized but the frontier release payoff will be high enough to keep the funding coming for the biggest players. Tokens/task is a bad metric, so we can use fully amortized compute (including across training/research costs) or whatever else you prefer.
This ignores a really interesting scenario where AI, being vastly cheaper and soon better than human coders, is able to write and test hugely complex software for a lot of these use cases that would be completely economically ridiculous today, but which will get cheaper over time, and then leash these to relatively low-intensity agents that use these tools. The simple argument is that instead of using Claude to compute 2+2 a million times, we just get Claude to code a calculator. You kind of dismiss this but I think a more fully featured version of this argument is actually quite compelling, especially when you count unfathomably wide-ranging improvements in token use efficiency that are coming not just for text but multimodal applications too. The US uses as much oil today (about 15-20 million barrels a day) as we did in the 1970s. Resource consumption numbers don’t just go up.
It’s sad, I’ve given it some of my recent posts and drafts (and random unpublished things I might get around to finishing at some point) and it doesn’t identify me (or a lot of other users here). There aren’t many (identified, I guess) NHS doctors in this sphere so I guess it’s a small world.
That's not the intention behind my argument really. People are using Claude to code a calculator (and that was something you could have done a year or two back), it just doesn't make sense when we already have perfectly adequate human-designed calculators.
But put your ears (?) to the grapevine and you'll see that people are making all kinds of toys, bespoke bits of standalone software that AI enabled them to do. Are they world-changing, yet? Probably not. But the proof of principle is there. Notice that I've called them toys, even if some of these things are legitimately valuable for their creator or people with similar, bounded but under-serviced use cases. I collect these things on X, though I'm too tired to present examples.
Of course, that is today AD. I have no reason to dispute the claim that in the near future, far more sophisticated and immediately compelling software artifacts won't be abundant, but their commercial moat will be nonexistent, since any other Claude Code Monkey should be able to replicate them in a fast-follower fashion.
And implicitly, I've accounted for larger models coordinating agentic swarms. Mythos 2 ordering around a bunch of Sonnet 5.2s and Haiku 5.1s to manage the grunt work. Humans already do this, and I've seen the benefits after a month of extensive practice with agentic orchestration.
Here, my reply would be that in the near to medium term (2-5 years), the human aspect will be severely deprecated. It won't be a lawyer writing an LLM brief that another judge uses an LLM to explain. That's a very transitional stage, though it's anyone's bet how long that state of affairs will last with protectionist and credentialist regulations at play. As someone who worried that ChatGPT can replace me at 80% of my job, I can't complain too hard about the extra time, money and job security.
This is the kind of inference that will die. Eventually. My point is that it's like people using email to send each other scanned documents, signing them, and sending them back. A short, stupid stage that won't last. But more streamlined and coherent systems only drastically increased the value of email.
You'd previously said you didn't want to know if it could identify you. I presume that's changed? Because it can. In incidental conversation, it knows who you are as "2rafa", and it definitely knows you're a woman. You crop up in discussions of the Motte all the time as a "valued contributor", a framing I can't disagree with at all. Beyond that, I've tried to respect your privacy and didn't outright check but I expect to see interesting things.
It's not even the NHS! I had a big debate with @Shrike about... alien civilizations. Just those samples of my text pegged me as self_made_human with Claude reporting a subjective 50-60% confidence. And guess what it gets it more right than not. I'm usually the top pick, even if it worries if it's missed someone else. Bridge Mormons? Oh, that's obvious too. I've tested on samples that minimize PII or obvious interests.
The problem is that everyone catches up fast. I don't know how closely you followed Mythos and recent events, but OAI made a big deal about how GPT 5.5 Thinking was just as good at cybersecurity, for much cheaper, and most importantly available to the public for $20. They joked that Dario was doing a big safety jig to avoid the uncomfortable possibility that Anthropic simply can't afford to sell Mythos at scale, they lack the compute. This might change with the new xAI tie up, Elon prefers that Dario wins if that means Altman loses.
Apologies if I've missed anything, wasn't kidding about the migraine, and the meds plus sequelae have me loopy. I hate my job. Take it away, as long as I get paid. Jokes aside, tell me Ive ignored something or looked over something else, I go above and beyond when talking to you.
More options
Context Copy link
More options
Context Copy link