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 -
Does anyone else feel like we're heading for a good old-fashioned, 2008-tier, financial crash? I recall people bringing up the possibility ever since Covid bucks started rolling out, but even though we were due for one, and even though the money printer was going brrrrr, the crash has so far failed to materialize.
At the time, I was of two minds about this. On one hand, all the libertarian theory I used to subscribe to said money-printing => boom, boom => bust. On the other hand, the problem for me was it never felt like a boom, and I think this is changing now. A key feature of the pre-crash boom is "malinvestment"; capital going into often downright deranged projects, that are later abandoned half-finished. Well, I feel like the datacenter craze qualifies, and with the wave of AI IPOs that are coming just as major indices are changing their rules, to allow for these companies' near-instantaneous inclusion (an investment so good, you can't pass up on it. Literally, if you're American), it seems like we're solidly in the "irrational exuberance" phase.
Add this to the list of things I hope I'm wrong about, because if we get a proper crash, the political fallout is going to be massive. The script writes itself: Trump / tech bros / capitalism bad, even more gay race communism now.
To whitepillers: is there an argument for why I'm wrong that doesn't boil down to "you don't get it, chud, it's the New Economy! The Singularity is just around the corner! All the rules are obsolete!"? This argument is verboten, because this is pretty much what people say with every bubble.
To fellow blackpillers: any ideas on how to brace for impact? Any IT guys here old enough to make it through the dotcom bubble? How did you do it? Any advice you would have given your past self?
If there's a crash, I predict it'll be due to energy prices and the war in Iran, not AI.
Anthropic is nearly profitable, or supposed to have been profitable in Q2. Some say that's phoney accounting and Anthropic says profits may not be maintained. But it seems that major investments are paying off. Furthermore, AI models are getting continually better as with Fable most recently. This trend will continue, bigger and better models working for longer need more compute to run them and so justify further intensifying investment. I guess that argument is forbidden by your post. But if it is The Singularity, if it is a New Paradigm then presumably that's good for stocks!
Anyway, energy is more important to the economy than AI right now. You can't just shut off a huge amount of oil and gas production without ramifications! The 1973 oil shock is a useful precedent. The market seems to have been expecting peace talks to advance more smoothly than they actually have been. Bombing has just resumed. Iran has announced they're re-closing the straits and possibly the other straits in the Red Sea too.
I'm suspicious about that given stories like this.
Maybe they had a revenue-positive quarter, but if it was buoyed by various companies overshooting their AI budget, this might not be an actual trend.
I've been suggested this was even an intentional move by an Anthropic investor to goose the company's valuation by artificially increasing their ARR (claims, not substantiated).
Not quite. The U.S. has been a net energy exporter since 2019. So domestic energy production can more than satisfy our needs for the foreseeable future.
That's before we get into the possibility of nuclear coming online.
Yes, higher oil prices impact the U.S. too, but we're not going to be the first ones to tap out, in either a literal or metaphorical sense.
Remember we just secured a massive source of nearby oil.
One thing about markets, they respond to shifts in supply and demand.
How do you even do that, are we supposed to believe that some company with $500 million to spend didn't know how to use cost tracking? Cost tracking comes with the pleb-tier management tools I get!
And if they were trying to goose the valuation then why would they admit it was unintentional, it makes Anthropic sound expensive? (They are expensive). It sounds more like an anti-Anthropic story to me.
The story sounds greatly exaggerated or misleading IMO. Who is this mystery company too, what are the details on this? More likely some company just consciously spent a lot of money on Claude AI and then some reporter fluffed it up into a narrative we all hear instead of a boring article that goes nowhere.
Regarding energy, even though the US is a major producer and isn't as badly affected as other countries like Australia, Europe or poor countries, the world economy is global. Problems in Asia will spread to America. The US is busily exporting oil, including the strategic reserve releases, to take advantage of price gaps and stabilize markets. But this is a temporary fix. The price gaps will narrow. The invisible hand of the market will slowly but surely squeeze the US economy if this price pressure remains. Even if oil producers profit, much more of the US economy relies on truck transport, plastics, feedstock, jet fuel, lubricants and all the other chemicals which support industrial civilization.
Venezuelan oil is low quality and requires years of patient investment and capable administration to realize much net gains from. Crude oil production there is barely half of what it was in 2016 and shows no signs of making up for the current supply loss. Nobody is a bigger fan of nuclear than me but nuclear energy is not going to lower fuel prices in a matter of months.
https://www.ceicdata.com/en/indicator/venezuela/crude-oil-production
One of my tiny clients was billed over $10,000 recently because their customer-facing chatbot got locked in a loop with somebody's openclaw agent.
An acquaintance works for a small to mid sized private software company, and he alleges that they put restrictions on AI use after they already hit seven figures this month.
Overall, I don't know if that story is true, but what I'm seeing on the ground strongly suggests that executives really don't have the slightest fucking clue what's happening until the bill lands on their desk.
This raises the question- how much does a token cost, and how much does it do? I'm given to understand the standard software engineering productivity line is 1 line of code/hour, is a token more or less than that(and does it cost more than an hour of a software engineer's time)?
This discussion of tokens is sort of like 'the carpet costs 40,000 Kazakhi tenges(to pick random foreign money I'm not familiar with)'. OK, sounds like a lot, but how much is a tenge? If it's worth as much as yen, that's not a bad deal.
[apologies if I'm answering a rhetorical question]
A 'token' is the ML equivalent of a syllable: the word, portion of a word, or symbol that represents the smallest viable input or output unit of effort. The exact value (and cost) depends on model, as well as whether it's input or output.
I'm ... skeptical to endorse line-of-code as a measure of programmer time -- I've spent days planning out business-critical logic that ended up five lines of code and needed to be absolutely correct, and spat out thousands of lines of text in an hour before when it was just boilerplate -- but the output side you can give a pretty good average. Depends on the tokenizer and your output language, but I'd expect less an average of less than 20 tokens per line in C++ or TypeScript, and I'd get worried if a human coder was regularly writing >80 tokens in a single line.
((edit: less so in java.))
So pessimistically, 12.5k LoC per million tokens, more realistically 25k.
Input is the high-variance part. If you're writing from scratch, the input is a few paragraphs and some design documents, maybe some scribbled image files if you feel spicy and the model supports it. I've done a few personal projects like that where it's been <2k tokens to get 20k line-of-code. If you have an existing codebase you want the model to adjust to, or an API document you need the model to learn, that can burn through a lot of tokens fast; the only real restriction is context window size, and most of the corporate APIs obfuscate that (tbf, often because they have an automated store and search strategy). I've blown through 50k in a single search once (thanks, Atmel, love your manual layout too). Input is typically cheaper and there's some strategies to reduce the cost of repeated input hits with the same content, but they're complicated and pretty specialized.
For some examples:
For smaller or more efficient models, inference is pretty cheap: Qwen3.6 35B's probably the weakest coding model I'd use in a professional environment (and borders the point where it might be better to run it locally, if only for privacy/security reasons), but there's a lot you can do.
That said, all of that can go out of the window when you start getting agentic options involved. Someone made a fun experiment of trying to let a local model figure out a display protocol by hooking a camera, an LLM, and a microcontroller together, and they got it mostly there overnight, which is really cool. It also probably burned tens of millions of tokens on output for an interface code block that should have ended up in the <1.5k line-of-code level.
So TL;DR, tokens are cheaper than basically all white collar workers in the modern west, but they get overused for fun projects?
More that they're cheaper than a code monkey, only weakly expensive if you're doing something hard or novel (or novel-to-you), and they can get ludicrously expensive if you just start firing the slop cannons, either to solve a problem by volume or by producing a lot of useless or specialized lines-of-code.
More options
Context Copy link
And because they allow brute-forcing problems in ways that weren't possible before, or tackling new problems. Or because the user cocks up, as in the case where they fail to notice their two bots getting into an infinite loop.
More options
Context Copy link
The problem is that a token is cheap, but the amount of tokens you need to do useful things can be very high.
For agentic programming, the agent needs to hold a non trivial amount of the codebase in the context. That can easily be millions of tokens. Then you have whatever pile of "skills" (read: markdown files) you use, then add the various layers of prompts, then add reasoning chains. It adds up very quickly.
Once you start adding parallel agents and loops, it can get insane.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link