Disagree with your second paragraph. Uber/lyft subsidized into deep negative margin territory around ~2015 or so. Anthropic (and OpenAI) are subsidizing but not losing money on these consumer plans.
The claim is that uber and lyft lost money on each ride but that openai and anthropic make money on inference. Just not enough to pay the cost of developing the models. The difference is that uber and lyft had to change their pricing (or payment) models to make money, where anthropic or openai could just sell enough inference (at some level of sales).
to clarify, I mean that the big model companies are charging consumers much less than equivalent API pricing but they're not actually losing money so its more of a steep at-cost discount for inference. It likely does not fully cover amortized R&D just like the other reply stated but it does still cover marginal inference cost (GPU/power/etc)
Uber and Lyft were paying drivers $X but charging users way less than $X so they were literally burning investor money to get market share.
> It likely does not fully cover amortized R&D just like the other reply stated but it does still cover marginal inference cost (GPU/power/etc)
Why does this point come up over and over again, pretending that you can truly separate training and inference costs. Yes, they are separate things but the value OpenAI and Anthropic are presenting to the world is "we're the absolute best, no one else comes close." Well, to keep that up you can't just not train for extended periods of time. You have to keep that engine going non-stop when there's free Chinese models nipping at your heels. You can be profitable on "just inference" all you want but if training expenses dwarf that, you're not going to be profitable overall, and that's the bottom line.
> does still cover marginal inference cost
Simply comparing to the larger models on OpenRouter implies that the pure hosting costs (equipment + power + minimal overhead) still exceed plan pricing if we assume all users always use their full weekly allowance.
So my conclusion is that it only works because the majority of users doen't fully utilize their allowance. Last month I used almost nothing of my Claude 20x plan (did use Codex though).
And I have always found that a little hollow, and depressingly capitalist within tech. Outside of tech, people who nearly exclusively value end-results are the same people taking pictures of their food, and the same people who can't sit in silence, or get irritable when a hike hasn't blessed them with a view yet. For me, it's almost always been the journey, and the result is just an ephemeral blip while work continues iteratively.
Sure, automating away things that would otherwise just be basically copy/pasted anyway have always been fine, but this is a very different scale.
It is interesting how people can differ in terms of disposition though.
Edit: I didn't mean to imply that one disposition is better or worse than another objectively, but rather that I subjectively find impatient, superficial, consumption-driven people extremely dull, which I also tend to associate with people who mostly care about the result. Additionally, I believe that this focus on the production of more and faster is a non-trivial contribution to producing things that have no longevity or are otherwise enshittified. The less we're involved in the intimate details of a thing, the less we're inclined to care, and the less it's even acceptable to care about quality.
I think I understand your point, and it certainly is true if you get in the right mindset with the right anectadata to back it up. But I feel it's leaving a lot on the table. Result oriented people are key to linking the technical with the real world. It's a bit of a wide statement to say that everyone who likes end result is impatient superficial or consumption driver.
I agree that there's a decent overlap here, and that your point stands. But there's a lot of merit in getting real products out there, vibe coded or not. Market forces then dictate value.
I think the primary goal is usually to sell things quickly and at bigger scales, not to make an empath feel something inside for the plight of the users. It's a lucky circumstance to be intimately working on a direct to customer product that you can see tangible gains from and the impact on someone's day. In those lucky positions I've been in where the user hasn't been obscured by business outcomes, part of the implied deal is still that I've found the mechanics of the job somewhat gratifying, otherwise it is indeed hollow in the sense that the middle part where you cut things with a metaphorical chisel goes away or changes sufficiently. It was the goal for many developers to be engaged in that way, not anyone else who didn't like that, and certainly not the product goal, but it was part of why those people did it, and it always created tension with those who exclusively care about numbers and had no technical expertise. If all you do is grind for outcome, then have at it, but I find that only minimally and rarely mentally engaging.
I find the current work extremely engaging. A customer can ask for a fairly complex feature and 25 minutes later, I’ll show it to them.
In the past, I’d have to put it on the roadmap, prioritize it, debate it. It’d take 25 days instead of 25 minutes. More often than not, we would have never done the feature because it was deemed not impactful enough for all customers.
In my early career I used to be the former. Now I'm the latter, and truthfully, probably have become a bit burned out by all the toil and friction, especially on the ops side of things.
Getting to the end product so much faster with so little friction really pushes the boundaries of what you can actually build. It's exciting (Also kind of ominous for obvious reasons)
Telling the AI to go, say... sort out all the IAM stuff for some new feature you're building is actually incredible.
Yes, your property tax is high ($17k) because of prop 13 - you are effectively subsidizing others who have sat on their residential and commercial property for 20-30+ years who are paying 25-50% of what you pay for any similar property in the same calendar year!
Your property tax goes up by 2% of your , guessing (relatively recent?) purchase price, and theirs goes up 2% of a much smaller base due to prop 13.
They have no incentive to move, downgrade. And they often use trusts to pass this tax base down to their kids. It’s feels like icky European feudalism.
Again i’m emphasizing that this goes for both residential and commercial buildings.
I agree with everything you’ve said, but one clarification: you don’t even need a trust to pass the tax break down to your kid when you die. Sure, a trust can streamline that process, but it’s not required. At least everyone lucky enough to inherit benefits, not just those whose parents had the foresight (and the means) to create a trust.
Who gets to define the market? Does the "market" consist of only advertising technology platforms, or everything involved in online advertising, or all advertising through every possible channel worldwide?
Because bad behavior tends to scale with company size, not because smaller companies are more virtuous, but because they just have less power to do harm than bigger companies.
Imagine that there were only two companies, and they each had 50% market share in every market that exists. Would you expect either to be well behaved?
