Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

I think Anthropic and OpenAI have found product-market fit

simonwillison.net

Post reply on HN
Print view View original

591–600 of 1001 posts

Re: I think Anthropic and OpenAI have found product-market fit

#591

Post by timmg »

sixhobbits wrote: post #216
aerhardt wrote: post #132

I find this analysis confusing. PMF for coding was likely reached some time last year. Profitability, which is different, we don’t know. The article kind of confuses both without making a strong economic case or using numbers in a compelling way. I don’t understand what the Uber case has to do with this either. The Uber COO clearly said that at least in terms of ROI he’s not seeing the results either. My take is the…

Pmf is this weirdly defined thing where "if you're not sure you have it then you don't". I think it was clearly useful for months to people who had tried it and taken the time to understand it, but now that knowledge has spread to the point where wallet holders are convinced it's not just passing fad or hype so now pmf can be "claimed". I agree it's weird to say "those people have pmf" though, usually it's something…

> Pmf is this weirdly defined thing where "if you're not sure you have it then you don't".

I'm not sure if this runs counter to your point or not, but: I don't see any future where LLMs aren't a core part of Software Engineering. The horse is out of the barn. There is no going back.

Re: I think Anthropic and OpenAI have found product-market fit

#592

Post by whatshisface »

thesz wrote: post #443
whatshisface wrote: post #227

Earlier quoted context omitted.

Inference has traditionally been far less expensive than training. One public example is the fact that hobbyists can run StableDiffusion ($600k training costs[1]) on their personal computers. Speaking to your point, inference being dramatically less costly than training would not be seen as a delta from the norm. The model of providing inference for anything near the operational costs (like a utility would), would th…

The difference between training and inference is 1) one have to keep intermediate results for backward pass in training and 2) computation for training double because of the backward pass. Training is also done over batches, which increase memory requirements by several orders of magnitude. This is why training needs costly compute. One of the ways out of this unfortunate situation is to use something like Stochastic…

That is an estimate of the relative cost of one training step, but you have to multiply it by the number of training steps, an unknown quantity.

Re: I think Anthropic and OpenAI have found product-market fit

#593

Post by simonw »

shimman wrote: post #589
simonw wrote: post #533

Earlier quoted context omitted.

What's the psychosis?

That leaders are completely fine with impoverishing vast swaths of American workers because of "progress?"

If they are then yes, that's psychotic. Not sure how it's relevant to my article about Anthropic and OpenAI's enterprise pricing though.

Re: I think Anthropic and OpenAI have found product-market fit

#594

Post by overgard »

brokencode wrote: post #43
antman wrote: post #33

The costs are exorbitant and most software is not produced by companies with such a huge moat. Anthropic made a profit through their recent bait amd switch pricing. There is zero useful insights online to indicate whether this might die due to commoditisation with good enough open models or fail the race to get more people subsidising unsustainable growth with other people’s money. Who knows? In any case they dont se…

Usage costs will come down with better hardware. Hardware is improving rapidly each generation.

A lot of the new hardware requires retrofitting existing datacenters for appropriate cooling, or is waiting to be installed because the new datacenters haven't been built yet. By the time they're installed it's likely a lot of Blackwell GPUs are going to be very out of date. Newer hardware is turning into huge capex bills along with the corresponding depreciation costs. Basically, it's not the same as plugging a new GPU into your desktop, the upfront investment is extremely expensive and all the numbers I'm seeing suggest that the newer GPUs are costing more to run, not less.

Re: I think Anthropic and OpenAI have found product-market fit

#595

Post by Gigachad »

spamizbad wrote: post #285
trjordan wrote: post #11

They've got, ballpark, $5t to $10t to make back in the next 5 years, or the hardware buildouts will start getting written down. This means we're going to need $1t+ per year in spending, per year, on tokens. 200m knowledge workers in the world, 30m developers. We're talking about a world where you need 5% of every knowledge workers salary to go into tokens. 20% if you're a developer. That's a _huge_ shift. Most people…

I will also tell you, as someone who works at a company that's trying to remain profitable, that token spend has caught the eyes of finance and much like cloud spend they've already started applying pressure to control costs. This May my team is protected to use 30% fewer tokens than we did in April - this was by intention. I suspect we'll drop more in June.

I expect in the future, when these AI companies stop subsidizing costs, the idea of spinning up 20 agents to work on some brain fart idea that you throw out after looking closer will come to an end. It'll be seen like assigning developers on work that hasn't been properly planned for or reviewed.

Re: I think Anthropic and OpenAI have found product-market fit

#596

Post by fragmede »

spamizbad wrote: post #285
trjordan wrote: post #11

They've got, ballpark, $5t to $10t to make back in the next 5 years, or the hardware buildouts will start getting written down. This means we're going to need $1t+ per year in spending, per year, on tokens. 200m knowledge workers in the world, 30m developers. We're talking about a world where you need 5% of every knowledge workers salary to go into tokens. 20% if you're a developer. That's a _huge_ shift. Most people…

I will also tell you, as someone who works at a company that's trying to remain profitable, that token spend has caught the eyes of finance and much like cloud spend they've already started applying pressure to control costs. This May my team is protected to use 30% fewer tokens than we did in April - this was by intention. I suspect we'll drop more in June.

Can't wait till June, when finance gives the team the choice: everyone gets double tokens if you choose to fire somebody.

Re: I think Anthropic and OpenAI have found product-market fit

#597

Post by somewhereoutth »

whatshisface wrote: post #173
trjordan wrote: post #11

They've got, ballpark, $5t to $10t to make back in the next 5 years, or the hardware buildouts will start getting written down. This means we're going to need $1t+ per year in spending, per year, on tokens. 200m knowledge workers in the world, 30m developers. We're talking about a world where you need 5% of every knowledge workers salary to go into tokens. 20% if you're a developer. That's a _huge_ shift. Most people…

Here are a few thoughts: - The publicly available information about how inference costs compare to training costs is conflicted. EEs involved in datacenters talk about power usage spikes during training runs as if they were a major factor in the designs, but academic papers discussing cost-optimal scaling confidently treat inference-time compute as a major factor. - On the side of the balance indicating that training…

Yes the huge discrete stepwise training spend is critical.

Maybe investors will realise that "the only winning move is not to play".

And so we are left with (as was) frontier models getting more and more out of date as whoever their post bankruptcy custodians are tries to eek pennies on the dollar for inference on their decaying property. Perhaps along with local and/or highly specialized models still feeding on the after-glow of the huge amount of training that was (and is no longer) done.

The next AI winter is going to be deep, savage, and long.

Re: I think Anthropic and OpenAI have found product-market fit

#598

Post by tredre3 »

sourcecodeplz wrote: post #547
jillesvangurp wrote: post #476

Earlier quoted context omitted.

Most of the corporate world in the EU or North America will be hesitant to rely on Chinese AI providers. There are some very real blockers for that for things like data security, compliance, etc. And recent geopolitics don't help. Legalities aside, you need to look not at the model quality but at the infrastructure needed to scale these models from tens (now) to hundreds (soon) of millions of users. Only a handful of…

Deepseek and all the other Chinese models have open-weights. You can host them yourself, no need to send data to China or rely on them.

There is still a risk of supply-chain attack. People give LLMs direct access to their entire infrastructure via tools, and never check the code produced. It's not difficult to steer an LLM during training so that they'd output malware only when prompted a certain way, and that wouldn't come up during the initial evaluation.

Personally I see no difference between China and America in terms of risks of them embedding "backdoors" so to speak, but I disagree when people claim that open-weight models are obviously safe just because they can be ran locally.

Re: I think Anthropic and OpenAI have found product-market fit

#599

Post by rogerrogerr »

nilamo wrote: post #451
seanp2k2 wrote: post #141

Earlier quoted context omitted.

>The bottleneck has moved from producing a thing that works to knowing that the thing was the right thing to build I would argue that that's been the case for quite some time before AI. As an example, what innovative amazing world-changing products have Google or Meta launched in the past decade with their very high numbers of very talented and highly-compensated engineers? The issue with most big tech companies are…

> As an example, what innovative amazing world-changing products have Google or Meta launched in the past decade Kubernetes is at 11 years ago, and is huge enough to be included there. The Google Pixel was just under 10 years ago. So... not nothing haha

Numbers I see put Pixel at less than 5% of iPhone sales. Not nothing, but world-changing? I doubt the world would look significantly different had Google not done Pixel.

Re: I think Anthropic and OpenAI have found product-market fit

#600

Post by decidu0us9034 »

dirck-norman wrote: post #187
specproc wrote: post #124

Earlier quoted context omitted.

I'm sorry, what the feck does "value creation" mean here? I live in a place where people are so, insanely squeezed from every angle. Wages are stagnant, prices rocketing. Where is the money to pay for this value going to come from? No one I know feels richer than they did a decade back. I've not been able to meaningfully put up my prices for a decade. People are tired and stressed and scared, particularly scared of a…

Feelings aren’t fact. A lot of data shows the doomerism is not reflected in the actual numbers and much of it has to do with rapid inflation and continued vibes. Consumption has risen, inflation adjusted wages have risen for blue collar and white collar alike. Most social mobility has been the middle class moving into the upper middle class, not moving to the lower class. The main thing holding people back is the hou…

The upper decile of income earners account for more than half of all consumption in the US. Household balance sheets and wealth have never looked stronger, again when you account for all the appreciated stocks and properties owned by the upper quartile. True incomes for the lowest decile rose significantly for the first time since 1970 in 2022 and then sort of stayed flat again. Sure, statistically significantly, not "significantly" as in personally meaningful after figuring in rising consumer costs. There is a narrative where you can see all this as hugely positive but this is also largely a "vibes" based narrative. I don't know why you'd expect most people to care about what the "vibes" are like for the best off in society, that's a bit removed from their daily concerns.
Post reply on HN