Post

Do you have recommended reading here? My understanding is that the subs are super subsidised indeed. But the actual API calls (which is what many enterprise customers are paying) don't immediately feel very subsidised to me. Are you saying AWS/GOOG are providing Anthropic with subsidised capacity?

I assume the plan is to get enough users that it gets embedded into so many things they essentially get a government bail out over and over.

Compare the cost of Microsoft’s Arm accelerator MAIA (coming up on gen 3) with general purpose inferencing infrastructure and you see why the compute cost of tokens is going down (but token usage is going up in probably a mirror image of the cost graph)

If this wasn't ironic, it still wouldn't be the worst theory I'd heard

IMO, we're 1-2 Chinese open-source quiet releases from everything collapsing. Just assuming some lab is going to come out with a Mythos-level model that can run on a Snapdragon out of nowhere, or we move on from transformer-based models into something completely different w/ diff compute needs.

The AI is going to invent cold fusion, and that will solve both the power and cooling issues!

Sam Altman ground Moore's law into a fine powder and snorted it, and thousands of credulous rich people across the country have decided that made him a genius.

Chips are getting faster, but they're not getting cheaper for it. The OG formulation of Moore's Law was cost per integrated component is cut in half every x months (the value of x fluctuated over the years).

I remember when there were huge billboards about $0 down mortgages and being like 'this seems like an issue' pre sub prime crash

And the answer from the True Believers was basically “volume!” which just meant the businesses based on wishcasting folded faster and more spectacularly, and billions of dollars of theoretical investment money evaporated overnight, which had, uh, some collateral effects

Can't talk about the economics of the whole thing but using the free Claude tokens via the web interface (no API, no agents, etc.) let me do in two months what before I couldn't do in 5 years, so it has been quite useful to me. Open-source projects, though, so I don't have any cost/benefit analysis.

I mean, Sam Altman did say their plan was to make ChatGPT smart enough to find a business model for them, and this does sound like the kind of business model ChatGPT would come up with

Ed Zitron, noted hater, has been yelling to anyone who will listen about this for ages. If Ed is correct, none of the maths makes any sense. www.wheresyoured.at/ais-economic...

Elect me and I’ll cut token prices by 600%!

In my field it is faster, better and (so far) cheaper than anything else before. We just structure data with it and it seems to have solved many parts of NLP. Smaller models are enough for this and it produces ROI in BPA. So, even without being a "fan" we have to use it.

It's not sustainable, unless you keep stealing money from the public purse which is what these techbros are doing with their alliance with Trump. Even that model is ultimately not sustainable but they will bleed us and planet dry until they get there (or we get there first with the guillotines).

Google "OpenAI training cost" or "Anthropic Operating Cost"

*targetes at enterprise

They don't want that, THEY want to control ALL OF THE DATA. If your data is stored offsite, it essentially belongs to whatever company is storing it.