This is the modern version of 2010's "Amazon makes no money in retail"

But isn't it the case that the cash cow for Amazon is actually AWS? That is, without AWS would we consider Amazon to be the success we see it as today?

Yes we would.

Keep in mind that while the store brings in a majority of revenue, it provides less than 20% of operating income. A majority of operating income comes from AWS (which provides less than 20% of revenue).

https://www.fool.com/investing/2024/01/10/amazon-e-commerce-...

Forget about AWS and simply compare Amazon with other offline/online/mixed stores.

"In 2023, Amazon accounted for 37.6% of all e-commerce sales in America, which was light-years ahead of second-place Walmart at 6.3%."

Amazon is highly successful even without AWS.

I see your overall point, i think it is the wrong comp. It compare more to SAP or Atlassian.

I'm curious how you put OpenAI, SAP & Atlassian all in the same category. Do you mind explaining your thinking?

It's a comparison, not a same category.

Leaders in their market, at first because of a small moat and large advance, then the advance shrink but the brand will stay first for at least a decade due to brand recognition.

The Amazon comp works less, unless openAI start to suddenly build the sell their own chips.

Amazon weren't making money because they were expanding rapidly and had a lot of capex.

OpenAI don't make money because they're losing money on inference. If you remove their capex - training costs, which you can't because they have to remain competitive, they'd still be losing money.

My understanding is that they sold a large chunk of the business to Microsoft in exchange for nearly unlimited Azure credits

Before I read this 32-minute long article, who is Edward Zitron and why should I care about his opinion or analysis?

From scrolling though his articles he is a guy who likes to write polemic articles declaring companies dead

Zitron is a blogger whose content/internet personality is centered around being anti-Big Tech, and very much falls in the "AI is dumb/useless and will die any day now".

He's a good writer, but his content is written through an extreme anti-AI lens, so take it with a pretty big grain of salt.

Ed Zitron has written many newsletters chronicling the business economics of OpenAI/other AI companies and their incentives.

They're the most comprehensive writeups in that aspect, although not as technically-oriented as most content on HN.

It would have been easier to Google him, or ask Claude/Mistral to write you a summary of his article.

I hadn't heard of the author before that article, but all the math and logic in it is sound. There is some guesswork on exact numbers (because OpenAI don't publish that sort of information) you can disagree with, but IMO it's fairly generous in their favour. And it still looks pretty bad for their financial prospects.

We're here for human interaction, it's pretty condescending to tell someone to "Google it".

If someone has 0 understanding of what an embedding model is used for, I seriously question their ability to predict winners and loser in AI.

> OpenAI has a "text embeddings" API that is used primarily for tasks where you want to identify anomalies or relationships in text, or classify stuff in text"