Great piece. Have you noticed the fury building amongst Anthropic's customer base, exhausted and stymied by its customer service robots? You're a CTO and front office have blown through their inference budgets by 300% and you need to discuss something with a human at Anthropic. Good luck.
The speed and efficiency of a firm’s ability to create new products is an indication of how well engineered the data is. AI/ML doesn’t change this.
The greater the ease and speed of new product creation, the more cost effective and profitable the business will be - this was understood by the software industry back in the 1900’s - I feel like all the excitement about AI and LLM has caused investors to forget the fundamentals.
Really interesting article, and a really strong insight about who wins in this model. I think you're right that buying outcomes will be the easier choice for most people looking to integrate AI, since it's a fixed cost, but it would be interesting to see how those cost structures would work.
Súper
Great piece. Have you noticed the fury building amongst Anthropic's customer base, exhausted and stymied by its customer service robots? You're a CTO and front office have blown through their inference budgets by 300% and you need to discuss something with a human at Anthropic. Good luck.
I’ve paid for a subscription, but don’t have access. Any suggestions? Tried the Substack chat, and did not have a great experience.
Tokens got cheaper, but nobody re-architected the workflows built when they were expensive—that’s the actual budget leak, not the API pricing.
The speed and efficiency of a firm’s ability to create new products is an indication of how well engineered the data is. AI/ML doesn’t change this.
The greater the ease and speed of new product creation, the more cost effective and profitable the business will be - this was understood by the software industry back in the 1900’s - I feel like all the excitement about AI and LLM has caused investors to forget the fundamentals.
Really interesting article, and a really strong insight about who wins in this model. I think you're right that buying outcomes will be the easier choice for most people looking to integrate AI, since it's a fixed cost, but it would be interesting to see how those cost structures would work.
This landed for me as a Lovable lover.
I’ve built three projects with it, and what I feel most clearly is the shift from tokens to outcomes.
But when the machinery disappears for the user, the ledger does not disappear.
It just moves somewhere harder to see.