I don’t disagree! But it’s a good go-to-market to cast a wide net and then optimize. To pretend like having poor unit economics for some use cases at launch is a guarantee for overall business failure is to be bad at analyzing businesses.

It’s definitely going to be a dance, but I use Claude Code all throughout the day and the throttling is fine from my perspective. They’ve done a good job at making it still good (enough)

I mean cheaper than value. The $200/mo is probably worth ~$4-5k/mo to us.

The inflated stock prices and the circular deals are definitely not going to turn out that well, but the idea that LLMs cost more to operate for valuable use cases is just nonsense.

I mean, I think the cost concerns on a lot of these are wildly overblown. Reminds me of all the think pieces about Facebook. Claude Code is dramatically cheaper than a developer that does the same work.

When I write, even advocacy pieces, I make significant efforts to represent counterpoints, and I do a lot of work to make sure I don’t misrepresent facts. Ed fails at both of these.

Anil’s piece isn’t wrong but also feels vague and selection-bias-impacted.

I dunno; this screed is completely anti-LLM, unfocused, and just wrong in many places: www.wheresyoured.at/the-case-aga...

My invitations to both remain open to show them how much efficiency gains LLMs provide in software development. (It’s a lot! And it’s changing how we build software, for the better [at least in terms of better software])

/3 I am informed it’s normal to get more intolerant as I get older and I think I am absolutely getting less tolerant of other people’s absolute bullshit.

Happy to pair on software dev and show you how Claude Code has changed how we do things (for the much better).

Such a wonderfully pithy driveby