12 ms·

If this can happen to Anthropic, imagine all the companies building on top of Claude Code for live products. Hopefully the industry is learning that competent p

If this can happen to Anthropic, imagine all the companies building on top of Claude Code for live products. Hopefully the industry is learning that competent problem solving human engineers are still very much needed when you have increasingly deceptive non-deterministic genies running your production stack.

gblargg 5mo ago
Maybe it will push companies to run them locally.
- tuwtuwtuwtuw 5mo ago
  Haha, good one.
- SilverElfin 5mo ago
  On what hardware? Like companies would buy up GPUs?
  - Hamuko 5mo ago
    Presumably you'd buy really beefy laptops. The price delta between buying the most basic MacBook Pro possible (14", M5, 16 GB unified memory, 1 TB SSD) and one with the M5 Max with 40 GPU cores, 128 GB unified memory, 2 TB SSD is $3400. How much Claude usage does that get you/in what time does it pay itself back?
    - squeaky-clean 5mo ago
      That doesn't get you any Claude usage. Claude models obviously aren't open, but equivalent models to Opus take about 400GB of memory to run. You can run the versions with fewer parameters or quantized weights but depending on how much quality you're sacrificing, now you'd have to compare the price against cheaper Claude models like Sonnet.
KaiShips 5mo ago
[dead]
samuelknight 5mo ago
It's not that simple. API is still up and there are multiple API providers. https://openrouter.ai/anthropic/claude-opus-4.7 https://openrouter.ai/anthropic/claude-opus-4.7
- varispeed 5mo ago
  The fact API is available, does not mean you will actually get the model it states you get. Today Opus 4.7 was noticeably dumber than yesterday. It performed worse than my local Qwen.
- subscribed 5mo ago
  I don't think there are many other companies serving Claude.
  - scottyah 5mo ago
    At least Google, Amazon, and Microsoft. What more do you want?
    - boleary-gl 5mo ago
      Came here to say this. At Kilo Code we aren’t impacted by this because of the other places that can run Claude
kinematikk 5mo ago
Sadly its "good enough" for execs

gblargg 5mo ago

Maybe it will push companies to run them locally.

tuwtuwtuwtuw 5mo ago
Haha, good one.
SilverElfin 5mo ago
On what hardware? Like companies would buy up GPUs?
- Hamuko 5mo ago
  Presumably you'd buy really beefy laptops. The price delta between buying the most basic MacBook Pro possible (14", M5, 16 GB unified memory, 1 TB SSD) and one with the M5 Max with 40 GPU cores, 128 GB unified memory, 2 TB SSD is $3400. How much Claude usage does that get you/in what time does it pay itself back?
  - squeaky-clean 5mo ago
    That doesn't get you any Claude usage. Claude models obviously aren't open, but equivalent models to Opus take about 400GB of memory to run. You can run the versions with fewer parameters or quantized weights but depending on how much quality you're sacrificing, now you'd have to compare the price against cheaper Claude models like Sonnet.

KaiShips 5mo ago

[dead]

samuelknight 5mo ago

It's not that simple. API is still up and there are multiple API providers. https://openrouter.ai/anthropic/claude-opus-4.7 https://openrouter.ai/anthropic/claude-opus-4.7

varispeed 5mo ago
The fact API is available, does not mean you will actually get the model it states you get. Today Opus 4.7 was noticeably dumber than yesterday. It performed worse than my local Qwen.
subscribed 5mo ago
I don't think there are many other companies serving Claude.
- scottyah 5mo ago
  At least Google, Amazon, and Microsoft. What more do you want?
  - boleary-gl 5mo ago
    Came here to say this. At Kilo Code we aren’t impacted by this because of the other places that can run Claude

kinematikk 5mo ago

Sadly its "good enough" for execs