4 ms·

Given the alleged recent extreme reduction in Claude Code usage limits (https://news.ycombinator.com/item?id=47739260 https://news.ycombinator.com/item?id=47739

Given the alleged recent extreme reduction in Claude Code usage limits (https://news.ycombinator.com/item?id=47739260 https://news.ycombinator.com/item?id=47739260), how do these more autonomous tools work within that constraint? Are they effectively only usable with a 20x Max plan? EDIT: This comment is apparently [dead] and idk why.

breakingcups 7mo ago
You seem to be vouched for now, no longer dead for me.
- minimaxir 7mo ago
  Hmm, I can't edit the original comment to retract that edit either. Either my account is flagged for something or HN is being weird.
  - TacticalCoder 7mo ago
    Everything looks good to me: you don't look like you have a flagged account (but then I don't work for HN).
  - 0123456789ABCDE 6mo ago
    this comment was made almost 1h after the one at the root of the thread you can make changes to your posts up to 10 minutes after they were originally created — see: https://news.ycombinator.com/newsfaq.html#:~:text=minutes%20you'd%20like.-,The%20maximum%20is%2010 https://news.ycombinator.com/newsfaq.html#:~:text=minutes%20...
    - minimaxir 6mo ago
      That is not what that feature does, it just is the time your comment is publically shown. Users have 2 hours to edit comments, and the button was gone within 1 hour.
giancarlostoro 7mo ago
I've been talking to friends about this extensively, and read all sorts of different social media posts on X where people deep dove things (I'm at work so I don't have any links handy - though I did submit one on HN, grain of salt, unsure how valid it is but it was interesting: https://news.ycombinator.com/item?id=47752049 https://news.ycombinator.com/item?id=47752049 ). I think the real issue stems from the 1 Million token context window change. They did not anticipate the amount of load it would give you. That first few days after they released the new token window, I was making amazing things in one single session from nothing, to something (a new .NET based programming language inspired by Python, and a Virtual Actor framework in Rust). I think since then they've been trying too many things to tweak things, whilst irritating their users. They even added a new "Max" thinking mode, and made "High" the old medium, which is ridiculous because you think you're using "High" but really you're not. There's a hidden config file to change their terrible defaults to let Claude be smarter still, and apparently you can toggle off the 1M tokens. I think the real fix, and I'm surprised nobody there has done this yet, is to let the user trim down their context window. Think about it, you used to have what? 350k tokens or so? Now Claude will keep sending your prompt from 30 minutes ago that's completely irrelevant to the back-end, whereas 3 months ago it would have been compacted by now. Others have noted that similar prompting for some ungodly reason adds tens of thousands of extra garbage tokens (not sure why). Edit looks like someone figured out that if you downgrade your version of Claude Code and change one single setting it unruins Claude: https://news.ycombinator.com/item?id=47769879 https://news.ycombinator.com/item?id=47769879
- dacox 7mo ago
  Yeah, I have been seeing lots of comments, tweets, etc, but given everything I have learned about these models - i do not think the change to 1M was innocuous. I'm not sure what they've claimed publicly, but I'm fairly certain they must be doing additional quantization, or at minimum additional quantization of the KV cache. Plus, sequence length can change things even when not fully utilized. I had to manually re-enable the "clear context and continue" feature as well.
  - giancarlostoro 7mo ago
imhoguy 7mo ago
AI race to the bottom is a debt game now. Once the party is over somebody will have to pay the bill.
- timacles 7mo ago
  It’s going to be crazy with the explanation they come up with why the us public has to pay to bail out AI for national security. In a way, it’s true if china has superior AI then it’s dominance over US will materialize. But it’s not hard to see how this scenario is being used to essential lie and scam into trillions of debt. Its interesting how the cutthroat space of big tech has manifested into an incidious hyper capitalist system where disrupting a system is it’s primary function. The system in this case is world order and western governments
  - joquarky 7mo ago
    "Move fast and break things" broke containment to the tech industry. Now you can see it everywhere.
stavros 7mo ago
It's not alleged: https://www.ghacks.net/2026/03/27/anthropic-reduces-claude-session-limits-during-peak-hours-for-free-pro-and-max-users/ https://www.ghacks.net/2026/03/27/anthropic-reduces-claude-s...
- minimaxir 7mo ago
  That's a separate change to what the linked HN post described.

breakingcups 7mo ago

You seem to be vouched for now, no longer dead for me.

minimaxir 7mo ago
Hmm, I can't edit the original comment to retract that edit either. Either my account is flagged for something or HN is being weird.
- TacticalCoder 7mo ago
  Everything looks good to me: you don't look like you have a flagged account (but then I don't work for HN).
- 0123456789ABCDE 6mo ago
  this comment was made almost 1h after the one at the root of the thread you can make changes to your posts up to 10 minutes after they were originally created — see: https://news.ycombinator.com/newsfaq.html#:~:text=minutes%20you'd%20like.-,The%20maximum%20is%2010 https://news.ycombinator.com/newsfaq.html#:~:text=minutes%20...
  - minimaxir 6mo ago
    That is not what that feature does, it just is the time your comment is publically shown. Users have 2 hours to edit comments, and the button was gone within 1 hour.

giancarlostoro 7mo ago

I've been talking to friends about this extensively, and read all sorts of different social media posts on X where people deep dove things (I'm at work so I don't have any links handy - though I did submit one on HN, grain of salt, unsure how valid it is but it was interesting: https://news.ycombinator.com/item?id=47752049 https://news.ycombinator.com/item?id=47752049 ). I think the real issue stems from the 1 Million token context window change. They did not anticipate the amount of load it would give you. That first few days after they released the new token window, I was making amazing things in one single session from nothing, to something (a new .NET based programming language inspired by Python, and a Virtual Actor framework in Rust). I think since then they've been trying too many things to tweak things, whilst irritating their users. They even added a new "Max" thinking mode, and made "High" the old medium, which is ridiculous because you think you're using "High" but really you're not. There's a hidden config file to change their terrible defaults to let Claude be smarter still, and apparently you can toggle off the 1M tokens. I think the real fix, and I'm surprised nobody there has done this yet, is to let the user trim down their context window. Think about it, you used to have what? 350k tokens or so? Now Claude will keep sending your prompt from 30 minutes ago that's completely irrelevant to the back-end, whereas 3 months ago it would have been compacted by now. Others have noted that similar prompting for some ungodly reason adds tens of thousands of extra garbage tokens (not sure why). Edit looks like someone figured out that if you downgrade your version of Claude Code and change one single setting it unruins Claude: https://news.ycombinator.com/item?id=47769879 https://news.ycombinator.com/item?id=47769879

dacox 7mo ago
Yeah, I have been seeing lots of comments, tweets, etc, but given everything I have learned about these models - i do not think the change to 1M was innocuous. I'm not sure what they've claimed publicly, but I'm fairly certain they must be doing additional quantization, or at minimum additional quantization of the KV cache. Plus, sequence length can change things even when not fully utilized. I had to manually re-enable the "clear context and continue" feature as well.
- giancarlostoro 7mo ago

imhoguy 7mo ago

AI race to the bottom is a debt game now. Once the party is over somebody will have to pay the bill.

timacles 7mo ago
It’s going to be crazy with the explanation they come up with why the us public has to pay to bail out AI for national security. In a way, it’s true if china has superior AI then it’s dominance over US will materialize. But it’s not hard to see how this scenario is being used to essential lie and scam into trillions of debt. Its interesting how the cutthroat space of big tech has manifested into an incidious hyper capitalist system where disrupting a system is it’s primary function. The system in this case is world order and western governments
- joquarky 7mo ago
  "Move fast and break things" broke containment to the tech industry. Now you can see it everywhere.

stavros 7mo ago

It's not alleged: https://www.ghacks.net/2026/03/27/anthropic-reduces-claude-session-limits-during-peak-hours-for-free-pro-and-max-users/ https://www.ghacks.net/2026/03/27/anthropic-reduces-claude-s...

minimaxir 7mo ago
That's a separate change to what the linked HN post described.