bcherny

3 ms·

31.
by bcherny 5mo ago
Hey, Boris from the team here. We did both -- we did a number of UI iterations (eg. improving thinking loading states, making it more clear how many tokens are being downloaded, etc.). But we also reduced the default effort level after eval
32.
by bcherny 5mo ago
Hey, Boris from the Claude Code team here. Normally, when you have a conversation with Claude Code, if your convo has N messages, then (N-1) messages hit prompt cache -- everything but the latest message. The challenge is: when you let a se
33.
by bcherny 5mo ago
Boris from the Claude Code team here. We agree, and will be spending the next few weeks increasing our investment in polish, quality, and reliability. Please keep the feedback coming.
34.
by bcherny 5mo ago
Hey, Boris from the Claude Code team here. People were getting extra cyber warnings when using old versions of Claude Code with Opus 4.7. To fix it, just run claude update to make sure you're on the latest. Under the hood, what was hap
35.
by bcherny 6mo ago
Thanks for the report! This was fixed in v2.1.92. Please: 1. Upgrade to the latest: claude update (seems like you did this already) 2. Start a new conversations (resuming an old convo may trigger this bug again in that convo)
36.
by bcherny 6mo ago
Thank you! Looking
37.
by bcherny 6mo ago
The docs are a good place to start: https://platform.claude.com/docs/en/build-with-claude/prompt...
38.
by bcherny 6mo ago
There were two issues with some other 3p harnesses: 1. Poor cache utilization. I put up a few PRs to fix these in OpenClaw, but the problem is their users update to new versions very slowly, so the vast majority of requests continued to use
39.
by bcherny 6mo ago
This warning is not enabled for modern models. No action needed. I'm digging into the report above as soon as they're able to /feedback.
40.
by bcherny 6mo ago
We are taking it seriously, and are continuing to investigate. We are not trusting the metrics.
41.
by bcherny 6mo ago
The number of people using Claude Code has grown very quickly, which means: - More configurations and environments we need to test - Given an edge/corner case, it is more likely a significant number of users run into it - As the ecosys
42.
by bcherny 6mo ago
That's weird. Would you mind running /feedback and sharing the id here next time you see this? I'd love to debug
43.
by bcherny 6mo ago
Ack, it is currently blue but we can make it red
44.
by bcherny 6mo ago
> 1hr -> 5min on March 6th This is not accurate. The main agent typically uses a 1h cache (except for API customers, which can enable 1h but it is not on by default because it costs more). Sub-agents typically use a 5m cache.
45.
by bcherny 6mo ago
Claude Code is the most prompt cache-efficient harness, I think. The issue is more that the larger the context window, the higher the cost of a cache miss.
46.
by bcherny 6mo ago
> past 256k it looks like your usage consumes your limits faster This is false. My guess is what is happening is #1 above, where restarting a stale session causes a 256k cache miss. That said, I hear the frustration. We are actively work
47.
by bcherny 6mo ago
Yes, we're trying a couple of experiments along these lines. Good intuition.
48.
by bcherny 6mo ago
I don't think that's accurate. The malware prompt has been around since Sonnet 3.7. We carefully evaled it for each new model release and found no regression to intelligence, alongside improved scores for cyber risk. That said, we
49.
by bcherny 6mo ago
Hey all, Boris from the Claude Code team here. We've been investigating these reports, and a few of the top issues we've found are: 1. Prompt cache misses when using 1M token context window are expensive. Since Claude Code uses a
50.
by bcherny 6mo ago
All environment variables can also be configured via settings files (in the “env” field). Our approach generally is to use env vars for more experimental and low usage settings, and reserve top-level settings for knobs that we expect custom
51.
by bcherny 6mo ago
Thanks for the feedback IDs — read all 5 transcripts. On the model behavior: your sessions were sending effort=high on every request (confirmed in telemetry), so this isn't the effort default. The data points at adaptive thinking und
52.
by bcherny 6mo ago
Christopher, would you be able to share the transcripts for that repo by running /bug? That would make the reports actionable for me to dig in and debug.
53.
by bcherny 6mo ago
Fair feedback, edited!
54.
by bcherny 6mo ago
Run this: /effort high
55.
by bcherny 6mo ago
Yep, exactly
56.
by bcherny 6mo ago
Yep totally -- think of this as "maximum effort". If a task doesn't need a lot of thinking tokens, then the model will choose a lower effort level for the task.
57.
by bcherny 6mo ago
From the contents of the issue, this seems like a fairly clear default effort issue. Would love your input if there's something specific that you think is unaddressed.
58.
by bcherny 6mo ago
There is a setting if you'd like to continue to see it: showThinkingSummaries. See the docs: https://code.claude.com/docs/en/settings#available-settings
59.
by bcherny 6mo ago
When you submit a /bug we get a way to see the contents of the conversation. We don't see anything else in your codebase.
60.
by bcherny 6mo ago
A mix of evals and vibes.

More ›