I agree with what I think you meant: it is not a bug in Claude. However "javascript on the backend" is exactly what Claude Code is! It's a node.js implementation of automation components that feed and pull from the Claude model.
I have a VERY hard time believing that they only use JSON serialization between the model and the TUI. If they're seriealizing JSON between agents, tools, or other components, then this problem is going to continue to persist for a very long time.
This symptom was affecting display of numbers in the TUI, but the real bug/design flaw is using jq or JSON as a transport mechanism between components or between tools. JSON's number range is famously poor and it even architecture-dependent. So if you use jq or JSON to connect components or tools, you will have this same problem silently occurring elsewhere without visibility in the UI.
This is a bigger deal than it seems like! A confidence-inducing fix would include a blog post describing a top-to-bottom audit of jq/JSON used as a transport layer between tools and components. Not just a patch to the most visible problem.
It made it to the front page at about #11. Then apparently the editors renamed it to the (less interesting/more convoluted) title of the page it linked to. I didn't cause or approve that change. Why does HN rewrite my post? Is it because it was negative to Claude Code?
Distributed Systems Engineers: Mid- and Senior- Levels
thatDot has created the world's first Streaming Graph, called "Quine" to make interpreting high-volume data streams dramatically faster and easier [1]. After 7 years of DARPA-funded research and development with a fantastic team of engineers, it's a game-changing core infrastructure technology. Our products are used by major security companies to do real-time behavioral analysis, monitor and triage their alarms, and automate remediation of security threats.
We're an early-stage venture backed company, and now is a great time to join! We believe strongly that each person should get more from their job than just a paycheck! It should be interesting, provide a chance to learn new things, help advance your career, work with amazing and kind people, and be fun! A person ≠ their opinions. We value individual people who know how to disagree productively with each other so we can all learn. We deliberately cultivate a space where learning new things is the goal. Non-traditional backgrounds (i.e. non-CS degrees) are strongly encouraged; philosophy majors wanted!
Tools in our stack - NOT REQUIRED if you're willing to learn:
- Scala. We happily use pragmatic functional programming on the JVM.
- Akka. The Actor model is at the heart of what we do.
- Akka Streams. Backpressure is the ideal way to approach streaming data.
Distributed Systems Engineers: Mid- and Senior- Levels
thatDot has created the world first Streaming Graph, called "Quine" to make interpreting high-volume data streams dramatically faster and easier [1]. After 7 years of DARPA-funded research and development with a fantastic team of engineers, it's a game-changing core infrastructure technology. Our products are used by major security companies to do real-time behavioral analysis, monitor and triage their alarms, and automate remediation of security threats.
We're an early-stage venture backed company, and now is a great time to join! We believe strongly that each person should get more from their job than just a paycheck! It should be interesting, provide a chance to learn new things, help advance your career, work with amazing and kind people, and be fun! A person ≠ their opinions. We value individual people who know how to disagree productively with each other so we can all learn. We deliberately cultivate a space where learning new things is the goal. Non-traditional backgrounds (i.e. non-CS degrees) are strongly encouraged; philosophy majors wanted!
Tools in our stack - NOT REQUIRED if you're willing to learn:
- Scala. We happily use pragmatic functional programming on the JVM.
- Akka. The Actor model is at the heart of what we do.
- Akka Streams. Backpressure is the ideal way to approach streaming data.
Distributed Systems Engineers: Mid- and Senior- Levels
Director of Engineering
thatDot has created the world first Streaming Graph, called "Quine" to make interpreting high-volume data streams dramatically faster and easier [1]. After 7 years of DARPA-funded research and development with a fantastic team of engineers, it's a game-changing core infrastructure technology. Our products are used by major security companies to do real-time behavioral analysis, monitor and triage their alarms, and automate remediation of security threats.
We're an early-stage venture backed company, and now is a great time to join! We believe strongly that each person should get more from their job than just a paycheck! It should be interesting, provide a chance to learn new things, help advance your career, work with amazing and kind people, and be fun! A person ≠ their opinions. We value individual people who know how to disagree productively with each other so we can all learn. We deliberately cultivate a space where learning new things is the goal. Non-traditional backgrounds (i.e. non-CS degrees) are strongly encouraged; philosophy majors wanted!
Tools in our stack - NOT REQUIRED if you're willing to learn:
- Scala. We happily use pragmatic functional programming on the JVM.
- Akka. The Actor model is at the heart of what we do.
- Akka Streams. Backpressure is the ideal way to approach streaming data.
- Databases like Cassandra, RocksDB, and more.
- …and the right tool for the job.
If you're interested in learning more, please reach out to: [email protected]
Distributed Systems Engineers: Mid- and Senior- Levels
Director of Engineering
thatDot has created the world first Streaming Graph, called "Quine" to make interpreting high-volume data streams dramatically faster and easier [1]. After 7 years of DARPA-funded research and development with a fantastic team of engineers, it's a game-changing core infrastructure technology. Our products are used by major security companies to do real-time behavioral analysis, monitor and triage their alarms, and automate remediation of security threats.
We're an early-stage venture backed company, and now is a great time to join! We believe strongly that each person should get more from their job than just a paycheck! It should be interesting, provide a chance to learn new things, help advance your career, work with amazing and kind people, and be fun! A person ≠ their opinions. We value individual people who know how to disagree productively with each other so we can all learn. We deliberately cultivate a space where learning new things is the goal. Non-traditional backgrounds (i.e. non-CS degrees) are strongly encouraged; philosophy majors wanted!
Tools in our stack - NOT REQUIRED if you're willing to learn:
- Scala. We happily use pragmatic functional programming on the JVM.
- Akka. The Actor model is at the heart of what we do.
- Akka Streams. Backpressure is the ideal way to approach streaming data.
- Databases like Cassandra, RocksDB, and more.
- …and the right tool for the job.
If you're interested in learning more, please reach out to: [email protected]
I have a VERY hard time believing that they only use JSON serialization between the model and the TUI. If they're seriealizing JSON between agents, tools, or other components, then this problem is going to continue to persist for a very long time.