DEV Community

Sam Sun profile picture

Sam Sun

Full-stack developer passionate about open source and clean code.

Location New York, NY Joined Joined on 
Unfingerprinted Generation Spans Cannot Be Diffed
codepro_3283 profile

Unfingerprinted Generation Spans Cannot Be Diffed

#ai #observability #debugging #python
Comments
6 min read
Join Tool Exits to Spans Before Trusting the Summary
codepro_3283 profile

Join Tool Exits to Spans Before Trusting the Summary

#ai #observability #debugging #python
Comments
6 min read
Wall-Clock Minus Exclusive Spans Is the Untraced Remainder
codepro_3283 profile

Wall-Clock Minus Exclusive Spans Is the Untraced Remainder

#ai #observability #debugging #python
Comments
6 min read
Map Every Working-Tree Delta to a Closed Tool Span
codepro_3283 profile

Map Every Working-Tree Delta to a Closed Tool Span

#ai #debugging #observability #python
Comments 1
7 min read
Off-Trace Token Counts Cannot Gate a Run
codepro_3283 profile

Off-Trace Token Counts Cannot Gate a Run

#ai #agents #observability #python
Comments
7 min read
A Trace You Cannot Replay Is Only a Log
codepro_3283 profile

A Trace You Cannot Replay Is Only a Log

#ai #agents #observability #testing
Comments
6 min read
Call the Agent Fixed Only After a Bounded Trace Diff
codepro_3283 profile

Call the Agent Fixed Only After a Bounded Trace Diff

#agents #observability #debugging #python
Comments 1
8 min read
Fail the Run If the Trace Forest Splits
codepro_3283 profile

Fail the Run If the Trace Forest Splits

#ai #agents #observability #python
Comments
7 min read
Stdout Cannot Halt an Agent Loop
codepro_3283 profile

Stdout Cannot Halt an Agent Loop

#observability #agents #tracing #python
Comments
7 min read
Treat Span Cardinality as a Hard Error
codepro_3283 profile

Treat Span Cardinality as a Hard Error

#agents #observability #python #debugging
Comments
6 min read
Causal Order Beats Timestamps in Agent Traces
codepro_3283 profile

Causal Order Beats Timestamps in Agent Traces

#ai #agents #debugging #python
Comments
8 min read
Gate Agent Success on Closed Spans
codepro_3283 profile

Gate Agent Success on Closed Spans

#agents #observability #debugging #python
Comments
6 min read
Fail the Tool Call If the Trace Cannot Cite a Source
codepro_3283 profile

Fail the Tool Call If the Trace Cannot Cite a Source

#ai #agents #debugging #observability
Comments
8 min read
Diff the Run Card, Not the Log Stream
codepro_3283 profile

Diff the Run Card, Not the Log Stream

#ai #debugging #python #testing
Comments
8 min read
The Data Boundary Problem: Using a Free Server Without Leaking Your Prompts
codepro_3283 profile

The Data Boundary Problem: Using a Free Server Without Leaking Your Prompts

#ai #security #python #webdev
Comments
4 min read
The 429 Cascade: Debugging a Code-Review Bot on a Free Endpoint
codepro_3283 profile

The 429 Cascade: Debugging a Code-Review Bot on a Free Endpoint

#ai #python #devops #llm
Comments
5 min read
MiniMax H3: Run a 30-Minute Adoption Gate Before You Commit
codepro_3283 profile

MiniMax H3: Run a 30-Minute Adoption Gate Before You Commit

#ai #llm #python #benchmarking
Comments
2 min read
Stdout Order Is Not Causality: A Parent-Span Contract for Agents
codepro_3283 profile

Stdout Order Is Not Causality: A Parent-Span Contract for Agents

#ai #debugging #python #agents
Comments
8 min read
Quota Is a Test Signal: Budgeting Free Model Calls Like CI Minutes
codepro_3283 profile

Quota Is a Test Signal: Budgeting Free Model Calls Like CI Minutes

#ai #testing #python #agents
Comments
5 min read
Free Tokens Won't Fix Black-Box Agent Runs: Build a Trace Loop
codepro_3283 profile

Free Tokens Won't Fix Black-Box Agent Runs: Build a Trace Loop

#ai #observability #llm #opensource
Comments
4 min read
Agent Summaries Are Claims, Not Logs: A Trace Audit Loop
codepro_3283 profile

Agent Summaries Are Claims, Not Logs: A Trace Audit Loop

#ai #observability #llm #testing
Comments
4 min read
A Free-Tier Trace Loop for Agent Run Regression
codepro_3283 profile

A Free-Tier Trace Loop for Agent Run Regression

#ai #observability #debugging #opensource
Comments 1
4 min read
Diff the Traces, Not the Code: A Bisection Loop for Agent Runs
codepro_3283 profile

Diff the Traces, Not the Code: A Bisection Loop for Agent Runs

#ai #python #opensource #tutorial
Comments
4 min read
Turn Agent Tool Calls Into Run Receipts
codepro_3283 profile

Turn Agent Tool Calls Into Run Receipts

#ai #agents #debugging #llm
1
Comments
4 min read
Agent Runs Are Black Boxes: A Trace-First Debug Loop
codepro_3283 profile

Agent Runs Are Black Boxes: A Trace-First Debug Loop

#ai #agents #observability #python
Comments 2
4 min read
Free Model Access and Free Hosting: A Decision Guide for Agent Workloads
codepro_3283 profile

Free Model Access and Free Hosting: A Decision Guide for Agent Workloads

#ai #agents #llm #architecture
Comments
3 min read
Free Endpoints Are a Contract, Not a Gift: A Fit Test for Agent Workloads
codepro_3283 profile

Free Endpoints Are a Contract, Not a Gift: A Fit Test for Agent Workloads

#ai #programming #llm #opensource
Comments 1
5 min read
Your Retry Loop Is a Token Incinerator: A Cascade Router for Mixed-Tier Endpoints
codepro_3283 profile

Your Retry Loop Is a Token Incinerator: A Cascade Router for Mixed-Tier Endpoints

#agents #python #architecture #ai
Comments
5 min read
Free vs Self-Hosted Models: A Break-Even Framework for Agent Workloads
codepro_3283 profile

Free vs Self-Hosted Models: A Break-Even Framework for Agent Workloads

#ai #llm #agents #opensource
Comments
5 min read
A Cost-Per-Token Benchmark Is Not Enough: A Reproducible Harness for Free and Cheap Model Endpoints
codepro_3283 profile

A Cost-Per-Token Benchmark Is Not Enough: A Reproducible Harness for Free and Cheap Model Endpoints

#ai #programming #tutorial #api
Comments
4 min read
A Preflight Harness for Tool-Using Coding Agents
codepro_3283 profile

A Preflight Harness for Tool-Using Coding Agents

#ai #security #programming #productivity
Comments 1
5 min read
loading...