The Replay Is Not a Copy Anymore

A candidate's session replay used to be a separate, hand-built approximation of the real interface. It now reuses the same components the candidate actually saw.

AlgoArena Team4 min

Sep 7, 2026Assessments

One Question, More Than One Agent

A hiring question can now grant a candidate more than one coding agent at once, with a tab per agent and a report that shows which one they actually reached for.

AlgoArena Team4 min

Sep 7, 2026Engineering

The Agent Is Already Warm When You Arrive

A candidate's coding environment now starts warming up in the background from the moment a session begins, so reaching a question no longer means watching it boot.

AlgoArena Team4 min

Subscribe to release notes, research posts, and original field notes.

RSS feed Latest changelog

Sep 6, 2026 AssessmentsThe Report That Shows Its Work

AlgoArena Team4 min Sep 5, 2026 ClassroomGrade the Drawing While the Question Is Still Open

AlgoArena Team4 min Sep 5, 2026 ClassroomQuizzes Grounded in Your Own Material

AlgoArena Team4 min Aug 5, 2026 Interview PrepHow We Match Problems to Companies

AlgoArena Team5 min Jul 28, 2026 VibecodingVibecode Something, Keep It, Let People Play It

AlgoArena Team4 min Jul 25, 2026 AssessmentsA Replay Should Show What the Candidate Actually Saw

Eli Young4 min Jul 22, 2026 AssessmentsReuse Beats Regenerate: Teaching Recruiting Copilot to Check the Shelf First

Eli Young4 min Jul 20, 2026 EngineeringPolish Is a Routing Problem

Andy Luu4 min Jul 17, 2026 VibecodingAn Agent Editor Built to Be Read Twice

Andy Luu8 min Jul 17, 2026 ClassroomAn Educator Assistant That Shows Its Work

AlgoArena Team7 min Jul 16, 2026 AssessmentsHow to Assess Developers Who Code With AI

Eli Young15 min Jul 12, 2026 ResearchRepairing a Problem Bank at Catalog Scale

AlgoArena Team6 min

Field stories

Research and evidence

Repairing a Problem Bank at Catalog Scale

How we judged every stored solution against every stored test across the full catalog, arbitrated whether the tests or the solutions were wrong, regenerated expected outputs by consensus, and shipped checkers for problems with many valid answers.

Read

Auditing the Arena 90 Campaign Path

How the 90-node campaign was pinned to existing problems, test-count coverage, topic balance, and a stable curriculum contract.

Read

The Five-Competency AI-Native Scoring Model

How assessment scoring moved from a raw score toward public competencies for problem solving, planning, AI direction, verification, and workflow autonomy.

Read

Ghost Replay Eligibility Without Publishing Raw Labels

The shipped gates that decide whether a completed run can become a self ghost, join the human ghost pool, or enter the solution bank.

Read

View all research

Changelog

AssessmentsJuly 8, 2026

Build skill people can inspect.

Start free Meet with the team

The Replay Is Not a Copy Anymore

One Question, More Than One Agent

The Agent Is Already Warm When You Arrive

Field stories

The Report That Shows Its Work

Grade the Drawing While the Question Is Still Open

Quizzes Grounded in Your Own Material

Vibecode Something, Keep It, Let People Play It

Research and evidence

Repairing a Problem Bank at Catalog Scale

Auditing the Arena 90 Campaign Path

The Five-Competency AI-Native Scoring Model

Ghost Replay Eligibility Without Publishing Raw Labels

Changelog

Recruiter fit controls and Discord identity

Course catalog, quizzes, and ranked stability

Shared agent workspace and attribution

Verification proofs and lifecycle emails

Build skill people can inspect.

The Replay Is Not a Copy Anymore

One Question, More Than One Agent

The Agent Is Already Warm When You Arrive

Field stories

The Report That Shows Its Work

Grade the Drawing While the Question Is Still Open

Quizzes Grounded in Your Own Material

Vibecode Something, Keep It, Let People Play It

Research and evidence

Repairing a Problem Bank at Catalog Scale

Auditing the Arena 90 Campaign Path

The Five-Competency AI-Native Scoring Model

Ghost Replay Eligibility Without Publishing Raw Labels

Changelog

Recruiter fit controls and Discord identity

Course catalog, quizzes, and ranked stability

Shared agent workspace and attribution

Verification proofs and lifecycle emails

Build skill people can inspect.

AlgoArena Blog

The Replay Is Not a Copy Anymore

One Question, More Than One Agent

The Agent Is Already Warm When You Arrive

Field stories

The Report That Shows Its Work

Grade the Drawing While the Question Is Still Open

Quizzes Grounded in Your Own Material

Vibecode Something, Keep It, Let People Play It

Research and evidence

Repairing a Problem Bank at Catalog Scale

Auditing the Arena 90 Campaign Path

The Five-Competency AI-Native Scoring Model

Ghost Replay Eligibility Without Publishing Raw Labels

Changelog

Recruiter fit controls and Discord identity

Course catalog, quizzes, and ranked stability

Shared agent workspace and attribution

Verification proofs and lifecycle emails

Build skill people can inspect.

AlgoArena Blog

The Replay Is Not a Copy Anymore

One Question, More Than One Agent

The Agent Is Already Warm When You Arrive

Field stories

The Report That Shows Its Work

Grade the Drawing While the Question Is Still Open

Quizzes Grounded in Your Own Material

Vibecode Something, Keep It, Let People Play It

Research and evidence

Repairing a Problem Bank at Catalog Scale

Auditing the Arena 90 Campaign Path

The Five-Competency AI-Native Scoring Model

Ghost Replay Eligibility Without Publishing Raw Labels

Changelog

Recruiter fit controls and Discord identity

Course catalog, quizzes, and ranked stability

Shared agent workspace and attribution

Verification proofs and lifecycle emails

Build skill people can inspect.