A 12-Day Codex Run to Build an Excel Clone

Matthew Berman describes giving Codex a brief goal and manually stopping it after more than 12 days, then demonstrates the resulting spreadsheet.

MBMatthew Berman연사 · Forward Future
채널
Matthew Berman
게시일
2026년 6월 24일
영상
0:52
읽기
1분
슬라이드
3
원문
영어
YouTube에서 열기
스크롤하면 영상이 오른쪽 아래로 작아져 계속 재생됩니다. 자막은 읽기 언어를 따릅니다.
읽기 언어2개 언어로 읽을 수 있음
이 글은 아직 한국어 번역이 없어 영어 원문으로 표시합니다.
요약

30초 요약

Matthew Berman says he asked Codex to clone Excel with full feature parity and let it run for more than 12 days before stopping it. He demonstrates a spreadsheet interface, a simple formula and sorting; the short video does not establish complete Excel compatibility.

  1. 1A long-running goalBerman says the Codex task ran for more than 12 days before he stopped it manually.
  2. 2Inspecting ExcelHe says Codex opened the desktop application and examined its features while building a clone.
  3. 3Spreadsheet demonstrationBerman demonstrates cell editing and a simple formula, and says columns can be adjusted and highlighted.
  4. 4SortingHe selects ascending order to demonstrate sorting, then asks how long the task might otherwise have continued.

핵심 정리

  1. Berman says a brief Excel-cloning goal ran for more than 12 days before he manually stopped it.
  2. He says Codex inspected features in the Excel desktop app while building its version.
  3. The demonstration shows a simple arithmetic formula and an ascending sort, but does not establish full feature parity.
전문

챕터별로 다듬은 전문 읽기

원본 자막만 보기
01

The prompt and the run

Berman gives Codex a goal of cloning Excel with full feature parity. He says the run continued for more than 12 days before he stopped it.

The Excel-cloning goal
슬라이드 1 / 3The Excel-cloning goalThe on-screen prompt asks Codex to clone Excel with full feature parity.

I gave Codex a short goal: `/goal Clone excel, full feature parity`. It ran for more than 12 days before I manually stopped it. I was impressed by the result.

Codex opened Excel on my computer, looked at features in the real desktop app and built its own version, I say. The resulting spreadsheet looks like Excel to me.

02

The resulting spreadsheet

He shows the interface and demonstrates a simple formula and sorting. His broad claims about completeness go beyond the examples shown.

A simple formula in the spreadsheet
슬라이드 2 / 3A simple formula in the spreadsheetThe demonstration displays the result of adding one and two.

I demonstrate adjustable columns and cell selection, then enter `=1+2` and show the result, 3. I describe the formulas and formatting as working and call the result feature-complete. The brief demonstration shows only a few examples, so it does not establish that all formulas and formatting work or that the clone matches every Excel feature.

Sorting options
슬라이드 3 / 3Sorting optionsThe spreadsheet shows an ascending-sort option in its menu.

I select the sort button and choose ascending order, showing the values sorted. After stopping the run myself, I wonder how long it would have continued otherwise.

같은 채널의 다른 글

슬라이드 12장34:21EN
Matthew Berman· 2026년 10월 27일Matthew Berman Examines Anthropic’s IPO Filing, AI Growth and GovernanceMatthew Berman, Sam Altman · Forward FutureThe presenter reviews the S-1’s financial metrics, compute strategy and voting structure while questioning the tension between rapid AI development and safety.AnthropicAI Safety엔터프라이즈 AI+2
슬라이드 6장25:50EN
Matthew Berman· 2026년 10월 24일Kimi K3.1 Pro Tested: Strong 3D Demos, Uneven Results and Competitive PricingMatthew Berman · Forward FutureMatthew Berman tests the open-weight model on coding, 3D rendering and visual explanations, then compares benchmark results and prices.Open-weight modelsBenchmarksAI cost & performance+1
슬라이드 2장32:20EN
Matthew Berman· 2026년 10월 23일Who Is Local AI For? The Costs, Trade-offs and a Hybrid FutureMatthew Berman · Forward FutureMatthew Berman weighs data control and independence against hardware cost, setup complexity and inference speed.Local inferenceOpen-weight modelsAI cost & performance+1
슬라이드 6장1:03:37EN
Matthew Berman· 2026년 10월 23일Matthew Berman’s Team Tests Kimi K3.1 on Animation, Games, and ChartsMatthew Berman, Alex, Brian · Forward FutureA live team AMA turns into an informal test of Kimi K3.1, from personal-agent workflows to motion graphics, a browser game, and a benchmark chart.코딩 에이전트Generative media평가(Evals)+2
0:53EN
Matthew Berman· 2026년 10월 22일Sam Altman on AI and the Discovery of New KnowledgeSam Altman · OpenAIAltman expects progress in mathematical discovery to extend to biology, materials and energy.AI for scienceMathematicsComputational Biology+1
슬라이드0:22EN
Matthew Berman· 2026년 10월 21일A Robot Bodyguard Video Shows Dramatic Rescue ScenariosMatthew Berman · Forward FutureMatthew Berman reacts to a short robot demonstration depicting fighting and vehicle-rescue scenarios.Robotics평가(Evals)