Skip to Content
logologo
AI Incident Database
Open TwitterOpen RSS FeedOpen FacebookOpen LinkedInOpen GitHub
Open Menu
発見する
投稿する
  • ようこそAIIDへ
  • インシデントを発見
  • 空間ビュー
  • テーブル表示
  • リスト表示
  • 組織
  • 分類法
  • インシデントレポートを投稿
  • 投稿ランキング
  • ブログ
  • AIニュースダイジェスト
  • リスクチェックリスト
  • おまかせ表示
  • サインアップ
閉じる
発見する
投稿する
  • ようこそAIIDへ
  • インシデントを発見
  • 空間ビュー
  • テーブル表示
  • リスト表示
  • 組織
  • 分類法
  • インシデントレポートを投稿
  • 投稿ランキング
  • ブログ
  • AIニュースダイジェスト
  • リスクチェックリスト
  • おまかせ表示
  • サインアップ
閉じる

レポート 428

関連インシデント

インシデント 3028 Report
Poor Performance of Tesla Factory Robots

Loading...
Elon Musk says a flufferbot caused the Model 3 delays.
slate.com · 2018

Flufferbot! Alberto E. Rodriguez/Getty Images

Falling behind production deadlines for its Model 3 sedan and rapidly burning cash, Tesla has been struggling mightily to speed up its assembly line. On a call with investors Wednesday, CEO Elon Musk said the company had identified an unlikely culprit in those delays: a “flufferbot.”

As Tesla scrambles to deliver on some 450,000 orders for the Model 3, its first mass-market vehicle, Musk has described the company’s predicament as “production hell.” Part of the problem has been an over-reliance on automation, which Musk recently acknowledged, adding in an April 13 tweet: “Humans are underrated.”

Musk gave the public an amusing peek into that hell on the company’s quarterly earnings call Wednesday. “We did go too far on the automation front and automated some pretty silly things,” he said. Musk proceeded to offer the following example, about a fiberglass mat that was designed to insulate noise from the battery pack:

We had these fiberglass mats on the top of the battery pack. They’re basically fluff. So we tried to automate the placement and bonding of fluff to the top of the battery pack. Which is ridiculous. So we had this weird flufferbot. Which was really an incredibly difficult machine to make work. Machines are not good at picking up pieces of fluff. Human hands are way better at doing that. So we had a super-complicated machine. Using a vision system to try to put a piece of fluff on a battery pack. … … The line kept breaking down because Flufferbot would frequently just fail to pick up the fluff. Or put it in a random location.

Musk said he asked his team whether the fluff was really necessary. The company tested a car both with and without the fiberglass battery insulation and found “no change in the noise in the cabin.” They concluded that the part was unnecessary and did away with the flufferbot.

Elon on "silliest" examples of over-automation at Tesla: "Fluffer bot would frequently fail to pick up the fluff." Rip Flufferbot — Caroline O'Donovan (@ceodonovan) May 2, 2018

Musk also gave an example of what he called “overgeneralizing the design” of the Model 3. Tesla has plans for a dual-motor, all-wheel-drive version of the vehicle. But right now, it’s only producing the single-motor, rear-wheel-drive version. Nevertheless, the company was building battery packs with both front and rear ports, and adding a sealed plate to the front port on each one. “It added cost, added manufacturing staff, and added a failure mode,” Musk said. “For something that is unnecessary.” He implied that Tesla has stopped including the front-drive port on the batteries for rear-drive vehicles.

Partly as a result of such changes, Musk said, “We’ve had a radical improvement in battery pack production” in just the past few weeks. Whereas assembling a pack used to take as long as seven hours, Tesla can now assemble them in “under 17 minutes,” he said.

While it looks like Tesla will come nowhere close to Musk’s initial goal of building 500,000 cars this year, the company said Wednesday that it still hopes to meet a significantly revised goal of building 5,000 Model 3s per week within the next two months. However, 2,270 vehicles is the most it has built in a week so far, the company reported. You can read its full quarterly earnings report here.

Tesla’s stock had plunged 4.8 percent in after-hours trading by the end of Wednesday’s earnings call.

情報源を読む

リサーチ

  • “AIインシデント”の定義
  • “AIインシデントレスポンス”の定義
  • データベースのロードマップ
  • 関連研究
  • 全データベースのダウンロード

プロジェクトとコミュニティ

  • AIIDについて
  • コンタクトとフォロー
  • アプリと要約
  • エディタのためのガイド

インシデント

  • 全インシデントの一覧
  • フラグの立ったインシデント
  • 登録待ち一覧
  • クラスごとの表示
  • 分類法

2024 - AI Incident Database

  • 利用規約
  • プライバシーポリシー
  • Open twitterOpen githubOpen rssOpen facebookOpen linkedin
  • e1b50cd