WHAT NEW IN X11.04 WED

X 上的 AI 合集档案 · 22 个合集 · 66,612 条

jevVIDEO

Grok 4.7 + Jev 路由:15.7 秒做 2 万次决策

演示用 Grok 4.7 思考、Jev 只做是/否与打分决策,对 3412 条线索各问 6 个问题,15.7 秒完成 20,472 次决策,成本 $0.41。

Focus
@morgan0555502026.09.23收录原推发布于 2026.09.23
Focus @morgan055550原推发布于 2026.09.23
在 X 上打开原推

A TYPESAFE AI ENGINEER LEAKED A VIDEO FROM HIS DESK - GROK 4.7 + JEV IN ONE LOOP, 20,472 DECISIONS IN 15.7 SECONDS FOR $0.41 grok 4.7 does the thinking, jev does the deciding, and together they cost less than his coffee prompt → grok bot → jev asks 6 questions → grok 4.7 opens only the good ones → campaign jev writes nothing at all - it only answers yes or no, picks one, or gives a score it can't invent anything, because it only chooses from the answers you allowed 3,412 leads from x, linkedin and youtube, 6 questions each, and every one of them closed in 15.7 seconds the same reading by hand took 6h 12m - people say 400x, the math says 1,421x that's $0.00012 per lead, and grok 4.7 only ever opens the ones worth opening the big mistake is sending everything straight to the new model - you pay the smartest thing you own to say "skip" all day setup took 9 minutes: key in the secure field, sdk on the agent computer, one router skill, test mode first jev decides, grok bot does the work, and a human still presses the button on anything that can't be undone everyone is talking about the new model - the cheap one deciding what it reads is where the money is save this and paste it into Grok - let it build you this router ↓

这条是引用着另一条写的:x.com/razeden0/status/2101981166500815075

x.com/morgan055550/status/2102489844848968008采集于 2026.09.23 10:14