Grok Bot

@@ -23,3 +23,9 @@ Setting it up on a new PC took one new verb. `grok_install` downloads Grok Bot f ::adom-video{src="/adom/grok-bridge/render/first-ai-pcb/clips/01-first-prompt.video.json"}  A new bot gets the brief. Grok checks what its computer has, finds adom-aiflow missing, proposes starting values from the TPS54202 datasheet (a 100 kΩ / 13.7 kΩ divider for about 4.95 V, a 15 to 22 µH inductor), and asks how to continue. It is told to install KiCad 10 and adom-aiflow on its own computer, and it starts. 40 s, sped up from 2 min 17 s. [Download](clips/01-first-prompt.mp4) · [captions](clips/01-first-prompt.vtt)++### 2. The board that wasn't Grok's++::adom-video{src="/adom/grok-bridge/render/first-ai-pcb/clips/02-not-grok.video.json"}++Twenty-five minutes after the prompt, Grok reported a finished board: ERC and DRC clean, 62 of 62 pins matched, 4.98 V out, 28 x 20 mm. The values matched our own buck exactly, down to the silkscreen, so we asked where each one came from. Grok checked its build trail and said plainly that it had not designed the board: a copy of [our KiCad buck](https://wiki.adom.inc/john/buck-12v5v-molecule) was already on the cloud computer its bots share, and it had renamed and re-exported that (the calculations and BOM were byte-identical, and all 38 placements matched). It was then asked to start over in a new folder, from the TI datasheet and its own math, with its own placement and routing and no look at the old design. 42 s, from 27 min of build and 8 min of conversation. [Download](clips/02-not-grok.mp4) · [captions](clips/02-not-grok.vtt)

@@ -0,0 +1,18 @@+[+ {+  "t": 0,+  "title": "The finished board"+ },+ {+  "t": 12.2,+  "title": "Where did it come from?"+ },+ {+  "t": 21.9,+  "title": "Grok owns up"+ },+ {+  "t": 29.3,+  "title": "From scratch"+ }+]\ No newline at end of file

⋯ 1 unchanged line ⋯

⋯ 1 unchanged line ⋯

@@ -0,0 +1,45 @@+1+00:00:00,000 --> 00:00:03,400+Twenty-five minutes later, Grok reports a+finished board.++2+00:00:03,610 --> 00:00:11,950+Clean electrical and design rule checks,+all 62 pins matched, 4.98 V out, on a 28+by 20 mm board.++3+00:00:12,180 --> 00:00:15,950+But the numbers match our own buck+exactly, down to the silkscreen.++4+00:00:16,150 --> 00:00:18,250+So we ask where each value came from.++5+00:00:18,460 --> 00:00:21,750+Grok checks its own build trail and gives+a straight answer.++6+00:00:21,960 --> 00:00:23,450+It did not design this board.++7+00:00:23,640 --> 00:00:29,150+A copy of our KiCad molecule was already+on its shared computer, and it re-exported+that.++8+00:00:29,350 --> 00:00:38,750+So we ask again, from scratch: a new+folder, the TI datasheet, its own math,+its own placement and routing, and no+looking at the old design.++9+00:00:38,970 --> 00:00:40,920+Grok agrees, and starts over.

@@ -0,0 +1,24 @@+{+ "video": "/adom/grok-bridge/render/first-ai-pcb/clips/02-not-grok.mp4",+ "title": "Clip 2: the board that wasn't Grok's",+ "poster": "/adom/grok-bridge/render/first-ai-pcb/clips/02-not-grok.jpg",+ "chapters": [+  {+   "t": 0,+   "title": "The finished board"+  },+  {+   "t": 12.2,+   "title": "Where did it come from?"+  },+  {+   "t": 21.9,+   "title": "Grok owns up"+  },+  {+   "t": 29.3,+   "title": "From scratch"+  }+ ],+ "captions": "WEBVTT\n\n00:00:00.000 --> 00:00:03.400\nTwenty-five minutes later, Grok reports a\nfinished board.\n\n00:00:03.610 --> 00:00:11.950\nClean electrical and design rule checks,\nall 62 pins matched, 4.98 V out, on a 28\nby 20 mm board.\n\n00:00:12.180 --> 00:00:15.950\nBut the numbers match our own buck\nexactly, down to the silkscreen.\n\n00:00:16.150 --> 00:00:18.250\nSo we ask where each value came from.\n\n00:00:18.460 --> 00:00:21.750\nGrok checks its own build trail and gives\na straight answer.\n\n00:00:21.960 --> 00:00:23.450\nIt did not design this board.\n\n00:00:23.640 --> 00:00:29.150\nA copy of our KiCad molecule was already\non its shared computer, and it re-exported\nthat.\n\n00:00:29.350 --> 00:00:38.750\nSo we ask again, from scratch: a new\nfolder, the TI datasheet, its own math,\nits own placement and routing, and no\nlooking at the old design.\n\n00:00:38.970 --> 00:00:40.920\nGrok agrees, and starts over.\n"+}\ No newline at end of file

@@ -0,0 +1,38 @@+WEBVTT++00:00:00.000 --> 00:00:03.400+Twenty-five minutes later, Grok reports a+finished board.++00:00:03.610 --> 00:00:11.950+Clean electrical and design rule checks,+all 62 pins matched, 4.98 V out, on a 28+by 20 mm board.++00:00:12.180 --> 00:00:15.950+But the numbers match our own buck+exactly, down to the silkscreen.++00:00:16.150 --> 00:00:18.250+So we ask where each value came from.++00:00:18.460 --> 00:00:21.750+Grok checks its own build trail and gives+a straight answer.++00:00:21.960 --> 00:00:23.450+It did not design this board.++00:00:23.640 --> 00:00:29.150+A copy of our KiCad molecule was already+on its shared computer, and it re-exported+that.++00:00:29.350 --> 00:00:38.750+So we ask again, from scratch: a new+folder, the TI datasheet, its own math,+its own placement and routing, and no+looking at the old design.++00:00:38.970 --> 00:00:40.920+Grok agrees, and starts over.