-
Notifications
You must be signed in to change notification settings - Fork 2
Expand file tree
/
Copy path2026-ca-origami-retreat-aicoding.html
More file actions
614 lines (468 loc) · 24.6 KB
/
Copy path2026-ca-origami-retreat-aicoding.html
File metadata and controls
614 lines (468 loc) · 24.6 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
<!doctype html>
<html>
<head>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0, maximum-scale=1.0, user-scalable=no">
<!-- Edit me start! -->
<title>A few words of intro into AI assisted coding</title>
<meta name="description" content="CA Origami Retreat. A brief introduction to agentic coding with Claude Code, spec-driven development workflows, and the AI coding maturity ladder.">
<meta name="author" content=" Yaroslav O. Halchenko ">
<!-- Edit me end! -->
<link rel="stylesheet" href="css/custom.css">
<link rel="stylesheet" href="reveal.js/dist/reset.css">
<link rel="stylesheet" href="reveal.js/dist/reveal.css">
<link rel="stylesheet" href="reveal.js/dist/theme/beige.css">
<!-- Theme used for syntax highlighted code -->
<link rel="stylesheet" href="reveal.js/plugin/highlight/monokai.css">
</head>
<body>
<div class="reveal">
<div class="slides">
<!-- Start of slides -->
<!-- S1. Title Slide -->
<section>
<section>
<a href="http://centerforopenneuroscience.org/"><img data-src="pics/con-ccn-dartmouth-letterhead.svg" style="height:150px;"></a>
<h2 style="margin-top:0.3em;margin-bottom:0.3em;">A few words of intro into<br/>AI assisted coding</h2>
<p style="margin-top:0.3em;margin-bottom:0.3em;font-size:0.75em;font-style:italic;">
<strong>From vibe coding to spec-driven compound engineering</strong><br/>
</p>
<div style="margin-top:0.3em;text-align:center">
<table style="border: none;">
<tr>
<td>
Yaroslav O. Halchenko<br>
<small><a href="https://twitter.com/yarikoptic" target="_blank"><img data-src="pics/twitter.png" style="height:30px;margin:0px" />@yarikoptic</a></small>
<small><a href="https://fosstodon.org/@yarikoptic" target="_blank"><img data-src="pics/mastodon-icon.svg" style="height:30px;margin:0px;margin-right:0.1em" />@yarikoptic@fosstodon.org</a></small>
</td>
<td></td>
</tr>
<tr>
<td>
<small><br><a href="http://centerforopenneuroscience.org/" target="_blank">Center for Open Neuroscience</a>
<br><a href="https://pbs.dartmouth.edu/" target="_blank">Department of Psychological and Brain Sciences</a>
<br><a href="https://www.dartmouth.edu/ccn/" target="_blank">Center for Cognitive Neuroscience</a><br>
<a href="http://www.dartmouth.edu" target="_blank">Dartmouth College<br>
New Hampshire, USA</a></small>
<img style="width:160px; margin: 0px" data-src="pics/2026-ca-origami-retreat-aicoding-qrcode.png"/>
</td>
<td>
</td>
</tr>
</table>
</div>
<small>
<strong>CA Origami Retreat — 2026</strong><br/>
Live slides/<a href="https://datasets.datalad.org/centerforopenneuroscience/talks/.git">Sources</a>:
<a href="https://datasets.datalad.org/centerforopenneuroscience/talks/2026-ca-origami-retreat-aicoding.html#/">https://datasets.datalad.org/centerforopenneuroscience/talks/2026-ca-origami-retreat-aicoding.html</a><br/>
<br>
<small>
<a href="http://datalad.org" target="_blank"> <img style="height:100px;margin:15px" data-src="pics/datalad_D.svg"/></a>
<a href="http://neuro.debian.net" target="_blank"> <img style="height:100px;margin:15px" data-src="pics/neurodebian.png"/></a>
<a href="http://repronim.org" target="_blank"> <img style="height:100px;margin:15px" data-src="pics/repronim-logo-vertical.svg"/></a>
<a href="https://open-brain-consent.readthedocs.io" target="_blank"> <img style="height:100px;margin:15px" data-src="pics/OBC_LogoCheck.svg"/></a>
<a href="https://dandiarchive.org" target="_blank"> <img style="height:100px;margin:15px" data-src="pics/dandi-logo-square.svg"/></a>
<a href="https://bids.neuroimaging.io/" target="_blank"> <img style="height:100px;margin:15px" data-src="pics/BIDS_Logo.png"/></a>
<a href="https://standforukraine.com/" target="_blank"> <img style="height:100px;margin:15px" data-src="pics/Ukrainian_Blue-Yellow_ribbon.svg"/></a>
</small>
</small>
</section>
</section>
<!-- S2. Reality Check / Disclaimer -->
<section>
<h2 style="margin-top:0.5em;margin-bottom:0.5em;">Reality Check</h2>
<div style="text-align:center; margin-top:0.5em;">
<img src="pics/borrowed/idiocracy-fixed.gif" style="width:50%; box-shadow: 10px 10px 8px #888888;"/>
</div>
<div style="margin-top:0.8em; padding:0.8em; background-color:rgba(255,255,255,0.9); border-radius:10px;">
<p style="font-size:0.85em; margin-bottom:0.3em;">
<strong>Disclaimer:</strong> This talk will <em>not</em> make you at once into AI coder.
</p>
<p style="font-size:0.8em; font-style:italic; color:#555; margin-top:0.3em;">
But it <em>will</em> point to a few approaches you can start using TODAY!<br/>
And you MUST NOT forget about social aspect of the retreat!
</p>
</div>
</section>
<!-- Book recommendation -->
<section>
<h3><a href="https://scifi.fandom.com/wiki/Avogadro_Corp" target="_blank">Avogadro Corp</a></h3>
<h4>The Singularity is Closer than it Appears</h4>
<div style="text-align:center; margin-top:0.5em;">
<a href="https://scifi.fandom.com/wiki/Avogadro_Corp" target="_blank"><img src="pics/borrowed/2011-book-avogadro-corp.png" style="max-height:700px; box-shadow: 5px 5px 8px #888;"/></a>
</div>
<p style="font-size:0.7em; color:#555; margin-top:0.5em;">William Hertling, 2011</p>
</section>
<!-- S3-S19: Main content in markdown sections -->
<section data-markdown data-separator="^\n----\n" data-vertical="^\n---\n">
<textarea data-template>
# WARNING!
[](https://hbr.org/2026/02/ai-doesnt-reduce-work-it-intensifies-it)
----
## Previously on...
### YODA Webinar: preserving ALL digital artifacts
<table style="width:100%;"><tr>
<td style="width:55%; vertical-align:top; font-size:0.75em;">
Where I left off — why we should preserve **all** digital research artifacts as an **AI-accessible knowledge base**
<a href="https://datasets.datalad.org/centerforopenneuroscience/talks/2026-repronim-YODA-BIDS-webinar.html#/" target="_blank">Slides</a> |
<a href="https://www.youtube.com/watch?v=1XbTbJ_P2x0" target="_blank">YouTube</a> |
<a href="https://datasets.datalad.org/repronim/ReproTube/web/#/channel/ReproNim/video/1XbTbJ_P2x0?tab=wide=1&t=3053&q=AI&cs=1&filter=1" target="_blank">AnnexTube (at AI section)</a>
<div style="margin-top:0.8em; padding:0.6em; background-color:#e3f2fd; border-left:4px solid #1976d2;">
<strong>Key idea:</strong> With AI assistance, raw materials (depth) become synthesized insights (surface).<br/>
But first you need to <strong>capture and organize</strong> those materials — that's where YODA comes in.
</div>
</td>
<td style="width:45%; text-align:center;">
<img src="pics/surface-depth-v2.jpg" style="max-height:400px; box-shadow: 5px 5px 8px #888;"/>
</td>
</tr></table>
----
### YODA Beyond Code & Data:<br/>Organizing ALL Digital Research Artifacts
<div style="padding: 1.5em; border-radius: 10px; margin-top: 0.5em;">
<table style="width:100%; font-size:0.75em; color:#333;">
<tr style="vertical-align:top;">
<td style="width:50%; padding-right:1em;">
<strong style="color:#2a5;">Traditional YODA scope:</strong>
<ul style="margin-top:0.3em;">
<li>Code (scripts, notebooks)</li>
<li>Data (raw, derivatives)</li>
<li>Containers (environments)</li>
<li>Papers (manuscripts)</li>
</ul>
</td>
<td style="width:50%; padding-left:1em; border-left:2px solid #666;">
<strong style="color:#e53;">Expanded YODA vision:</strong>
<ul style="margin-top:0.3em;">
<li><strong>Communication:</strong> Slack, Discord, Matrix channels</li>
<li><strong>Meetings:</strong> Zoom recordings, transcripts</li>
<li><strong>Media:</strong> YouTube videos, captions, comments</li>
<li><strong>Development:</strong> GitHub issues, PR discussions, CI logs</li>
<li><strong>Computation:</strong> Shell histories, terminal sessions</li>
<li><strong>Research:</strong> PACS imaging, instrument logs</li>
</ul>
</td>
</tr>
</table>
<div style="margin-top:1em; padding:1em; background-color:#fffacd; border-left:4px solid #f90; font-size:0.8em; text-align:left; color:#333;">
<strong style="color:#f60;">Why preserve ALL this?</strong>
<ul style="margin-top:0.5em; margin-bottom:0.3em;">
<li><strong>Then:</strong> Archived for troubleshooting ("what was discussed?", "why didn't it work?")</li>
<li><strong>Now:</strong> <span style="background-color:#ffeb3b; padding:0.2em 0.4em;">AI-accessible knowledge base</span></li>
<li>With AI assistance: Raw materials (depth) → Synthesized insights (surface)</li>
<li><strong>YODA enables:</strong> Comprehensive organization, federated composition, version control at scale</li>
</ul>
</div>
<p style="font-size:0.7em; margin-top:0.8em; font-style:italic; color:#555;">
<strong>See vision:</strong> <a href="https://github.com/con/ceptualization/issues/2" target="_blank" style="color:#05a;">con/ceptualization#2</a> —
Configuration-driven archival with DataLad subdatasets for discovery & agentic access
</p>
</div>
----
### [con/serve](https://github.com/con/serve): Archiving Everything
<small>https://github.com/con/serve</small>
<div class="mermaid">
%%{init: {'theme': 'base', 'themeVariables': {'primaryColor': '#2d3748', 'primaryTextColor': '#fff', 'lineColor': '#718096', 'fontSize': '18px'}}}%%
flowchart LR
subgraph IN["INBOUND"]
direction TB
comm["Slack, Matrix, Email"]
media["YouTube, Zoom"]
code["Issues, Wikis, PRs"]
ai["AI Sessions"]
pubs["Citations, PDFs"]
cloud_in["Cloud Storage"]
comm ~~~ media ~~~ code ~~~ ai ~~~ pubs ~~~ cloud_in
end
subgraph HUB["THE VAULT"]
direction TB
ga["git-annex"]
dl["DataLad"]
org["YODA / STAMPED"]
surfaces["Hugo, HedgeDoc, LLM agents"]
ga --- dl
dl --- org
dl --- surfaces
end
subgraph OUT["OUTBOUND"]
direction TB
archives["OpenNeuro, DANDI, OSF"]
backup["S3, Glacier, Dropbox"]
webpub["GitHub Pages"]
archives ~~~ backup ~~~ webpub
end
IN ==>|"archive"| HUB
HUB ==>|"publish"| OUT
classDef inbound fill:#2b6cb0,stroke:#2c5282,color:#fff,stroke-width:2px
classDef hub fill:#d69e2e,stroke:#b7791f,color:#1a202c,stroke-width:3px
classDef outbound fill:#2f855a,stroke:#276749,color:#fff,stroke-width:2px
class comm,media,code,ai,pubs,cloud_in inbound
class ga,dl,org,surfaces hub
class archives,backup,webpub outbound
style IN fill:#ebf8ff,stroke:#2b6cb0,stroke-width:2px,color:#2b6cb0
style HUB fill:#fefcbf,stroke:#d69e2e,stroke-width:3px,color:#744210
style OUT fill:#f0fff4,stroke:#2f855a,stroke-width:2px,color:#2f855a
</div>
</textarea></section>
<!-- AI Coding Maturity Ladder — grouped as vertical slides -->
<section>
<section data-markdown><textarea data-template>
## AI Coding Maturity Ladder
[<!-- .element: style="height:550px" -->](https://www.youtube.com/watch?v=6W_-YWHKwZ4)
<small>
Credit: [Jo Van Eyck "You are Not Ready: Agentic coding in 2026"](https://www.youtube.com/watch?v=6W_-YWHKwZ4)
— **Where are YOU on this ladder?**
</small>
</textarea></section>
<section data-markdown><textarea data-template>
### Level 1: Chat
<div style="font-size:0.85em;">
- ChatGPT / Claude open in a browser — copy-paste code, ask questions
- **Skills:** prompt engineering + [context engineering](https://www.anthropic.com/engineering/effective-context-engineering-for-ai-agents)
- What to feed into context window — and what **not** to
- **Mastery:** consistent, decent-quality output from your prompts
- **Time:** hours to weeks
- **Hint:** use multiple models in parallel first via e.g. [Open WebUI](https://github.com/open-webui/open-webui) ([chat.dartmouth.edu](https://chat.dartmouth.edu))
</div>
<div style="margin-top:1em; padding:0.8em; background-color:#fff3cd; border-left:4px solid #ffc107; font-size:0.8em;">
<em>"Not rocket science, can pick up basics in hours but take weeks to become efficient"</em>
</div>
</textarea></section>
<section data-markdown><textarea data-template>
### Level 2: Mid-Loop Generation
<div style="font-size:0.85em;">
- LLM meets you **in your IDE** — "autocomplete on steroids"
- Type a code comment → get 1-3 alternative implementations
- **Skills:** evaluating alternative designs, critically selecting the best option
- **Time:** relatively quick to pass through
</div>
<div style="margin-top:1em; padding:0.8em; background-color:#d1ecf1; border-left:4px solid #17a2b8; font-size:0.8em;">
<em>Key: learn to <strong>critically evaluate</strong> suggestions, not just accept them</em>
</div>
</textarea></section>
<section data-markdown><textarea data-template>
### Level 3: In-the-Loop Agentic Coding
<div style="font-size:0.85em;">
- The **"babysitting phase"** — you actively watch the agent work
- You see it struggle, doom-loop, and you intervene
- **Skills:** recognizing doom loops, installing guard rails, building prompt/skill libraries, extracting reusable skills & meta-prompts
- **Time: 2-3 months minimum** regardless of seniority
- This is where most people should spend serious time
</div>
<div style="margin-top:1em; padding:0.8em; background-color:#f8d7da; border-left:4px solid #dc3545; font-size:0.8em;">
<em>"You need to see them struggle and you need to get frustrated with these struggles"</em>
</div>
</textarea></section>
<section data-markdown><textarea data-template>
### Level 4: On-the-Loop Agentic Coding
<div style="font-size:0.85em;">
- You **spec the work, hand it off, go have coffee**
- Come back to high-quality artifacts, verify, move on
- Running **multiple agent sessions in parallel** (each in own [worktree](https://git-scm.com/docs/git-worktree))
- **Prerequisites:** all prior skills + reusable [skills](https://code.claude.com/docs/en/skills), [MCP servers](https://code.claude.com/docs/en/mcp), test automation, security scanners
</div>
<div style="margin-top:1em; padding:0.8em; background-color:#d4edda; border-left:4px solid #28a745; font-size:0.8em;">
<em>"Not a lot of people are here today"</em>
</div>
</textarea></section>
<section data-markdown><textarea data-template>
### Level 5: Multi-Agent Coding
<div style="font-size:0.85em;">
- The frontier — [Claude Code Teams](https://code.claude.com/docs/en/agent-teams), [Devin](https://devin.ai/)-style orchestration
- **Everyone is still figuring this out**
- *"If you have any tips, share/discuss with colleagues!"*
</div>
<div style="margin-top:1em; padding:0.8em; background-color:#f8d7da; border-left:4px solid #dc3545; font-size:0.8em;">
<strong>Warning:</strong> <em>"If you start here without the foundations, you will shoot yourself, your organization, and your friends in the foot"</em>
<br/><br/>
<strong>Do your reps. Climb deliberately.</strong>
</div>
</textarea></section>
</section>
<!-- Continue with remaining slides -->
<section data-markdown data-separator="^\n----\n" data-vertical="^\n---\n">
<textarea data-template>
## [Claude Code](https://code.claude.com/docs/en/overview) — What Is It?
<div style="font-size:0.85em;">
- **CLI-based agentic coding tool** by Anthropic ([source-available](https://github.com/anthropics/claude-code), [not yet truly open source](https://github.com/anthropics/claude-code/issues/8517))
- Works in your terminal: reads/edits files, runs commands, plans, iterates
- Key: it's an **agent**, not a chatbot — it has a loop with tools
</div>
<div style="font-size:0.8em; margin-top:0.8em;">
| Concept | What it does |
|---------|-------------|
| [`CLAUDE.md`](https://code.claude.com/docs/en/memory) | Project-level instructions / persistent memory |
| [Skills / commands](https://code.claude.com/docs/en/skills) | Reusable automation patterns (slash commands) |
| [Hooks](https://code.claude.com/docs/en/hooks) | Guard rails that run automatically (pre/post tool use) |
| [Plan mode](https://code.claude.com/docs/en/common-workflows) | Explore → design → get approval → implement |
| [MCP servers](https://code.claude.com/docs/en/mcp) | Connect to external tools (GitHub, databases, ...) |
</div>
----
## Claude Code — In Practice
<div style="font-size:0.85em;">
- **Plan mode** → Implementation → Review cycle
- Works with any language/framework, integrates with git
- `@pytest.mark.ai_generated` convention for AI-written tests
</div>
<div style="font-size:0.75em; margin-top:0.8em; padding:0.8em; background-color:rgba(255,255,255,0.95); border-radius:8px;">
```bash
$ claude "implement the retry logic from the spec"
> I'll read the specification first...
> Reading .specify/specs/retry-logic.md
> Planning implementation: 3 files to modify
> [Plan mode] Here's my approach:
> 1. Add RetryPolicy class to core/retry.py
> 2. Integrate into api_client.py
> 3. Add tests with @pytest.mark.ai_generated
> Shall I proceed?
```
</div>
----
## [con/yolo](https://github.com/con/yolo) — Sandboxed Claude Code
<div style="font-size:0.85em;">
- **Containerized (Podman)** setup for running Claude Code in isolation
- Enables "YOLO mode" (permission bypass) **safely** inside a container
- Preserves your config and working directory paths
- Seamless compatibility with native Claude Code sessions
</div>
<div style="margin-top:1em; padding:0.8em; background-color:#e3f2fd; border-left:4px solid #1976d2; font-size:0.8em;">
<strong>Why?</strong> For <strong>on-the-loop</strong> work (level 4): you need safe autonomy to let agents work unsupervised while limiting blast radius.
</div>
----
## The 5-Stage Development Loop
<div class="mermaid">
%%{init: {'theme': 'base', 'themeVariables': {'fontSize': '18px'}}}%%
flowchart TD
IDEA["Idea"] --> S1["1. Overall Design / Spec"]
S1 --> S2["2. Aspect Design / Prompt"]
S2 --> S3["3. AI Work"]
S3 --> S4["4. Review / Test"]
S4 -->|"iterate"| S2
S4 --> S5["5. Compound into Spec"]
S5 -->|"next aspect"| S2
S5 -->|"adjust / memorize"| S1
style IDEA fill:#9c27b0,color:#fff,stroke:#7b1fa2,stroke-width:2px
style S1 fill:#1565c0,color:#fff,stroke:#0d47a1,stroke-width:2px
style S2 fill:#2e7d32,color:#fff,stroke:#1b5e20,stroke-width:2px
style S3 fill:#e65100,color:#fff,stroke:#bf360c,stroke-width:2px
style S4 fill:#c62828,color:#fff,stroke:#b71c1c,stroke-width:2px
style S5 fill:#283593,color:#fff,stroke:#1a237e,stroke-width:2px
</div>
----
### Mapping the Approaches
<div style="font-size:0.85em;">
| Approach | Stages used | What's missing |
|----------|------------|----------------|
| **Vibe coding** | 2 → 3 → 4 (loop) | No initial spec, no condensing — knowledge is lost |
| **Spec-driven** | **1** → 2 → 3 → 4 → 2... (→1) | Loops back to prompt, ideally also updates spec |
| **Compound engineering** | 1 → 2 → 3 → 4 → **5** → **1**+2... | **Optimal:** each cycle makes the next one easier |
</div>
<div style="margin-top:1em; padding:0.8em; background-color:#e8eaf6; border-left:4px solid #3f51b5; font-size:0.8em;">
<strong>Key insight:</strong> Each unit of engineering work should make subsequent units easier — not harder.
<br/>Stage 5 (compound) is what turns experience into <strong>reusable specifications</strong>.
</div>
----
### Spec-Driven Tools Ecosystem
<div style="font-size:0.75em;">
| Tool | What it does |
|------|-------------|
| [spec-kit](https://github.com/github/spec-kit) | GitHub's spec-driven dev toolkit: constitution → specify → plan → tasks → implement |
| [OpenSpec](https://github.com/Fission-AI/OpenSpec) | Lightweight, tool-agnostic spec framework. Proposal / Specs / Design / Tasks |
| [Compound Engineering](https://github.com/EveryInc/compound-engineering-plugin) | Claude Code plugin: plan → work → review → compound cycle |
| [LAD](https://github.com/chrisfoulon/LAD) (ad-hoc) | Project-specific dev guidelines adapted per repo |
</div>
<div style="font-size:0.75em; margin-top:0.8em;">
- Video: ["This Claude Code Plugin Transforms Vibe Coders Into Senior Engineers"](https://www.youtube.com/watch?v=IQ1_5jPiQoE)
- Real example: [dandi-cli PR #1805](https://github.com/dandi/dandi-cli/pull/1805) — adapted LAD specs to project practices
</div>
----
### AI-Assisted Projects — Real Examples
<div style="font-size:0.85em;">
These are **real research tools**, not toy demos:
| Project | What it is |
|---------|-----------|
| [mykrok](https://github.com/mykrok/mykrok) | Personal knowledge base ([example](https://mykrok.oneukrainian.com/), requires auth) |
| [AnnexTube](https://github.com/con/annextube) | YouTube archival into git-annex/DataLad ([ReproTube](https://datasets.datalad.org/repronim/ReproTube/web)) |
| [con/serve](https://con.github.io/serve/) | The digital artifact archival system itself |
| [citations-collector](https://github.com/con/citations-collector) | DOI → citation discovery ([Zotero](https://www.zotero.org/groups/5774211/dandi/library), [TSVs](https://github.com/dandi/dandi-bib/tree/master/citations)) |
| [dandi-cli](https://github.com/dandi/dandi-cli) | DANDI Archive client with [LAD specs](https://github.com/dandi/dandi-cli/tree/master/.lad) & AI-generated tests |
</div>
<div style="margin-top:1em; padding:0.8em; background-color:#fff3cd; border-left:4px solid #ffc107; font-size:0.8em;">
All of these use <strong>spec-driven workflows</strong>, <code>CLAUDE.md</code>, and/or reusable skills to maintain quality.
</div>
----
### Reusable Skills: [con/skills](https://github.com/con/skills/)
<div style="font-size:0.85em;">
Claude Code skills — **write once, use across all projects:**
| Skill | What it automates |
|-------|------------------|
| [github-project-status](https://github.com/con/skills/tree/master/github-project-status) | Repository health reports |
| [introduce-codespell](https://github.com/con/skills/tree/master/introduce-codespell) | Spell-checking setup |
| [introduce-git-bug](https://github.com/con/skills/tree/master/introduce-git-bug) | Distributed bug tracking setup |
| [issue-triage](https://github.com/con/skills/tree/master/issue-triage) | Intelligent issue analysis |
| [pr-review-update](https://github.com/con/skills/tree/master/pr-review-update) | PR review management |
| [scan-projects](https://github.com/con/skills/tree/master/scan-projects) | Multi-project scanning & reporting |
| [tinuous-analyzer](https://github.com/con/skills/tree/master/tinuous-analyzer) | CI failure investigation |
</div>
<div style="margin-top:0.8em; padding:0.8em; background-color:#e8eaf6; border-left:4px solid #3f51b5; font-size:0.8em;">
This <strong>IS</strong> stage 5 (compound/condense) in action — the more you condense your workflow into skills, the more efficient every future project becomes.
</div>
----
# Thank you!
## Don't forget about YODA and
### Let the AI agents be with you
<a href="https://github.com/myyoda/poster/blob/master/ohbm2018.pdf" target="_blank"> <img style="height:450px;margin:20px" data-src="pics/yoda.svg"/></a>
<!-- .element: style="font-size:0.6em" -->
</textarea></section>
</section>
<!-- background slide , Needs tune up of config below
<section data-background="pics/con-ccn-dartmouth-letterhead.svg" data-background-opacity="100%" data-background-size="95%" data-background-position="top">
</section>
-->
<!-- End of slides -->
</div>
</div>
<script src="reveal.js/dist/reveal.js"></script>
<script src="reveal.js/plugin/notes/notes.js"></script>
<script src="reveal.js/plugin/markdown/markdown.js"></script>
<script src="reveal.js/plugin/highlight/highlight.js"></script>
<script src="reveal.js/plugin/search/search.js"></script>
<script src="reveal.js-mermaid-plugin/plugin/mermaid/mermaid.js"></script>
<script>
// More info about initialization & config:
// - https://revealjs.com/initialization/
// - https://revealjs.com/config/
Reveal.initialize({
hash: true,
// The "normal" size of the presentation, aspect ratio will be preserved
// when the presentation is scaled to fit different resolutions. Can be
// specified using percentage units.
width: 1400,
height: 1050,
//width: 1280,
//height: 960,
//width: 1024,
//height: 768,
// Factor of the display size that should remain empty around the content
margin: 0.1,
// Bounds for smallest/largest possible scale to apply to content
minScale: 0.2,
maxScale: 1.0,
progress: true,
history: true,
center: true,
controls: true,
slideNumber: 'c',
// uncomment for the slide with background banner
// controls: false, // true
// slideNumber: false, // 'c'
pdfSeparateFragments: false,
pdfMaxPagesPerSlide: 1,
pdfPageHeightOffset: -1,
transition: ['slide', 'fade'], // none/fade/slide/convex/concave/zoom
// Learn about plugins: https://revealjs.com/plugins/
// RevealSearch: Press CTRL+Shift+F to search slide content
plugins: [ RevealMarkdown, RevealHighlight, RevealNotes, RevealSearch, RevealMermaid ]
});
</script>
</body>
</html>