summaryrefslogtreecommitdiffstats
path: root/2025/captions/emacsconf-2025-llm--emacs-editors-and-llm-driven-workflows--andrew-hyatt--main.vtt
blob: ea969e1e24141d0f972bdc3e6a656cfce1889271 (plain) (blame)
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
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
WEBVTT captioned by amitav

NOTE Introduction

00:00:01.040 --> 00:00:03.079
Hi, I'm Andrew Hyatt.

00:00:03.080 --> 00:00:09.399
I'm going to talk to you today about Emacs and AI,

00:00:09.400 --> 00:00:10.879
and where things are right now

00:00:10.880 --> 00:00:12.119
in the world of Emacs and AI,

00:00:12.120 --> 00:00:14.159
via large language models,

00:00:14.160 --> 00:00:16.999
and where things might be going,

00:00:17.000 --> 00:00:22.699
and what it means for the future of Emacs.

00:00:22.700 --> 00:00:27.279
I think what we're seeing with Emacs is interesting.

00:00:27.280 --> 00:00:29.399
We've seen a lot of different things

00:00:29.400 --> 00:00:31.559
come around in the past year,

00:00:31.560 --> 00:00:33.119
in the past several years.

00:00:33.120 --> 00:00:35.079
There's lots of different solutions.

00:00:35.080 --> 00:00:36.759
But in the past year, things have been very interesting.

00:00:36.760 --> 00:00:39.679
I think there's new and interesting questions

00:00:39.680 --> 00:00:43.279
about what does it mean to use Emacs?

00:00:43.280 --> 00:00:45.479
What does it mean to use any editor?

00:00:45.480 --> 00:00:47.279
I'm going to be talking about Emacs,

00:00:47.280 --> 00:00:50.359
and I'm going to show you various Emacs packages

00:00:50.360 --> 00:00:53.079
as demonstrations of these ideas.

00:00:53.080 --> 00:00:59.839
But there's the general question of

00:00:59.840 --> 00:01:03.719
what does it mean to use any editor, not just Emacs?

00:01:03.720 --> 00:01:06.239
What does it mean to do work?

00:01:06.240 --> 00:01:10.719
And I think the industry in general is facing these challenges

00:01:10.720 --> 00:01:13.279
of we don't really know where things are going to end up,

00:01:13.280 --> 00:01:16.919
but we do know the direction they're going.

00:01:16.920 --> 00:01:20.039
Emacs is a reflection of that.

00:01:20.040 --> 00:01:23.239
I think the answer for Emacs might be

00:01:23.240 --> 00:01:25.719
a little bit different than everything else,

00:01:25.720 --> 00:01:28.599
but I do want to show you what's out there

00:01:28.600 --> 00:01:33.319
so we can explore what are the possibilities

00:01:33.320 --> 00:01:41.119
of Emacs, AI, and generally how we get things done.

00:01:41.120 --> 00:01:44.719
Thanks. Let's dive right into it.

NOTE Copilot

00:01:44.720 --> 00:01:48.079
We're going to start by showing you

00:01:48.080 --> 00:01:51.039
some things that are pretty well integrated,

00:01:51.040 --> 00:01:55.279
that look a lot like what you see in Emacs

00:01:55.280 --> 00:01:58.679
and fit in with the kinds of editing

00:01:58.680 --> 00:02:02.639
that you normally do in Emacs.

00:02:02.640 --> 00:02:06.579
So this is just kind of like, it's well integrated.

00:02:06.580 --> 00:02:08.779
So we're going to talk about Copilot and Semext.

00:02:08.780 --> 00:02:12.679
Copilot is by Microsoft via GitHub,

00:02:12.680 --> 00:02:14.759
and Semext is just my personal demo,

00:02:14.760 --> 00:02:18.039
but they're both showing you, you know,

00:02:18.040 --> 00:02:24.399
this kind of thing. Let's start with Copilot.

00:02:24.400 --> 00:02:31.919
Let's try out Copilot on just a standard bit of Elisp.

00:02:31.920 --> 00:02:38.439
We're going to write a Fibonacci function.

00:02:38.440 --> 00:02:43.079
Let's try out Emacs on a standard bit of Elisp.

00:02:43.080 --> 00:02:49.279
We're going to write a Fibonacci function.

00:02:49.280 --> 00:02:53.159
And you can see like as soon as we even start typing it,

00:02:53.160 --> 00:02:56.339
we get everything as a completion.

00:02:56.340 --> 00:02:59.879
So you can just press Tab here,

00:02:59.880 --> 00:03:02.159
and you've just completed

00:03:02.160 --> 00:03:06.799
a significant bunch of Emacs Lisp code.

00:03:06.800 --> 00:03:09.919
It will do this no matter where you are.

00:03:09.920 --> 00:03:14.799
So, pretty useful. It will just keep suggesting things.

00:03:14.800 --> 00:03:16.439
Do you want to do this?

00:03:16.440 --> 00:03:17.479
I'm not sure.

00:03:17.480 --> 00:03:22.839
But it usually is offering pretty reasonable things.

00:03:22.840 --> 00:03:29.299
So you could do this with code,

00:03:29.300 --> 00:03:32.119
of course, any code.

00:03:32.120 --> 00:03:33.919
You don't really even have to have a mode for it, right?

00:03:33.920 --> 00:03:36.679
That's kind of the beauty of AI is that

00:03:36.680 --> 00:03:38.519
you don't need any Emacs functionality for this,

00:03:38.520 --> 00:03:39.519
except for Copilot.

00:03:39.520 --> 00:03:41.679
It doesn't need to know the structure of your code.

00:03:41.680 --> 00:03:45.279
It doesn't need anything except for the text itself

00:03:45.280 --> 00:03:51.239
and whatever AI integration that this is.

00:03:51.240 --> 00:03:53.739
We can look at, you can do the same thing with Org-mode.

00:03:53.740 --> 00:03:57.999
So we could say create, no,

00:03:58.000 --> 00:04:02.919
how about let's, let's do, you know, spring cleaning.

00:04:02.920 --> 00:04:10.839
It's actually the fall, but still we'll say spring cleaning.

00:04:10.840 --> 00:04:12.767
And it'll start suggesting things that, you know,

00:04:12.768 --> 00:04:15.439
maybe at first, it doesn't really know what to do to

00:04:15.440 --> 00:04:16.433
clean up all code.

00:04:16.434 --> 00:04:18.400
It thinks I need to clean up code, but no,

00:04:18.401 --> 00:04:21.839
this is going to be actual, you know,

00:04:21.840 --> 00:04:31.567
clean hood over range. Clean out pantry.

00:04:31.568 --> 00:04:33.879
These are all really reasonable suggestions.

00:04:33.880 --> 00:04:38.319
You just keep going here.

NOTE Semext

00:04:38.320 --> 00:04:40.559
I'm going to demonstrate Semext,

00:04:40.560 --> 00:04:43.879
which is a package I have on GNU Elpa,

00:04:43.880 --> 00:04:48.719
that is designed to integrate AI in a very Emacs-like way.

00:04:48.720 --> 00:04:50.999
And so what you could do is you could do a

00:04:51.000 --> 00:04:54.799
semext-search-forward.

00:04:54.800 --> 00:04:58.719
The UI looks just like other Emacs commands,

00:04:58.720 --> 00:05:02.379
but you can search for anything.

00:05:02.380 --> 00:05:06.279
There's really no way to express what I'm about to,

00:05:06.280 --> 00:05:08.679
what I'm trying to demonstrate

00:05:08.680 --> 00:05:12.359
in Emacs's normal search commands.

00:05:12.360 --> 00:05:15.399
You could really ask for anything.

00:05:15.400 --> 00:05:18.759
And it takes a little while, which is not Emacs-like,

00:05:18.760 --> 00:05:20.033
but everything else is sort of like

00:05:20.034 --> 00:05:21.719
it's designed to be like Emacs,

00:05:21.720 --> 00:05:23.519
except way more powerful.

00:05:23.520 --> 00:05:27.119
You don't need any mode to be active for this.

00:05:27.120 --> 00:05:32.039
You  just need the library

00:05:32.040 --> 00:05:34.759
and an AI provider of some sort, either locally

00:05:34.760 --> 00:05:41.199
or, you know, your favorite cloud provider.

NOTE Integrated AI experiences: gptel, ellama, chatgpt-shell, etc.

00:05:41.200 --> 00:05:43.679
Now we're going to move on to a different way

00:05:43.680 --> 00:05:46.399
of interacting with AI and Emacs.

00:05:46.400 --> 00:05:52.319
This way is less like the normal editing experience.

00:05:52.320 --> 00:05:56.999
So you lose some familiarity. However, in exchange,

00:05:57.000 --> 00:05:58.079
it is a lot more powerful.

00:05:58.080 --> 00:06:00.119
And there's a whole suite of these tools.

00:06:00.120 --> 00:06:02.479
I'm going to demonstrate gptel,

00:06:02.480 --> 00:06:05.779
which is the most popular one.

00:06:05.780 --> 00:06:06.399
But there are many.

00:06:06.400 --> 00:06:08.479
And I think different people have

00:06:08.480 --> 00:06:11.759
their own preferences of what they like to use.

00:06:11.760 --> 00:06:12.999
We're going to try now something

00:06:13.000 --> 00:06:15.079
that is a step away from just editing.

00:06:15.080 --> 00:06:19.839
And we're going to, I'm actually using gptel.

00:06:19.840 --> 00:06:22.799
There are several packages that are going to be

00:06:22.800 --> 00:06:25.959
doing the same sort of thing as I'm going to show you.

00:06:25.960 --> 00:06:29.999
gptel has sort of become the most popular one.

00:06:30.000 --> 00:06:32.199
So that's why I'm showing that to you.

00:06:32.200 --> 00:06:39.319
But let's just highlight everything and say gptel rewrite.

00:06:39.320 --> 00:06:42.399
And gptel basically just has a few things.

00:06:42.400 --> 00:06:45.119
There's different ways of thinking about this.

00:06:45.120 --> 00:06:49.999
With just a few very configurable menus,

00:06:50.000 --> 00:06:53.959
you can do a large variety of things.

00:06:53.960 --> 00:06:59.819
So let's give rewrite instructions.

00:06:59.820 --> 00:07:06.600
"Turn this into an iterative program

00:07:06.601 --> 00:07:12.199
instead of a recursive program."

00:07:12.200 --> 00:07:17.799
In Elisp, you really should not be using recursion.

00:07:17.800 --> 00:07:20.359
So we could say "return to be ready".

00:07:20.360 --> 00:07:21.119
Do we accept it?

00:07:21.120 --> 00:07:24.519
Yes, we accept it. Or we could iterate and say, no, no,

00:07:24.520 --> 00:07:26.799
that's not what we meant. We meant something else.

00:07:26.800 --> 00:07:29.159
Or you did something a little something wrong.

00:07:29.160 --> 00:07:29.879
Please fix it.

00:07:29.880 --> 00:07:31.879
So this is all very powerful.

00:07:31.880 --> 00:07:33.799
Is this editing?

00:07:33.800 --> 00:07:40.279
Well, it's in the editor.

00:07:40.280 --> 00:07:42.759
You could do this while editing, while deleting,

00:07:42.760 --> 00:07:44.959
you could be doing some sort of traditional editing.

00:07:44.960 --> 00:07:47.679
And then this, which is editing

00:07:47.680 --> 00:07:48.919
in the sense that it's in your editor,

00:07:48.920 --> 00:07:51.039
you might have to highlight

00:07:51.040 --> 00:07:52.799
some parts of the file and do things,

00:07:52.800 --> 00:07:54.719
but generally you don't even need to,

00:07:54.720 --> 00:07:59.879
or you go to a spot and you say, put code at this spot.

00:07:59.880 --> 00:08:01.959
It's kind of like editing.

00:08:01.960 --> 00:08:05.839
I would say it's not exactly editing,

00:08:05.840 --> 00:08:10.159
but it's at least something that must happen in an editor

00:08:10.160 --> 00:08:12.359
and it's well integrated into Emacs.

00:08:12.360 --> 00:08:14.759
As you can tell, it used very sort of

00:08:14.760 --> 00:08:18.239
modern standard Emacs UI paradigms

00:08:18.240 --> 00:08:20.759
and it's all written in Elisp.

00:08:20.760 --> 00:08:23.779
Everything is happening in Elisp here.

00:08:23.780 --> 00:08:25.959
So this is just very much an Emacs experience.

00:08:25.960 --> 00:08:27.679
It's just not exactly editing

00:08:27.680 --> 00:08:29.879
because the thing doing the editing

00:08:29.880 --> 00:08:32.519
is the AI and not you.

00:08:32.520 --> 00:08:36.039
You're just kind of telling it what to do.

NOTE Outside the editor

00:08:36.040 --> 00:08:41.119
Now we're going to go and look at a way of interaction

00:08:41.120 --> 00:08:43.239
that's even more powerful

00:08:43.240 --> 00:08:46.279
and even more disconnected from the normal editing experience.

00:08:46.280 --> 00:08:47.919
In fact, it's so disconnected

00:08:47.920 --> 00:08:52.399
that most people are using this without an editor.

00:08:52.400 --> 00:08:57.879
These are things like Claude Code

00:08:57.880 --> 00:09:01.079
or the sort of open source equivalent, Aider.

00:09:01.080 --> 00:09:05.039
There's a few other things that follow this pattern as well.

00:09:05.040 --> 00:09:07.479
But it's very interesting in the sense

00:09:07.480 --> 00:09:09.839
that while you can integrate these with the editors,

00:09:09.840 --> 00:09:12.039
and I'm going to show you an Emacs integration,

00:09:12.040 --> 00:09:13.519
you don't need to.

00:09:13.520 --> 00:09:16.939
And that's not the way most people are using them.

00:09:16.940 --> 00:09:19.759
And I find it very interesting that sort of

00:09:19.760 --> 00:09:23.719
we're going back kind of full circle where, you know,

00:09:23.720 --> 00:09:31.959
in the 1960s or 70s, we were using Ed from the terminal

00:09:31.960 --> 00:09:35.639
to edit files, but then we created editors,

00:09:35.640 --> 00:09:37.959
and that was a really good idea.

00:09:37.960 --> 00:09:40.167
It is a lot easier to edit files

00:09:40.168 --> 00:09:42.499
when you have an actual UI.

00:09:42.500 --> 00:09:46.879
But now it's 2025, and we're back in the terminal,

00:09:46.880 --> 00:09:50.799
and we're editing files through the terminal,

00:09:50.800 --> 00:09:53.599
and you know what, it's great,

00:09:53.600 --> 00:09:56.899
but I think it's even better with Emacs.

00:09:56.900 --> 00:10:00.279
On the other hand, it comes with some trade-offs,

00:10:00.280 --> 00:10:04.733
as you can see, as we will see.

NOTE Outside Experiences: claude-code.el, aidermacs, eca

00:10:04.734 --> 00:10:07.467
Okay, we're going to look at

00:10:07.468 --> 00:10:20.320
[audio glitch] Claude Code IDE, aidermacs, ECA.

00:10:20.321 --> 00:10:22.639
Last time, I didn't show you all the variants.

00:10:22.640 --> 00:10:26.839
I do want to show you eca, which points to,

00:10:26.840 --> 00:10:29.799
it is a very similar tool in what it does,

00:10:29.800 --> 00:10:32.739
but does have a different

00:10:32.740 --> 00:10:37.239
and I think better type of Emacs integration.

00:10:37.240 --> 00:10:42.599
All right, we're going to demonstrate Claude Code IDE,

00:10:42.600 --> 00:10:46.839
which is one of three Claude Code packages.

00:10:46.840 --> 00:10:47.719
It's a bit confusing.

00:10:47.720 --> 00:10:52.039
One of them will be demoed by another presenter

00:10:52.040 --> 00:10:54.639
at the Emacs conference, so stay tuned for that.

00:10:54.640 --> 00:10:56.439
Here I'm just going to give you a little taste

00:10:56.440 --> 00:10:58.759
of what these packages look like.

00:10:58.760 --> 00:11:03.339
So if we say Claude Code IDE,

00:11:03.340 --> 00:11:06.839
it presents us with basically

00:11:06.840 --> 00:11:09.039
almost exactly what you would get

00:11:09.040 --> 00:11:11.519
when you're running this in the terminal.

00:11:11.520 --> 00:11:13.933
And essentially there's a terminal interface.

00:11:13.934 --> 00:11:16.659
You can see that there's a vterm.

00:11:16.660 --> 00:11:20.699
But here we're going to say, "In scratch.el"...

00:11:20.700 --> 00:11:23.400
let's say what we want to happen.

00:11:23.401 --> 00:11:32.133
[In scratch.el, there is a fibonacci function.

00:11:32.134 --> 00:11:39.567
Can you add all normal elisp headers

00:11:39.568 --> 00:11:43.859
and footers to this file?]

00:11:43.860 --> 00:11:45.840
So, we just say what's going to happen,

00:11:45.841 --> 00:11:48.399
and this is going to do things in the background.

00:11:48.400 --> 00:11:50.979
It's not going to do things through Emacs.

00:11:50.980 --> 00:11:54.079
That said, there is an integration with Emacs,

00:11:54.080 --> 00:12:00.659
so that it can do things like show you these nice ediffs.

00:12:00.660 --> 00:12:03.199
My screen is not really wide enough

00:12:03.200 --> 00:12:04.699
to show you a really great ediff here,

00:12:04.700 --> 00:12:06.239
but you can kind of see what it's doing,

00:12:06.240 --> 00:12:09.079
and you can see, yeah, that looks good,

00:12:09.080 --> 00:12:14.120
so you could say yes, yes, accept the changes,

00:12:14.121 --> 00:12:25.299
and if we... Just need to revert the buffer.

00:12:25.300 --> 00:12:28.459
We can quit the printout of this.

00:12:28.460 --> 00:12:33.019
We see that it just did everything I asked it to.

00:12:33.020 --> 00:12:36.139
Is everything exactly right?

00:12:36.140 --> 00:12:39.159
Probably not. It's reasonable for a start though.

00:12:39.160 --> 00:12:40.959
But you could ask it to do anything.

00:12:40.960 --> 00:12:45.339
You could say, write unit tests for this, and it will.

00:12:45.340 --> 00:12:49.019
You could say, write me a suite of functions

00:12:49.020 --> 00:12:52.579
like Fibonacci, and it'll probably do something reasonable.

00:12:52.580 --> 00:12:54.900
But you can see this is not editing.

00:12:54.901 --> 00:12:58.659
There's nothing editing-like about this.

00:12:58.660 --> 00:13:07.159
That said, there is something that is editing.

00:13:07.160 --> 00:13:08.599
You need to give it instructions.

00:13:08.600 --> 00:13:10.959
You need to tell it what to do.

NOTE Org files

00:13:10.960 --> 00:13:19.619
And what you could do is... You could have a project.org,

00:13:19.620 --> 00:13:23.899
and what you could do is you could have functions.

00:13:23.900 --> 00:13:26.659
The way I've done things often is ....

00:13:26.660 --> 00:13:28.439
You could say something like,

00:13:28.440 --> 00:13:36.199
unit tests for Fibonacci. How do you spell Fibonacci?

00:13:36.200 --> 00:13:40.479
I don't remember. But then you could say that this is,

00:13:40.480 --> 00:13:47.159
you could clock it, basically. org-clock.

00:13:47.160 --> 00:13:48.879
What I've done is...

00:13:48.880 --> 00:13:50.399
You could add custom commands to Claude Code,

00:13:50.400 --> 00:13:53.119
and you could just say, look, here's my Org file,

00:13:53.120 --> 00:13:57.879
read it and do the thing that I'm clocked in as.

00:13:57.880 --> 00:14:01.159
And then you can write a bunch of instructions here, like,

00:14:01.160 --> 00:14:07.039
I like to use ert for tests. Tests should, like, whatever.

00:14:07.040 --> 00:14:08.639
You should just say... everything

00:14:08.640 --> 00:14:10.999
you need to kind of specify.

00:14:11.000 --> 00:14:13.199
As you get to more complicated tasks,

00:14:13.200 --> 00:14:16.679
it's harder and harder to give it all the context

00:14:16.680 --> 00:14:17.799
it needs for a task,

00:14:17.800 --> 00:14:22.299
and Org Mode is actually a pretty good way to do this.

00:14:22.300 --> 00:14:24.079
I find that this works pretty well,

00:14:24.080 --> 00:14:26.699
and you can even have it instruct Claude

00:14:26.700 --> 00:14:29.333
to just mark things done in your Org file

00:14:29.334 --> 00:14:30.679
when they're done.

00:14:30.680 --> 00:14:32.867
And it knows how to do this, of course.

00:14:32.868 --> 00:14:37.959
So, let's just clock out.

00:14:37.960 --> 00:14:45.239
That's one way to do things.

NOTE ECA

00:14:45.240 --> 00:14:49.499
So one other thing I'd like to show you is eca,

00:14:49.500 --> 00:14:52.879
which, compared to Claude Code, ECA is open source.

00:14:52.880 --> 00:14:54.239
It's very nice in that respect.

00:14:54.240 --> 00:14:57.839
It doesn't have to use Anthropic's models.

00:14:57.840 --> 00:15:00.279
You can use local models,

00:15:00.280 --> 00:15:07.619
but it has the advantage of integrating very well with Emacs.

00:15:07.620 --> 00:15:08.559
I'm not going to demonstrate it,

00:15:08.560 --> 00:15:11.159
because it works essentially the same thing you could do

00:15:11.160 --> 00:15:14.119
approximately the same kinds of things

00:15:14.120 --> 00:15:15.479
you could do with Claude Code.

00:15:15.480 --> 00:15:17.439
You just write what you want to happen

00:15:17.440 --> 00:15:18.639
and it will make it happen.

00:15:18.640 --> 00:15:21.879
It again does not do this through Emacs,

00:15:21.880 --> 00:15:23.039
but what it does do is

00:15:23.040 --> 00:15:25.119
it gives you a much better Emacs interface

00:15:25.120 --> 00:15:26.919
that's not terminal-based,

00:15:26.920 --> 00:15:29.639
because you're not using it through the terminal,

00:15:29.640 --> 00:15:31.239
or not even through comint,

00:15:31.240 --> 00:15:35.599
you are using it through a backend

00:15:35.600 --> 00:15:37.499
that is exchanging structured information

00:15:37.500 --> 00:15:40.999
with this process that is doing all the work.

00:15:41.000 --> 00:15:41.900
But other than that,

00:15:41.901 --> 00:15:44.519
it's the same model as Claude Code

00:15:44.520 --> 00:15:52.059
and projects of that nature.

NOTE Editing

00:15:52.060 --> 00:15:56.159
We've seen in the demos that I gave

00:15:56.160 --> 00:15:58.639
that there are AI experiences

00:15:58.640 --> 00:16:01.279
that are very natural in the world of editing.

00:16:01.280 --> 00:16:05.339
because they, like Copilot, just offers completion,

00:16:05.340 --> 00:16:09.479
it fits very well with what we all do in Emacs.

00:16:09.480 --> 00:16:14.279
And it's truly, yes, it's kind of a cheat in a sense

00:16:14.280 --> 00:16:15.639
for editing experiences,

00:16:15.640 --> 00:16:20.159
because it can do so much, but it's just editing.

00:16:20.160 --> 00:16:25.259
Whereas things like gptel and those kinds of tools,

00:16:25.260 --> 00:16:29.799
they are clearly in an editor and using editor,

00:16:29.800 --> 00:16:35.319
they're using Emacs, but they represent sort of like, well,

00:16:35.320 --> 00:16:37.759
you can edit for a while, then you could use these tools

00:16:37.760 --> 00:16:39.479
to do something that is not editing,

00:16:39.480 --> 00:16:45.899
this AI just changing the buffer for you. And that's fine.

00:16:45.900 --> 00:16:48.399
It's still... It may not be editing,

00:16:48.400 --> 00:16:52.033
but it's still clearly something that

00:16:52.034 --> 00:16:55.567
is useful to do in Emacs

00:16:55.568 --> 00:16:57.039
and belongs in Emacs.

00:16:57.040 --> 00:17:01.859
But the new tools like Claude Code and things like that

00:17:01.860 --> 00:17:02.639
are kind of different.

00:17:02.640 --> 00:17:06.639
Yes, they will get better integrated with Emacs,

00:17:06.640 --> 00:17:11.639
but it's not clear that they really need to.

00:17:11.640 --> 00:17:15.479
They can do a lot of things without editing.

00:17:15.480 --> 00:17:19.239
In a sense, editing is obsolete in some sense.

00:17:19.240 --> 00:17:23.459
For as many tasks, you don't need to edit anymore.

00:17:23.460 --> 00:17:26.439
And that's a nice thing.

00:17:26.440 --> 00:17:30.579
No one really knows when all this will end,

00:17:30.580 --> 00:17:36.879
how far things will go. It could be that in a decade or so,

00:17:36.880 --> 00:17:41.039
no one's really editing for work anymore.

00:17:41.040 --> 00:17:43.159
Maybe you're just writing instructions.

00:17:43.160 --> 00:17:44.319
You could do that with anything.

00:17:44.320 --> 00:17:47.439
You don't need Emacs or any special editor.

00:17:47.440 --> 00:17:50.439
We could all be using Notepad. That would be bad.

00:17:50.440 --> 00:17:58.039
But... I think it could go that far,

00:17:58.040 --> 00:18:01.839
but it could be that, well, for many specialized things,

00:18:01.840 --> 00:18:04.359
people are still using editing for certain tasks,

00:18:04.360 --> 00:18:07.000
but most tasks are getting fed to just...

00:18:07.001 --> 00:18:08.839
AI is just doing those things.

00:18:08.840 --> 00:18:15.759
In any case, I think it's clear that editing is diminishing,

00:18:15.760 --> 00:18:17.959
the need for editing itself is diminishing.

00:18:17.960 --> 00:18:21.879
And in such a world, It's interesting to think

00:18:21.880 --> 00:18:24.799
where Emacs is headed, especially in relation to

00:18:24.800 --> 00:18:26.359
all the other editors.

00:18:26.360 --> 00:18:28.599
I think people will use Emacs less.

00:18:28.600 --> 00:18:31.639
But I think other editors, like VS Code,

00:18:31.640 --> 00:18:37.999
may simply disappear or be a relatively fringe tool.

00:18:38.000 --> 00:18:42.719
And Emacs is going to follow its own path.

00:18:42.720 --> 00:18:44.679
It's very extensible. It could do anything.

00:18:44.680 --> 00:18:47.919
If there's one thing Emacs can do, it's adapt.

00:18:47.920 --> 00:18:51.679
Emacs has been around for a long time.

00:18:51.680 --> 00:18:54.799
It's pretty clear that Emacs will be around for a long time.

00:18:54.800 --> 00:18:58.879
It might be that in the future,

00:18:58.880 --> 00:19:04.339
editing is some sort of like an artisanal activity that we do.

00:19:04.340 --> 00:19:05.599
It's kind of weird to think about it.

00:19:05.600 --> 00:19:07.679
It's not like baking bread.

00:19:07.680 --> 00:19:10.079
But it is the sense that AI might be

00:19:10.080 --> 00:19:12.399
churning out code in the way, you know,

00:19:12.400 --> 00:19:14.199
the factories are turning out bread,

00:19:14.200 --> 00:19:17.139
but if you really want the good stuff,

00:19:17.140 --> 00:19:20.999
you'll have to do it yourself.

00:19:21.000 --> 00:19:23.959
I don't know if it'll be exactly like that,

00:19:23.960 --> 00:19:29.519
but it could be that Emacs survives and thrives

00:19:29.520 --> 00:19:33.559
in a very kind of specialized ecosystem of people

00:19:33.560 --> 00:19:35.599
who contribute and use it in the way

00:19:35.600 --> 00:19:39.539
it has survived and thrive right now.

00:19:39.540 --> 00:19:46.139
And I think that's a really nice way for all this to end up.

00:19:46.140 --> 00:19:48.719
There's the whole sense of how society will end up

00:19:48.720 --> 00:19:50.759
if all this happens. I don't know,

00:19:50.760 --> 00:19:54.639
but Emacs will be there for us when whatever happens.

00:19:54.640 --> 00:20:00.079
So thank you, and let's help make Emacs the best it can be

00:20:00.080 --> 00:20:04.880
to survive and thrive in the next decade.