|
00:06
erooke left
00:18
nuno left
00:22
erooke joined
02:16
mehbark left
02:23
erooke left
02:34
hulk joined
02:35
kylese left
02:39
erooke joined
03:15
hulk left,
kylese joined
04:30
stanrifkin_ joined
04:32
stanrifkin left
04:40
erooke left
04:56
erooke joined
05:29
vasko4535586169 left,
vasko4535586169 joined
05:52
hurufu joined
06:14
belluzj joined
06:19
belluzj left
06:31
stanrifkin_ left,
stanrifkin_ joined
06:48
human-blip left
06:50
timo left
06:56
erooke left
06:58
iwuvmika left
07:09
erooke joined
07:12
iwuvmika joined
07:32
human-blip joined
07:47
Sgeo left,
human-blip left
07:50
human-blip joined
08:27
human-blip left,
human-blip joined
09:12
erooke left
09:29
erooke joined
09:30
belluzj joined
09:46
belluzj left
10:49
samebchase left
10:50
samebchase joined
11:29
erooke left
11:47
erooke joined
12:54
belluzj joined
13:14
belluzj left
13:23
Voldenet left
13:34
Voldenet joined
13:48
erooke left
14:02
erooke joined
14:23
dg left
14:27
dustinm` left
14:34
dustinm` joined
|
|||
| lucs | I truly appreciate the help I get from people here to my usually simple questions. | 14:50 | |
| But for more involved stuff, where would you recommend I go?: Stack Overflow, Reddit, some mailing list? | |||
| disbot | <antononcube> LLMs. | 14:52 | |
| lucs | No thanks. | ||
| But thanks anyway :) | |||
| disbot | <antononcube> Ok. Damn! I was looking for an excuse to finish an AI-skill that peruses the docs.raku.org for answers... | 14:53 | |
|
14:54
snonux left
|
|||
| disbot | <antononcube> That should be done anyway -- many/most systems for technical computations have those kind of enhancements. | 14:54 | |
|
14:55
snonux joined
14:59
stanrifkin_ left
15:51
hurufu left
15:52
hurufu joined
16:01
erooke left
16:14
Voldenet left
16:15
Voldenet joined
16:18
erooke joined
|
|||
| tbrowder | antononcube: how would such a thing be better than asking google | 16:27 | |
| i use it all the time with pretty good results | 16:29 | ||
| or as i tell friends “ask dr google” | 16:31 | ||
| disbot | <antononcube> Yes, I agree "Google knows" Raku well enough. Do you trust those directions, though? Also, very often I want to have concrete documentation page and section or paragraph that describe what I am asking for (about Raku language). | 16:34 | |
| refactus | another resource I've found valuable at times is the search feature on this channel's log | 16:36 | |
| disbot | <antononcube> Another important reason is to use (relatively) small locally run models. I can force them to understand better the relevant Raku documentation and form an answer on a smaller (narrower) amount of text. | ||
| <antononcube> The AI-agent skill I have mind is primarily retrieving relevant pages of docs.raku.org. (Then those can be used in whatever workflows.) | 16:37 | ||
| lizmat | .oO( too bad you still need humans to write documentation for LLMs to rip ) |
16:52 | |
| tbrowder | can’t argue about. finding authoritative docs for publication. but as lizmat is saying i think ai has skipped the old stuff where we are expected to credit our sources for our published stuff | 16:58 | |
|
16:58
dmvrtx left
16:59
dmvrtx joined
|
|||
| tbrowder | pres j f kennedy’s famous speech (ask not what your country…) was not truly original, his speech writer took it from a much earlier speech by us army general omar n bradley. | 17:10 | |
| that sentence that is | 17:11 | ||
|
17:21
coleman left,
coleman joined
|
|||
| disbot | <antononcube> > Ask not what can LLMs do for you, but what can you for the LLMs. … I guess. | 17:38 | |
| <antononcube> > Ask not what can LLMs do for you, but what can you do for the LLMs. … I guess. | 17:39 | ||
|
17:55
abraxxa joined
|
|||
| tbrowder | any idea how ai avoids copyright infringement? | 18:00 | |
| consulting my lawyer son… | 18:05 | ||
|
18:19
erooke left
|
|||
| disbot | <antononcube> Everything LLMs is publicly available. Similar case were already brought up against Google. There is (almost) 1-to-1 correspondence. | 18:25 | |
| <antononcube> Everything LLMs are trained with is publicly available. Similar case were already brought up against Google. There is (almost) 1-to-1 correspondence. | 18:27 | ||
|
18:35
erooke joined
|
|||
| tbrowder | ok, good to know, thanks | 19:02 | |
| my son says for there’s a suit by New York Times against one of the AI firms for copying all their data. | 19:39 | ||
| japhb | tbrowder: As your lawyer son may have already told you: This whole area of regulation and jurisprudence is in heavy flux. Worldwide things are changing month to month. | 19:40 | |
| (Of course, how individuals, projects, and corporations respond to that flux is all over the map as well.) | 19:41 | ||
| tbrowder | yep, he did. he also said he expected ai users to copyright artistic things they create. | 19:44 | |
|
19:49
dustinm` left
19:51
dustinm` joined
20:02
iwuvmika left
|
|||
| disbot | <scullucs> Turns out I tried Claude (first time) and it solved my problem perfectly 🙂 | 20:06 | |
| japhb | As a side note on the searching question, I believe Google recently publicly said something along the lines of "You can't use Google *without* using AI at this point." | 20:12 | |
| So Google "instead of" an LLM isn't really a thing, I don't think. | |||
| japhb remembers when it was purely algorithmic ... sigh. | 20:13 | ||
|
20:26
dustinm` left
20:27
dustinm` joined
|
|||
| tbrowder | japhb: think you are correct. i've been referring to "him" as "dr google" for some time now. | 20:29 | |
| japhb | nod | ||
|
20:35
erooke left
20:52
erooke joined
21:11
hurufu left
21:34
hurufu joined
22:33
abraxxa left
22:51
erooke left
23:00
hurufu left
23:05
erooke joined
|
|||