Logs: liberachat/#haskell
| 2025-10-30 18:18:34 | → | L29Ah joins (~L29Ah@wikipedia/L29Ah) |
| 2025-10-30 18:20:49 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 2025-10-30 18:25:40 | → | Googulator55 joins (~Googulato@2a01-036d-0106-03fa-9dbb-a0af-2124-a319.pool6.digikabel.hu) |
| 2025-10-30 18:25:40 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 256 seconds) |
| 2025-10-30 18:25:46 | × | Googulator54 quits (~Googulato@2a01-036d-0106-03fa-9dbb-a0af-2124-a319.pool6.digikabel.hu) (Quit: Client closed) |
| 2025-10-30 18:31:47 | → | poscat joins (~poscat@user/poscat) |
| 2025-10-30 18:33:33 | × | poscat0x04 quits (~poscat@user/poscat) (Ping timeout: 252 seconds) |
| 2025-10-30 18:36:37 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 2025-10-30 18:40:56 | × | L29Ah quits (~L29Ah@wikipedia/L29Ah) (Read error: Connection timed out) |
| 2025-10-30 18:41:25 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 264 seconds) |
| 2025-10-30 18:43:56 | × | pr1sm quits (~pr1sm@24.91.163.31) (Remote host closed the connection) |
| 2025-10-30 18:46:37 | → | Tuplanolla joins (~Tuplanoll@91-159-187-167.elisa-laajakaista.fi) |
| 2025-10-30 18:47:50 | → | haltingsolver joins (~cmo@2604:3d09:207f:8000::d1dc) |
| 2025-10-30 18:52:01 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 2025-10-30 18:55:08 | → | tromp joins (~textual@2001:1c00:3487:1b00:5978:a504:f2fd:26f) |
| 2025-10-30 18:55:50 | × | tromp quits (~textual@2001:1c00:3487:1b00:5978:a504:f2fd:26f) (Client Quit) |
| 2025-10-30 18:56:58 | → | ttybitnik joins (~ttybitnik@user/wolper) |
| 2025-10-30 18:57:04 | → | tromp joins (~textual@2001:1c00:3487:1b00:5978:a504:f2fd:26f) |
| 2025-10-30 18:58:50 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 256 seconds) |
| 2025-10-30 19:07:34 | → | L29Ah joins (~L29Ah@wikipedia/L29Ah) |
| 2025-10-30 19:08:13 | × | synchromesh quits (~john@2406:5a00:2412:2c00:a151:32b5:2959:c671) (Read error: Connection reset by peer) |
| 2025-10-30 19:08:16 | × | L29Ah quits (~L29Ah@wikipedia/L29Ah) (Read error: Connection reset by peer) |
| 2025-10-30 19:09:26 | → | synchromesh joins (~john@2406:5a00:2412:2c00:a151:32b5:2959:c671) |
| 2025-10-30 19:10:03 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 2025-10-30 19:10:07 | → | L29Ah joins (~L29Ah@wikipedia/L29Ah) |
| 2025-10-30 19:14:41 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 244 seconds) |
| 2025-10-30 19:17:50 | → | rvalue- joins (~rvalue@about/hackers/rvalue) |
| 2025-10-30 19:18:55 | × | rvalue quits (~rvalue@about/hackers/rvalue) (Ping timeout: 264 seconds) |
| 2025-10-30 19:18:58 | × | tromp quits (~textual@2001:1c00:3487:1b00:5978:a504:f2fd:26f) (Quit: My iMac has gone to sleep. ZZZzzz…) |
| 2025-10-30 19:20:22 | × | haltingsolver quits (~cmo@2604:3d09:207f:8000::d1dc) (Ping timeout: 256 seconds) |
| 2025-10-30 19:23:38 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 2025-10-30 19:27:16 | rvalue- | is now known as rvalue |
| 2025-10-30 19:28:49 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 264 seconds) |
| 2025-10-30 19:35:34 | × | opencircuit_ quits (~quassel@user/opencircuit) (Remote host closed the connection) |
| 2025-10-30 19:36:44 | → | opencircuit joins (~quassel@user/opencircuit) |
| 2025-10-30 19:39:28 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 2025-10-30 19:44:44 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 256 seconds) |
| 2025-10-30 19:45:25 | → | segfaultfizzbuzz joins (~segfaultf@23-93-74-222.fiber.dynamic.sonic.net) |
| 2025-10-30 19:46:04 | × | Sgeo quits (~Sgeo@user/sgeo) (Read error: Connection reset by peer) |
| 2025-10-30 19:48:52 | → | Sgeo joins (~Sgeo@user/sgeo) |
| 2025-10-30 19:55:15 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 2025-10-30 19:56:33 | → | peterbecich joins (~Thunderbi@172.222.148.214) |
| 2025-10-30 19:57:01 | <segfaultfizzbuzz> | what are the norms these days regarding using "ai" to code among good quality professional programmers. is it fine to use or do i need to type everything into my keyboard myself |
| 2025-10-30 19:59:08 | <EvanR> | you will be ridiculed for using your keyboard at all. Voice input to an LLM is the only way to signal how up to date you are |
| 2025-10-30 20:00:06 | <segfaultfizzbuzz> | hahaha... but seriously...? |
| 2025-10-30 20:00:08 | <EvanR> | in the same way that handwriting is not a thing anymore |
| 2025-10-30 20:00:10 | <haskellbridge> | <sm> it varies a lot |
| 2025-10-30 20:00:14 | <haskellbridge> | <loonycyborg> Direct neural uplink better |
| 2025-10-30 20:00:19 | <geekosaur> | ai gets haskell very wrong still |
| 2025-10-30 20:00:22 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 260 seconds) |
| 2025-10-30 20:00:27 | <segfaultfizzbuzz> | lol yeah sorry i forgot to mention using neuralink while using tesla self driving |
| 2025-10-30 20:00:31 | <geekosaur> | even js needs to be checked |
| 2025-10-30 20:00:36 | <segfaultfizzbuzz> | geekosaur: oh? nice :-) |
| 2025-10-30 20:00:43 | <haskellbridge> | <sm> I don't think you can generalise, it depends what you're doing |
| 2025-10-30 20:00:44 | <segfaultfizzbuzz> | js is awful,... rust is like,... not bad i find |
| 2025-10-30 20:01:02 | <EvanR> | I use it for C and it still needs to be checked, obviously |
| 2025-10-30 20:01:22 | <geekosaur> | the question is what it was trained on. if you have a lot of blog posts by people who're still learning the language, the code the AI will produce will mostly be at their level |
| 2025-10-30 20:01:57 | <haskellbridge> | <sm> yes also the model, the ai-based coding tool, the context, the prompts all matter |
| 2025-10-30 20:01:57 | <segfaultfizzbuzz> | geekosaur: there also is how you prompt,... if your language is better you get better results i think |
| 2025-10-30 20:02:04 | <geekosaur> | keep in mind that current AI still doesn't understand anything; it's a Markov bot with a smarter notion of how language fits together |
| 2025-10-30 20:02:24 | <haskellbridge> | <sm> the coding tools and chat bots are no longer just that |
| 2025-10-30 20:02:30 | <segfaultfizzbuzz> | and then there is architecting your application so that you can kind of limit the damage that can happen, but i would imagine that's the same as structuring code for writing on a team |
| 2025-10-30 20:02:34 | <geekosaur> | which means it's only as good as the Markov chains it can build from its training data |
| 2025-10-30 20:02:47 | <segfaultfizzbuzz> | hahaha markov chains :-) you might not be wrong there |
| 2025-10-30 20:03:10 | <EvanR> | if you start pasting large amounts of code generated by the LLM into the project without understanding any of it, well, it will start to break down, and there's plenty of memes about where this leads |
| 2025-10-30 20:03:23 | <geekosaur> | seriously, it explains a lot of things, including why LLMs in the first place |
| 2025-10-30 20:03:35 | <segfaultfizzbuzz> | geekosaur: why LLMs in the first place? explain? |
| 2025-10-30 20:03:43 | <geekosaur> | and makes a lot of sense if you think about it |
| 2025-10-30 20:04:11 | <segfaultfizzbuzz> | EvanR: yeah i would say that "without understanding any of it" isn't what i do, but it can save me a lot of round trips back and forth from documentation and also it can sometimes stich things nicely (type conversions, etc) |
| 2025-10-30 20:04:17 | <geekosaur> | why large language models are what led to something that comes across as "actual AI" |
| 2025-10-30 20:04:54 | <segfaultfizzbuzz> | roughly speaking if you can write the type signature of your function then ai seems like it can do decently at filling in the rest 50% to 75% of the time... |
| 2025-10-30 20:05:00 | <EvanR> | lol |
| 2025-10-30 20:05:01 | <segfaultfizzbuzz> | or at least that's what i find |
| 2025-10-30 20:05:13 | <EvanR> | no |
| 2025-10-30 20:05:21 | <segfaultfizzbuzz> | EvanR: oh? |
| 2025-10-30 20:05:44 | <EvanR> | the type signature is usually not enough to judge what you want it to do |
| 2025-10-30 20:05:52 | <EvanR> | maybe you mean the carefully chosen name of the function |
| 2025-10-30 20:05:56 | <segfaultfizzbuzz> | type signature plus a description/comment |
| 2025-10-30 20:06:17 | <segfaultfizzbuzz> | yeah sorry, not type signature in isolation, i meant type signature as a hard restriction on the validity of the LLM output, given a reasonable prompt |
| 2025-10-30 20:07:35 | <EvanR> | yeah, trying its own output against the type checker before producing anything, and retrying until it works, explains a lot of the performance I've seen |
| 2025-10-30 20:07:45 | <EvanR> | but I'm not sure if that's a thing |
| 2025-10-30 20:07:53 | <segfaultfizzbuzz> | EvanR: can you elaborate on "a lot of the performance I've seen" |
| 2025-10-30 20:08:14 | <EvanR> | on a particular product preview I was working with |
| 2025-10-30 20:08:19 | <EvanR> | integrated into the IDE |
| 2025-10-30 20:08:30 | <segfaultfizzbuzz> | by performance do you mean like memory and cpu usage? |
| 2025-10-30 20:08:32 | <EvanR> | the haskell code generator started slowing down to unusable after a while |
| 2025-10-30 20:09:14 | <EvanR> | I'm sure cpu usage and memory usage was high but no |
| 2025-10-30 20:09:33 | <EvanR> | code per minute |
| 2025-10-30 20:10:00 | <fgidim> | haskell code into an LLM? :O |
| 2025-10-30 20:10:07 | <segfaultfizzbuzz> | oh you think people are calling some centralized resource over the web at a high rate...? |
| 2025-10-30 20:10:12 | <EvanR> | into, and out of |
| 2025-10-30 20:10:25 | <segfaultfizzbuzz> | what is "the haskell code generator" otherwise? |
| 2025-10-30 20:10:27 | <geekosaur> | I hear of a lot of people using Claude… |
| 2025-10-30 20:10:34 | <EvanR> | I'm not sure how much this product was dependent on remote resources |
| 2025-10-30 20:10:37 | <geekosaur> | Or Copilot |
| 2025-10-30 20:10:38 | <segfaultfizzbuzz> | geekosaur: yep, i do... more specifically claude code |
| 2025-10-30 20:10:43 | <EvanR> | it was more like copilot |
| 2025-10-30 20:10:54 | <EvanR> | but some proprietary thing |
| 2025-10-30 20:10:54 | <segfaultfizzbuzz> | copilot used to be pretty bad, it improved more recently... claude code is better |
| 2025-10-30 20:11:03 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
All times are in UTC.