TLDR: This all should be more about how a necessarily diverse forum communicates to and with each other (forum culture) and arrives at a protocol on how to treat user needs for particular ‘solutions’ (which can also touch on the need for scripting) … than about the wholesale horizons / ethics of (non-)GAI, expert systems in critical contexts, or the general fate of engineering culture(s) …
First, I think this is a really high-value discussion, and I learn a lot – especially about how the general social discourse about fundamental questions of AI (projective and analytical, re. history of ‘lived’ computer culture(s), STS-wise, in terms of ethical and normative even philosophical considerations, cultural diagnostics, etc.) is trickling down into individual positions and micro-discursive greenhouses, like the DT-forums. So, big thanks to everyone in here sharing views. I listen and sometimes learn.
But I generally feel this is a very high-level discussion. I enjoy it in a lot of ways. But I think the generalizing and ‘vested positions’ manner also has its pitfalls. And loses its contextual grounding with respect to critical things like some pragmatic guidance about how the community coalesces around shared and mutual practices given a dualistic situation:
– 1) the fact that some people can/want to (learn to) script and some never will; and 2) the fact of computer- or ‘AI’-assisted programming, and the routes this opens up to everyone – and in different ways – from pro-coders to companies, to ‘individual pro-sumer users’. 3) The fact that there regularly arises the challenge for ‘ordinary’ (read here: ‘non-coding’) users to mediate the onboard capabilities of DT (including its very helpful provided script library) with a situation where these do not suffice.
The situation that I think should be the start of a user forum discussion is that – these individuals can either turn to the community of ‘script angels’; – let the problem remain unsolved; – turn to a feature request (though there is no culture of ‘acknowledging FRs here at DT’); – or turn to the hacking means modern state-of-the-art consumer technology provides to them (actually: pushes to them).
The answer can’t be, as some people here convincingly lay out, “let’s treat ‘hacked code’ as ‘regular coding’” (as done by some – somehow – professionalized coders). I think the answer at the same time can’t simply be “anyone has to learn to code”. That is socially just somewhere between impossible and precocious. – I also wouldn’t expect/anyone ask to do a course in social moderation or nonviolent communication, even though these by now belong to the standard repertoire of social interaction, resp. will be critical to human survival.
Given that, I think @uimike hit the nail by really rendering this – on top of layers like the question of augmentation, human-machine interaction, professional ‘state of affairs’ – as an issue of communication and (necessarily negotiated) identities in such a situation of ‘encounter’ (which really is what a ‘forum’ by its very name is). This always also means collective negotiations and the need for a culture of mutual respect and some attention for diversity. And the attempt to create shared understandings and protocols, without forcing anyone into any other identity than the one he/she ‘naturally’ brings. (I think transformation of personality is important; but it shouldn’t be the business of a software forum…)
In the collaborative contexts I value, all these challenges – as well as the new affordances, like AI-assisted programming, as a special case of, hm… “democratized”, expert systems – really are treated as a reason to deliberate a shared practice given a heterogeneous situation and distributed competencies. Where I come from, these are ‘solved’ by negotiating things collectively and mutually given a shared set of overarching aims. Which I think in this case would be to maximize the efficient and helpful ways people can make use of DT, given different starting situations, work contexts, styles of doing things, etc.
The result of such shared deliberation normally – in the best case – would be a shared understanding (culture, protocol) of how to treat the use of AI-assisted coding as a potential fact in a situation that is mainly defined by people trying to make the best (individual) use of DT, and secondly be a cohort of people that do not know each other personally (mostly), and are comprised of coders as well as non-coders, and some people in between. I’d say this is just the sociological reality of the crowd here.
So, in reference to prior discussion, and stating the approach of one ‘non-coding’ thorough DT-user, I can say:
• I value the script culture around DT, and all that it provides, normally being very helpful to non-initiated users (even if at times a little in danger of being paternalistic and elitist in implied tonality; see examples given above and elsewhere). I also see the script layer as an additional layer, and in normal/standard cases, people shouldn’t be necessitated/expected to write extra scripts when they enter into the scene as users buying into a GUI-software.
• When I use ‘AI’-assisted programming, which I do only in rare cases, I don’t count/qualify it as ‘proper coding’, but rather see it as a hack – in a repertoire of these hacks we all do daily to get things done in systems we can’t fully master (and nobody really can do that). And I do not see/treat/evaluate it as using augmentation means in life-critical contexts (I think the discussion goes a little off the rails here). – I also see it as a test of what I can achieve individually, without using/relying on external sources, and a way to just ‘check’ things and new ways that might be useful and might not be.
• When I do so (hacking code in this way), I’d never expect anyone to ‘fix’ it, and honestly, I haven’t seen an approach of people here “coaxing” others to do their work beyond any normal mutual exchange situation.
• I also think, fundamentally, the community overall has to cater to and find a meaningful, mutually respectful social intercourse vis-a-vis the facts that a) some people can code; b) some never will; everyone still needs to – respectfully – talk to each other, with a view of maximizing what the community can wrench from DT
… what would help here is a shared understanding of how this kind of (necessary) negotiation happens in terms of a culture (and maybe protocols) of constructive interaction. And mutual respect, also encompassing respect for different competencies lying with different individuals. That is – IMHO – the real challenge to a “forum” of users(!) in a situation of mixed competencies, capabilities and needs.
PS @jonmoore – sorry for picking a particular subset of the touched-upon questions, here strictly from such a particular perspective of a ‘non-coder’ and ‘ordinary user’. I am aware this is not 100% in alignment with the original intent of the thread. But I also think it aligns with a lot of the issues, (implicit) questions, and sentiments that are/were raised here. … and finally, it touches on the sheer reality of something like Claude Code now being generally available, and what it means for a DT-forum context. – I am very sure other aspects will be continued in the discussion, even if I share/insert a very particular highlighting here…
