Hearing the feelings in this rant, which does touch a nerve, I can’t help think about how different the developer community reaction to the LLM push might be if the focus were on quality instead of efficiency.
-
But that’s me; I don’t think my ethical concerns are shared widely enough for companies to have to be ramming AI down developers’ throats the way they are. The token quotas etc are a symptom of something large and deep.
Maybe that post about MTBF vs MTTR helps explain it.
/end
@inthehands To briefly interact with this, while devs usually do go after something shiny, it is usually the shiny things propped up by major tech corporations that make it through without much resistance - See react, express, angular, etc.
The current state of tech has a serious evangelism problem. I think most people have seen "tech evangelism" in the wild in some form prior but now I feel it reached a new level with the current marketing and push of LLMs, the inherent trust the tech corportions got for free from many of them, the way they focused on management figures to promote it instead of people who would actively scrutinise this.
Given the state where selling software solutions was becoming difficult for many and all the major corporations ensuring that everyone had to race to the bottom or devalue software as much as possible to then ensure that no one competes, it isn't exactly surprising that they wanted to turn the industry into a casino where they benefit from the licensing of the pokies and all the devs have to play the game.
Combine this with an economy that has been cooked by these people, developer distrust andtrust from other organisations only being placed in large tech corporations (or those tech corps will lobby it because they have the means to do so).
Developers who are concerned just don't really have any ability to resist internally or have the bandwidth to do so unless there are financial punishment to organisations that lean into this rubbish (which... there will be but likely for most of them, they will go bankrupt before it is too late).
There is also an asynchronous game as well, if you received slop and have to handle it, you have been given the difficult work that is also not valued and it is now at such a large volume.
-
I can’t think of another time when software devs had to be •forced• en masse to use a new technology that was supposed to help them. Usually we’re kind of stupid for the shiny new things: jamming them in when they solve nothing, doing unnecessary rewrites just to use the new hotness because it’s so cool and fun. Usually we’re the one trying to shove it down mgmt’s throat (or sneak it by them) rather than the reverse.
But not this time.
7/
@inthehands Yep. The fact that AI is being *forced* on people is maybe the biggest red flag of them all how terribly, horribly wrong things are in this bubble. Like you said, people and developers in particular, will flock to a new useful technology on their own, it's more that you might need to hold them back. Now, we're being extorted to use it or lose your bonuses, maybe even your job! It makes no effing sense whatsoever.
-
The trouble is, as Doctorow points out, that this vision makes AI a multi-billion dollar industry, not a multi-trillion dollar industry.
Even if you can claim that your ML / LLM thinger can reduce software bug rates or failure rates by 10x — which would be •wild• — demand for that is simply not going to fund data centers the size of Manhattan.
But make the claim of •speeding up• by 10x — an even wilder claim, but one some people are desperate to believe! — and all the money in the world will beat a path to your door.
5/
@inthehands I have another problem with this vision of the AI-assisted good centaur: It doesn't dilute the costs enough. Even accounting for the highly VC subsidized prices now, people expect that it will be fine to pay $600/month for Claude when price dumping ends, because it will save you more than $601 of work. But that would be only if everyone and their uncle is buying.
The "good kind of centaurs" would not be that many. The price might end up like 10k/m. And then it's just not proffitable
-
@thomholwerda
I refuse to believe that anyone who's actually used google translate before can seriously believe that it can replace human translation. -
The “efficiency” pitch I’m describing upthread isn’t really “go faster;” it feels more like “making good things doesn’t matter, what you cared all along about doesn’t really matter, and we don’t think •you• matter.
We always just wanted to built absolute shit, and you always tried to stop us. But now at long last we can.”
9/
@inthehands This is a big part of it, I think. But also, study after study says the efficiency gains aren't even real, once you factor in review and big fixing and describing the problem in enough precision for the machine, and downtime.
The management class has been trained to not do real work, but to have the appearance of work. LLM code has the appearance of hyper productivity. Whether it does or not is irrelevant to them.
-
Consider those two different versions of the radiologist’s role: one as a valued human augmented by a machine, doing a job they believe in better than they’ve ever done it — and the other as a cog in a corporate process whose job is to perpetually deal with the machine’s mistakes.
Consider the parallels in software development. All vibe coding and “agentic” stuff points to the second: developers as slop wranglers, as accountability sinks, as exhausted and expendable workers on a code assembly line.
3/
@inthehands The "code assembly line" makes me wonder where all that code is going. Is there a breakdown of all those developer hours by whom they're for? Internal financial services apps? The 100th rewrite of thegap.com? Insurance claim processing? I'm constantly amazed at how the quantity of code seems unrelated to the quantity of *real stuff going on*, especially social-value-producing real stuff like making solar panels and batteries. Maybe there's a bubble in the aggregate code factory?
-
@thomholwerda @janeishly @inthehands for sure. Seen similar thing in historic fields. Shared history is going be different in different languages.
-
RE: https://hachyderm.io/@mitchellh/116580433508108130
Hearing the feelings in this rant, which does touch a nerve, I can’t help think about how different the developer community reaction to the LLM push might be if the focus were on quality instead of efficiency.
1/
@inthehands the first thing I bring up in my ai coding workshops (I don’t give all that many, but they’ve been spanning from 2024 to now) is to focus on engineering quality and using a human centered “communication analysis”. If you can map who communicates with whom in order to build the final software (which includes everybody from user to customer to support to design to qa to management to developers), then it becomes much more easily apparent where to use ai tools. It can be the main code path but it’s usually much more effective to build better ticketing systems or design experimentation tools etc…
It also means that you can be an “ai assisted developer” without using any ai tools yourself, because you now get better tickets due to a vibecoded chrome extension that attaches the right internal docs, or something.
Of course not many large scale companies will change their approach, but I like to think that there’s a few teams out there who get to follow that framing and find renewed purpose (I know I do) by seeing how much quality can be improved by reshaping the narrative around the technology.
-
But that’s me; I don’t think my ethical concerns are shared widely enough for companies to have to be ramming AI down developers’ throats the way they are. The token quotas etc are a symptom of something large and deep.
Maybe that post about MTBF vs MTTR helps explain it.
/end
@inthehands Might be fruitful to question the assumption that companies are mandating the tools in response to practitioners' resistance to them. Mandates might just be the instinctive tactic of profit-obsessed managers / leaders regardless, and this has just been an opportunity to indulge in them. Mandates express the same impositionist attitude that LLM tech does, the attitude that results in the disasters that you mention drove you away from that tech.
-
The energy and water usage used to be a problem for mainframes. Watt's separate condenser in 1765 was essentially a water-cooled heat exchanger — the whole efficiency gain came from using cold water to condense steam without cooling the cylinder itself.
So the water usage argument has been around a while longer.
Data sourcing? You're trying to sort out which little calf and piglet went into your Weisswurst. Copyright in its current form has ruined the music industry, so let's spread the joy around
The problem with AI is the same as the Mirror Test for animal cognition. Y'all can stop screaming at the mirror - it's just you in there.
@tuban_muzuru @inthehands
The data sourcing problem isn't as simple as copyright.What AI companies are doing is:
1. taking from a public resource (public websites, source code of FOSS projects, knowledge-sharing forums, etc)
2. destroying the public resource in the process (overloading the websites, flooding FOSS projects with slop PRs, disrubting the social dynamic required for forums to work)
3. renting you access to a worse version of the public resource they have destroyed
-
E em0nm4stodon@infosec.exchange shared this topic
