Zot Nobot
Larry GarfieldI may regret this at some point, but I felt the need to put down in writing how I feel about this moment in the tech industry.
It is not kind. You may well be insulted by it. If you are... then you really should question yourself.
William Whitlow@Crell Thank you for quantifying the various aspects that it takes to keep AI running. It’s one of those details that is becoming increasingly apparent, but difficult to grasp the scope.
Would you care to elaborate on the “This is how societies die.” comment? Is that primarily in the sense of social apathy? In the lack of respect for others/foresight over long term consequences? The disruptive tactics of numerous tech companies that have eliminated many norms? Or something else?
@wwhitlow All of the above.
Major empires aren't destroyed from without, but by their own greed, infighting, and incompetence.
In this case, the "it is what it is" attitude is unworthy of someone in a democracy. That's how the billionaires and pedo-fascists were able to take over.
Then add short-sightedness about global warming for the last 50 years, and AI is just the latest part of it. That will kill us all. "It is what it is."
Dávid Bárdos@Crell @wwhitlow AI sometimes reminds me of lead.
The ancient Romans knew it was poisonous, especially in anything that touched food or drink. Yet they used it in their aqueducts anyway. Cheap, easy to mine (= comfortable to access), so no one cared about the consequences.
Sure, lead was not the only cause of Rome’s collapse or the thousand years of barbarism that followed. But cognitively impairing entire urban populations surely took its toll on political competence. 🤷
toni 
@david_bardos @Crell @wwhitlow Similar to how every covid infection impairs cognitive ability, but we have accepted that fighting it with air filters and masks would be too expensive and awkward.
Fabian Egli@david_bardos @Crell @wwhitlow
Edited after reading:
https://en.wikipedia.org/wiki/Roman_lead_poisoning_theory
Well, I learned something today.
I agree with you that the widespread use of led has striking parallels to today's AI use.
Jean-David Bar@fabianegli @david_bardos @Crell @wwhitlow AI, plastics and microplastics, pesticides... Many modern lead versions we have to survive to...
Derick Rethans@Crell @wwhitlow I think the societal aspect goes further than that. Because we are ending up not teaching anything person to person any more.
Take StackOverflow. Although there were often simple questions (and answers) with little thought, there was also a large amount of great questions, with equally great answers — so detailed that you learned the actual basics. This is now gone due to AI.
LLMs have only learned old content from there, anything newly created projects won't even be taught.
ekes@derickr @Crell @wwhitlow Writing and maintaining code - not one off projects for x - but all the rest that they are built on - is a social process in and of itself. What happens when the discussion and the collective learning and memory about it is lost?
Feels minor in relation to many of the global and localized impacts mentioned in the post, but one that I've not seen thought through elsewhere.
SeanBurlington 🌈 🕊️@derickr @Crell @wwhitlow I hear a lot of people stating that AI will get better with time - I see no evidence for this
It has run out of data to consume - data that it needs in order to improve
It has poisoned the well of knowledge
and since it is currently making massive financial losses - we cannot evaluate future pricing
Dagnabbit, Pascaline! 💥And since it is adding to the well of knowledge itself, using old knowledge, this additional information is a rehash of what is already there. It's like the 'websites' written by Ai, after which Ai uses them to add to its 'knowledge'. The worst is that these get more traction and search results. But they feel like voids echoing one another, not adding anything new, except if they find new human input to milk and bleed.
Wilfried Klaebe"AI" is pissing into the well of knowledge we all are trying to drink from.
Dan Schnaui was thinking about this, and I'm not sure AI has consumed all the data. It's consumed all the *public* data. But with companies wiring up their private codebases to LLMs, there's lots and lots more data to feed from.
El Duvelle@elduvelle @wwhitlow @Crell @derickr Not yet, but its traffic is down to its first year of existence. With less people on there, good questions and answers won’t be generated anymore.
@toni @wwhitlow @Crell @derickr
Or maybe only the most useful or complex questions will be asked & answered there? Or, it could become a "AI-free" space where people only interact with people? It might mean a reduction in users but it would still be extremely valuable to the community. I guess it depends how the people in charge decide to orient it. But I don't think it will ever die, it's too valuable.
@elduvelle @toni @wwhitlow @Crell I do support for my Xdebug tool on there, and instead of 2-3 per day, it's now less than one a month: https://stackoverflow.com/questions/tagged/xdebug — people are just not using it any more.
@derickr @elduvelle @toni @Crell
What can be said for StackOverflow, it is certain that the landscape of knowledge is changing. Looking to Wikipedia is probably the best approach for the AI free writing policy. https://en.wikipedia.org/wiki/Wikipedia:Signs_of_AI_writing
This new article from Anthropic might be of interest. It test skill acquisition of Junior Devs with and without AI. The sample size is somewhat small, but may offer some insight. https://www.anthropic.com/research/AI-assistance-coding-skills
Ken Butler
Greg Harvey 🌍@derickr @Crell @wwhitlow Coincidentally, I posted Larry's original blog post on LinkedIn as well and the silence is palpable. In case you were in any doubt it's techbro central over there. 😂 I'm sure there are people who share the sentiment, but they're not going to stick their head above the parapet, clearly...
James Widman@greg_harvey @derickr @Crell @wwhitlow
> And what happens when nothing's feeding AI except other AI??
The Burger²
Dominik@derickr @Crell @wwhitlow StackOverflow is gone because AI doesn't tell you that this question is a duplicate of a question solved 15 years ago where literally nothing applies because 15 years is like 2 centuries in IT time.
Love or hate AI, but StackOverflow was just waiting for something to replace it because of how bad it became.
mathew
Not a Spring Onion
Reddog
Guillaume Rossolini@Crell @wwhitlow there was a podcast by Paul Cooper titled Fall of Civilizations, which focused on precisely that aspect
And as I recall, it was exactly as you describe for all past civilizations: greed and power struggles
(Perhaps amusingly, the podcast itself stopped publishing new episodes shortly after increasing its own scope to video instead of audio)
Alejandro Celaya@Crell Thanks for writing this down. Now I have something well-written to share when someone asks how I feel about this topic.
Erin Dalzell (He/Him) 🇨🇦@Crell thank you for writing this
|7eter l-|. l3oling 🧰@Crell great analysis. Lots to think about. Many of us in software are now between a rock and a hard place - exactly as the billionaires want it.
Alessandro Lai@galtzo I think that "rock and a hard place" is the situation in which the majority of us feels right now, and contributes to the overall scenario.
I personally feel there too, and as a father I'm only a bit less crushed than @Crell probably only because my son old enough to start teaching him about things like these.
My hope is that the bubble will burst soon, making the job market bounce (removing the hard place) and skirting the AI fad to only where it's really useful.
Florian Engelhardt@alessandrolai @galtzo @Crell I do believe that this is an astonishing huge bubble (bigger than dot-com and crypto) and that it will burst.
I doubt it'll burst soon, just because the amount of VC money being around is unbelievable.
I doubt it'll burst soon, just because the amount of VC money being around is unbelievable.
@galtzo @Crell @flowcontrol that's probably true, but I still see that money is also the biggest bottleneck of this technology, so it seems a race to the bottom to me...
Claudio Zizza 🦜@alessandrolai @galtzo @Crell Sadly that won't happen so soon. To many companies are still on the silver bullet train.
@galtzo @Crell @SenseException yeah, but what if the bullet costs too much / more than what you can save with it?
Alessandro Lai (@[email protected])
@[email protected] @Crell @flowcontrol that's probably true, but I still see that money is also the biggest bottleneck of this technology, so it seems a race to the bottom to me...
Rubén Rubio@alessandrolai @galtzo @Crell @SenseException That is what Ed Zitron is saying since a while: https://www.wheresyoured.at
It is my go-to place to understand the AI bubble. It looks like it will burst this year.
@alessandrolai @galtzo @Crell Once a company is fixated that something is a silver bullet, the harder it gets for them to drop it.
Kotking@Crell as someone who didn't work in tech directly but like to read or see stuff from other experts. I saw that LLM is a blight pretty early myself (I think).
I heard that it's a must have as many innovations humankind made... While person will never bother looking up how many innovations humankind made be they failing along the way or being cruelest but efficient. LLM hammered in into everyone as it's a toy, playful stuff so you shouldn't worry about costs. The longer it goes the worse it gets but those who get burned are miniscule compared to those who only started using LLM.
We as society know about microplastics, covid, LLM, global warming but society will still use plastic everyday, won't wear mask in public transport, will play around with LLM, will contribute to organisations that burn our ecology to the ground.
There also wars be it your neighbour, yourself or far away from you. So there that whole aspect that affects our planet but you... The individual is the problem for not changing your habits. But it always comes that one does what everyone do, so few people masking doesn't mean everyone in bus will mask. You being vegan won't make others eat less. You not smoking won't stop assholes from smoking in public place. Social drinking is a must for most cultures.
You have to go against flow, you don't need to do all of the good stuff but changing to be vegan or to avoid use LLM... I think it's pretty easy to do one then other.
oatmealTalking about #AI while ignoring the ethical and societal issues involved is like talking about #Palestine without mentioning the ongoing Israeli occupation or colonialism.
[…] It only is what it is because you're OK with what it is, and aren't putting in the work to make it otherwise. Those that don't give a fuck about fair copyright application, about poor people, about our planet, are putting in the work to make it so. And you're letting them.
I do not forgive you.
https://www.garfieldtech.com/blog/selfish-ai
#ai #palestine #theaicon #aihype
C/c @Crell
Jim Gardner@Crell I completely agree with everything you’ve written here. Personally, I held out using any LLM services until it became a requirement from management, at which point I chose to fall in line because I’d prefer to have a job. But I’ve never been comfortable or happy about it. This tech, in the way it’s being built by the big companies, is anti-human collective mania. So while I did yield, I’m not going to say “it is what it is”. Unless what “it is” is a catastrophe.
Alerta! Alerta!@Crell Thank you for putting my thoughts into such elaborate words!
The challenge I see is that - especially after COVID - we are now living in a world of "Me" instead of a world of "We".
@heiglandreas That's been building for 50 years. Every conservative businessman in the 70s that decided to embark on this decade long destruction of the world so they could get even richer.
There is no circle of hell deep enough for them.
Craig Wright 🐵
sarah tonin 
@Crell i've done enough "compromises" (read: being forced into something i don't want) as is, i don't want to be forced into more
krig@Crell Here's something that rubs me the wrong way: "At some point soon, I will have to figure out how to work with AI coding tools" - No, you don't. That's a choice you are making. In the end, you are making the same choice as those who say "It is what it is". You choose personal convenience over your convictions. Someone has to be the first to refuse. Someone has to be the one to shoulder the burden. What makes you think that someone isn't you?
@krig @Crell While I don't disagree, that's a huge ask. If someone does a thing to pay the bils and everyone else is doing that thing faster with AI, those "convictions" could land them and their family without pay or benefits. That's a big gun to stare down the barrel of. That's the fear. That's what needs to be addressed. It may well be a false narrative, time will tell, but it's a big leep of faith to ask anyone to take.
@JustinMac84 @Crell First off, other people are already losing their jobs regardless of the choices they make personally, so to even have a choice puts them in a privileged position. Second, what are their principles worth if all it takes is to threaten with hypothetical (not real) joblessness? How about making an attempt at a principled stance before caving in? Third, the whole point of the article is that he looks down on those choosing personal convenience over conviction, yet here he is doing the same.
Lars Marowsky-Brée 😷@krig @JustinMac84 @Crell Yes - but also "not quite".
We'd be asking an individual to shoulder a systemic problem.
I respect everyone who makes that choice (if they feel they can without existential risks; or *despite* existential risks).
But that doesn't address the systemic problem.
That's why strikes aren't an individual refusing to work; but individuals unionizing and *collectively* protesting.
And why we regulate and legislate and organize.
@larsmb @JustinMac84 @Crell I don’t know if I agree. It feels like an easy cop-out. If every problem is systemic, surely there is no individual responsibility at all then? I think it is reasonable to say that if you are going to rant about someone elses individual choices you should be honest to yourself about the choices available to you.
@krig @JustinMac84 @Crell No. There are individual choices (and IMNSHO, obligations): organizing, voting, using one's platform, or at least *somehow* contributing towards collective action.
Given that we all kinda have to eat and pay for housing because capitalism, and possibly have non-obvious setups, asking someone to make a decision that has a significant risk of ruining their individual livelihood is a tough sell.
Organizing is how we get out of the prisoner's dilemma.
@larsmb @JustinMac84 @Crell I am not asking them to make a different choice (that is up to them. I make choices for myself). I am asking them to state clearly that they are making the choice, and not frame it as ”I have no choice”.
@krig @JustinMac84 @Crell Ah, that's indeed fair and a good requirement. (I think.)
Though holding that cognitive dissonance in one's head is not necessarily always healthy either (ask my therapist about it, because I can't *not*), especially when (it feels like) they have no *realistic* choice.
But yes, I'm also a strong advocate for at least admitting the situation we're in.
@larsmb @JustinMac84 @Crell I’ll also concede that this topic is not great for my own mental health, but I feel like so many of us are far too quick to jump to helplessness when widespread refusal (a strike if you will) would make a real difference.
@krig I feel that.
I've been wrestling a lot with that in the last 12 months in particular, and am still trying to fully make up my mind.
@krig @larsmb @JustinMac84 I am no stranger to collective action. I'm on the board of a voting reform NFP in my state.
There was a time where a conscientious dissenter could just not shop at Amazon.com. But because so many didn't care, it's now "use AWS or don't use the Internet."
Telling someone "OK, so just don't use the Internet" is not helpful. If we want to engage in society, we have to have some shared tools. Even if those tools are shit.