What price?

@adactio a little. Not really.

@baldur A lot of people don’t really care about their own integrity or perhaps never really had the chance to develop any in the first place. Being a consumer and not a maker is part of this I suspect, which is why they can so easily be sold the illusion of being creative. Also, some people (especially a lot of men who never got taught to care or share) are very entitled and selfish. Toxic masculinity isn’t all guns and rolling coal, it’s also smug patriarchal entitlement to exploit and disregard the consequences (you see this among men who think they’re more rational than everyone else, they are often HIGHLY self unaware people who are convinced they’re the good guy because their bigotry is pseudo-polite).

@adactio thanks to LLMs, everyone has forgotten about neural networks, and the nifty things that they can do. ☹️

Or worse, work like this Zooniverse project might now have a harder time attracting volunteers, because LLMs have poisoned the perception of machine-learning.

https://academic.oup.com/mnras/article/472/2/1315/3979473

A transient search using combined human and machine classifications

@baldur it is not going to get better because it is already pretty clear that copyright protections don’t apply to the training scenario. So you can 1) advocate for Congress to do something about this (I support this but am not sanguine) or 2) move on to new forms. Maybe text, image, audio, moving image, even code - all for which there are copious training sets (do you have a music library?) are all old media and it is time for creatives to move on. (As an artist, I am not afraid!)

@baldur what do I mean about moving on? Things like: live music and performance, site specificity, making, deep hang outs, going outside, getting a kiln and playing, and for me, distinguishing between artists who use models trained by others from artists who gather our own data and train our own models. #artisinalAI and collaborating with #AI (our new strangers… I actually welcome them.)

Maybe I will be among the early to bring a project with my own data and training:-) ?

ai artisinalai

@stalbaum @baldur The main reason not to be afraid as an artist is that so-called AI is an illustration machine not an art one and art hasn’t been about “realistic representation” since cameras were invented. There’ll be a couple of people (and are already) making bespoke AI artworks but I highly suspect that will also fizzle out to a large degree as novelty wanes (as is often the case with technological tools). This isn’t really that interesting a space in terms of technological/digital artworks or aesthetics* in my opinion but that doesn’t mean that some interesting work can’t be made (generally by subverting and breaking the technology).

*Other than talking about the aesthetics of fascism and it’s relationship to fantasy illustration and hyperrealism.

@adactio Hmm… so mad LLM disease… MLLM disease has nice symmetry… but “Mad Slop Disease” has a certain pungent ring to it

@adactio I really like this piece. A lot. I use Claude most often for coding assistance for my personal projects trying to get my data analytics chops in better shape again. Anthropic says the right things about their models being far less computationally expensive and sustainability is one of the key factors in their LLM product’s product decisions. Claude warns uses about the costs and efficient practices.

But, I’ve watched teams running multiples of these tools poorly and with poor results.

@adactio When this piece popped on yesterday I thought it was going to be about web development and poor practices with crazy sustainability costs associated with them. But, no…

The past week or two I’ve been digging for data and finding sites with full datasets in my browser in >2MB JSON strings. But, >2MB in front end frameworks (multiples for one page) for layout and light design. Then looking at >3MB in ads.

This is utterly crazy.

A couple of days ago I linked to a post by Robin Sloan called Is it okay?, saying:

Robin takes a fair and balanced look at the ethics of using large language models.

That’s how it came across to me: fair and balanced.

Robin’s central question is whether the current crop of large language models might one day lead to life-saving super-science, in which case, doesn’t that outweigh the damage they’re doing to our collective culture?

Baldur wrote a response entitled Knowledge tech that’s subtly wrong is more dangerous than tech that’s obviously wrong. (Or, where I disagree with Robin Sloan).

Baldur pointed out that one side of the scale that Robin is attempting to balance is based on pure science fiction:

There is no path from language modelling to super-science.

Robin responded pointing out that some things that we currently have would have seemed like science fiction a few years ago, right?

Well, no. Baldur debunks that in a post called Now I’m disappointed.

(By the way, can I just point out how great it is to see a blog-to-blog conversation like this, regardless of how much they might be in disagreement.)

Baldur kept bringing the receipts. That’s when it struck me that Robin’s stance is largely based on vibes, whereas Baldur’s viewpoint is informed by facts on the ground.

In a way, they’ve got something in common. They’re both advocating for an interpretation of the precautionary principle, just from completely opposite ends.

Robin’s stance is that if these tools one day yield amazing scientific breakthroughs then that’s reason enough to use them today. It’s uncomfortably close to the reasoning of the effective accelerationist nutjobs, but in a much milder form.

Baldur’s stance is that because of the present harms being inflicted by current large language models, we should be slamming on the brakes. If anything, the harms are going to multiply, not magically reduce.

I have to say, Robin’s stance doesn’t look nearly as fair and balanced as I initially thought. I’m on Team Baldur.

Michelle also weighs in, pointing out the flaw in Robin’s thinking:

AI isn’t LLMs. Or not just LLMs. It’s plausible that AI (or more accurately, Machine Learning) could be a useful scientific tool, particularly when it comes to making sense of large datasets in a way no human could with any kind of accuracy, and many people are already deploying it for such purposes. This isn’t entirely without risk (I’ll save that debate for another time), but in my opinion could feasibly constitute a legitimate application of AI.

LLMs are not this.

In other words, we’ve got a language collision:

We call them “AI”, we look at how much they can do today, and we draw a straight line to what we know of “AI” in our science fiction.

This ridiculous situation could’ve been avoided if we had settled on a more accurate buzzword like “applied statistics” instead of “AI”.

There’s one other flaw in Robin’s reasoning. I don’t think it follows that future improvements warrant present use. Quite the opposite:

The logic is completely backwards! If large language models are going to improve their ethical shortcomings (which is debatable, but let’s be generous), then that’s all the more reason to avoid using the current crop of egregiously damaging tools.

You don’t get companies to change their behaviour by rewarding them for it. If you really want better behaviour from the purveyors of generative tools, you should be boycotting the current offerings.

Anyway, this back-and-forth between Robin and Baldur (and Michelle) was interesting. But it all pales in comparison to the truth bomb that Miriam dropped in her post Tech continues to be political:

When eugenics-obsessed billionaires try to sell me a new toy, I don’t ask how many keystrokes it will save me at work. It’s impossible for me to discuss the utility of a thing when I fundamentally disagree with the purpose of it.

Boom!

Maybe we should consider the beliefs and assumptions that have been built into a technology before we embrace it? But we often prefer to treat each new toy as as an abstract and unmotivated opportunity. If only the good people like ourselves would get involved early, we can surely teach everyone else to use it ethically!

You know what? I could quote every single line. Just go read the whole thing. Please.