An appeal to companies doing AI

> Distressingly, the people most passionate about AI often express a not-so-subtle disdain for humanity.

I have noticed something similar: those who are ultra-passionate about AI are often Extremely Online, and it seems like their values tilt too far away from humanity for my taste. The use of AI is treated almost as an end in and of itself, which perpetuates a maximalist AI vision. This is also probably why they give off this weird vibe of having their personality outsourced.

> But there is a feedback loop: If you change the incentive structures, people’s behaviors will certainly change, but subsequently so, too, will those incentive structures.

This is a good point, and somewhat subtle too. Something that worries me is the acceleration of the feedback loop. The Internet, social media, smartphones, and now generative AI are all things that changed how information is generated, consumed and distributed, and changing that affects the incentive structures and behaviors of the people interacting with that information.

But information is spread increasingly faster, in higher amounts and with higher noise, and so the incentives landscape keeps shifting continuously to keep up, without giving people time to adapt and develop immunity against the viral/parasitic memes that each landscape births.

And so the (meta)game keeps changing under our feet, increasingly accelerating towards chaos or, more worryingly, meta-stable ideologies that can survive the continuous bombardment of an adversarial memetic environment. I say worryingly, because most of those ideologies have to be, by definition, totalizing and highly hostile to anything outside of them.

So yeah, interesting times.

Tech companies, at least those who weren't founded on AI, have a significant number of people internally who have exactly the opinions in this blog post.

Outside of the tech ecosystem, most people I encounter either don't care about AI, or are vaguely positive (and using it to write emails/etc). There are exceptions, writers and artists for example, but they're in the minority. To be clear, I'm also not seeing "normal" people raving about it much either, but most people really do not share the opinion of this post.

I realise it may not seem like it, but most big tech companies are not designing for high earning valley software engineers because they are not a big market. They're designing for the world, and the world hasn't made its mind up about AI yet.

I've tolerated AI autocomplete in VSCode but I am a bee's dick away from turning it off, because it so often generates a huge chunk of code that is ALMOST correct, and determining where it is wrong is as much a chore as writing it would have been. But, it's like I've got a junior-coder sidekick who doesn't take any feedback. Not great.

> What I found most interesting is that the multiple choice options included a lot of “I found the Terminator movies scary”, “I read too much Ray Kurzweil”, and/or “I am or was a SF Bay Area rationalist” undertones, but actual ethical objections were strangely absent.

After reading Sarah Wynn-Williams' book and see the current state of democracy in the USA and some European countries (apart from the fact that democracies are too slow to regulate anyway), I see little hope for the future.

> I do not actually believe “The Singularity” is a realistic threat due to every system that exhibits exponential growth encountering carrying capacities, which converts it into an S-curve.

Depending on the parameters of the curve, an S-curve may be effectively the same as an exponential curve. For instance if the IQ of AIs reach a plateau of 500 rather than exponentially increasing to infinity, we may not be around to see the plateau.

It's not even clear "superintelligence" is a meaningful concept. It could be that the broad conflicts in our society are largely intractable and arbitrary, in which case "superintelligence" can do little but complain about the contradictions its passed (something I suspect many on this forum can understand).

Perhaps the calculator is as close as we'll ever get to "superintelligence".

My only concern about AI is that it will get stuck again for another few decades and I'll become elderly before I see what's next.

20 years ago my dream was to see a nimble robot running up the mountain path live. I hope this event is not another 20 years away. Future comes so horribly slowly.

I'm very much into AI, in terms of judicious and sprinkled usage across disciplines -- but I'm a computer guy, I produce and support a lot of code, art, audio, video, social, multimedia etc content, and I've incorporated bits and pieces of AI here and there to achieve a vision or get work done (LLMs, diffusion, etc) -- my main issue with it, like anything technical, is that the mainstream interpretation / instinct by the average bozo is that the machine will do my work for me, and nothing could be further from the truth.

There's a lot of anti-AI sentiment from the mainstream, but I'm noticing the pro-AI mainstream sentiment comes from people who are either technically-minded grifters looking to deploy automated solutions to snake $$$ from people's pockets, or lazy / disengaged worker drones who just want the computer to do their middling work for them. And it will, up to a point, where your "work" plateaus into a mess of predictable, non-novel banality lol, unless you invest the time to master the tool (which, for what it's worth, isn't like introducing the toothed saw, it's more like a Dremel or a SAWZALL that have specific purposes but casual users won't ever master them)

Buying a digital ELPH in 2001 didn't make you a photographer unless you were a photographer with an open mind. Squarespace doesn't make you a web designer unless you've studied the system and understand the tradeoffs. AWS doesn't solve infrastructure unless you learn how to architect a solution that works for your use-case. AI, by the same property, is shit until you research how it works, experiment with solutions, and find novel workflows to get something out the other end that's new, fresh and exciting.

Companies just rolling "AI" into their products aren't gonna win over customers and users unless they use the tool to deliver something of exceedingly-needed value. If it's a short-term grift or "hail mary", good luck! You'll need it!

I actually think that many of the concerns the author doesn’t have are the more concerning ones and some of the concerns they do have are more likely to be not much of a big deal.

The privacy concern I find to be particularly overstated. This is an identical concern to ones that have existed before AI ever entered the fray. Anytime you send data to a system that someone else controls you run those exact same risks. I also think there’s an overstated fear that an app focused on private data (something similar to Signal) would just add some kind of AI functionality one day out of the blue and suddenly ship your data off to a hive mind.

Any app that is willing to cross that line already has done so (e.g., Facebook).

It also seems to be technologically simple to perform a lot of AI tasks without compromising privacy. E.g., chips with local-first AI computational ability are reaching consumer level devices. Even the much-maligned Windows Recall feature specifically emphasizes how it never sends information to Microsoft servers nor processes data in the cloud.

> why we dislike AI

For some completely unspecified group of “we”. At least the post itself says “why I personally dislike AI”.

I followed the HN-targeted link at the top ("It gets exhausting") and Oh. My. Good. Lord:

>> Something Awful had a particular response to furries. After creating a subforum specifically for furries to post in, everyone who used it was marked with a custom yellow star avatar, then banned.

https://archive.md/fX8Jo

I used to think the early internet was less shite than the current internet. Maybe I've been mollycoddled, wrapped in the HN blanket where behaviour like that would never be condoned. Maybe I need to see more social media. But, really? Yellow stars? OK wow.

Ahead of the inevitable Luddite comments. Here’s your daily reminder that the Luddites were not just technophobes, but were in fact artisans who were concerned about the leverage technology was providing capital to suppress worker rights while eroding the quality of the products. This tension should resonate with us.

https://www.history.com/articles/industrial-revolution-luddi...

The AI conversation tends to split folks along similar “passionate engineer craftsman” vs. “temporarily embarrassed billionaire” lines.

> If you do not understand people, you will fail to understand the harms that AI will unleash on the world.

Case in point: the official white house social media account regularly posts low-effort AI meme propaganda.

Here’s the thing…as long as we live in a capitalist society, money will be put in front of people because the very essence of capitalism is to turn a higher profit and anything that impedes that is evolved away. So the question to ask is whether we will remain a capitalist society. If you believe the answer is yes, the only thing to do is adapt, whether you like it or not. Resisting the change will only put you behind the curve. I’m not proclaiming a stance on AI here, but I think it’s prudent to be a realist.

[flagged]

> I’m concerned about the kind of antisocial behaviors that AI will enable.

> Coordinated inauthentic behavior

> Misinformation

> Nonconsensual pornography

> Displacing entire industries without a viable replacement for their income

The first three of these existed and occurred before the arrival of AI. Perhaps AI makes doing the first 3 easier. If there are not laws governing the first three post-AI, do we need laws governing them? If so, what do those look like?

As for "displacing entire industries without a viable replacement for their income" - yea, as a civilization we need to retrain and reeducate those whose livelihoods are displaced by automation. This too has been true forever...

Disabling ai at this point is like using your app in off line mode. Not a lots of people want to build for that

He is misplacing his concerns. He and the ones like him, to fulfill the "we" there.

Oh, in US, for US citizens, intelligence agencies can't do dragnet surveillance. But it seem to be perfectly OK that it does that for the rest of the world.

It enables people to do non-consensual pornography. It may have democratized the realistic video part of it, but already was available the fan fiction part. Is the problem the democratization or the ability on who have enough budget or sponsored agenda for it? At some moment you have to cut somewhere and define where is the realistic line.

Same goes for misinformation, what is wrong is the democratization and not that the people with enough resources can do it?

About displacing industries, it depends, but was already a big abuse of some of those industries to people. Some will adapt. Some will become obsolete as it happened with the industries they replaced in their own turn.

AI is a tool. And as any tool, it empowers people using it, for good and bad. Is the people that you keep giving power the main ones misusing them. Those are the elephants in the room that you refuse to see.