HilliTech's Newsletter: AI is a nuanced problem

Siri AI might be the closest thing we have to ethical AI

You would think with how extreme things have become in politics and social problems that we'd have a little more grace elsewhere. That increasingly hasn't been the case when discussing tech, especially AI.

The AI debate is stupid

Let's get this out of the way at the top. Artificial Intelligence is a bad name for an interesting technology that exists today because of theft and environmental destruction at a scale we've never seen before. It's an easy technology to hate, and with good reason!

It's also difficult to ignore the technological advances enabled by sufficiently controlled systems. AI for the sake of AI is stupid. Making your entire personality into an AI grifter is idiotic and a waste of a human life.

I'm not a fan of the technology the same way I'm not a fan of hammers. I'm also aware of the total destruction of the internet as we know it thanks to companies like Google and OpenAI chasing fake endpoints like sentience in the name of the future of humanity.

It's all bullshit. The grifting, the CEOs proclaiming it is the future, and even the users claiming life-changing experiences when using the plagiarism machines.

That's the problem though. We've been trying to build this technology so long that the idea behind Large Language Models, (LLMs) and thinking computers goes back to the 1980s. It isn't a new idea, it's just one we can actually execute with today's computer platforms.

How we arrived at the technology is also problematic. Scraping every part of the web, even destroying books, just to train a bot to be a bit more capable of referencing data given an input. It's horrific and should be viewed as a major casualty of the modern era.

Even the way companies talk about AI is batshit. They want you to think of their computer programs the same way you do humans with anthropomorphized language, which has lead to psychosis-like behavior in some people.

Don't lose the nuance

All that said, I think it is a mistake to dismiss AI as a useless, terrible technology on its face. We could have arrived at it later without all of the theft and destruction, but the billionaires couldn't stand going a more normal route.

The technology is here, it's not going away, and I believe that even though much of it exists through theft and environmental destruction, it isn't something we can expect humanity to put away. Much of the world exists today through destructive and hateful practices, but we're not trying to undo existence because our modern society exists because of murder, racism, slavery, and colonialism.

This isn't an excuse. I think we should be very aware of how AI was built just the same way we need to be aware of how the United States was built. That awareness should give you pause, and make you consider exactly how you might approach the technology and the world at large.

Hell, at a minimum, the knowledge that much of AI and the companies that built it are corrupt entities run by man babies with delusions of grandeur should, at the least, make you selective in which AI you use.

If I hear you've tested Grok for any reason, that's an instant loss of nearly all respect that I'm not sure you can recover from, for example. ChatGPT is the Meta of AI, while Claude is more like the Google. Gemini is fine but really just use Siri AI if you have to use anything of that nature.

I personally don't have much need for AI. I avoided ever installing Gemini, ChatGPT, Claude, or any of them on my devices. When Apple revealed its first very limited version of Apple Intelligence, I gave it a shot considering it was on-device and Apple at least attempted ethical sourcing. (Yes, AppleBot scraped the web for AI training before Apple told everyone it was doing so, that's inexcusable)

Apple paid publishers for access to books and articles for training. Even when it crawled the internet for data, it excluded anything deemed personal information or copyrighted data. Even today, Apple is seeking to partner with publishers for delivering news content when someone makes a request via Siri AI.

No AI is ethical really, but I do think Apple's is the closest to what an ethical AI might look like. Rebuilding the Apple Foundation Models by distilling a Gemini Frontier Model does complicate this discussion, but if we're going to talk about using AI, I doubt you can find a less harmful one than Siri AI.

I am talking about this because I understand the "no AI ever" people and their position. I even kind of understand the pro AI crowd, even if I think they're too far on that extreme too.

The point is, extremes are always bad. Nuance in any discussion is important, which brings me to a tangential topic this week.

The Nazi bar is full of Nazis

I don't want to get into specifics, you all reading this know what happened, and if you don't, then be happy in your bliss. I'm going to stick to generalities when referring to the events.

First off, the elephant in the room as I discuss this topic: Yes, AppleInsider still has an account on X. No, that wasn't my decision, and if it were up to me, we'd have deleted it when I deleted my account. I have also repeatedly made my feelings known on the matter.

The thing is, I don't hate people for staying on X even if I believe they should delete their accounts and never think about it again. The reasons for staying are complex, from wanting to ensure no one steals your username to having a unique user base that isn't present anywhere else. Running a business in the post-SEO world is tough, and not everyone can make decisions based on principles alone.

Elon Musk ruined a perfectly terrible and flawed platform and fast-forwarded it to its seemingly natural conclusion. It's basically 4Chan with more Nazis.

The thing is, staying, and even engaging with the platform, is problematic on its own, but doesn't necessarily paint that person as evil. They're attempting to tolerate the Nazis in the bar in the hopes that one day the Nazis will get bored and leave. It's not going to happen, but I can't really blame them for being hopelessly optimistic.

Then there's the various exoduses that took place after Elon's acquisition. I quite X, joined Mastodon, then made a Bluesky account as it became public. Bluesky is now my main social media presence. I've made it known, repeatedly, how I feel about X and hope that my discussions around it, and helping raise awareness about the pro-fascist platform filled with hateful content and a chatbot willing to make CSAM has been part of it.

If you can stomach your content being there, like Apple and its executives seemingly can, go for it I guess. I will happily not participate in the platform and pray for its death, or at the least, an ownership transfer that drives out the Nazis. It won't happen, but one can wish.

Blocking is always an option

The problem we've encountered this week is two-fold. My earlier AI discussion alluded to this, but the extremes between the pro-AI crowd and anti-AI crowd has created a problem. The pro-AI crowd doesn't seem to understand how blocking works on Bluesky, so when they get insulted, they cry and run back to their friends at the Nazi bar for sympathy.

"X never treated me this way for liking AI, Bluesky doesn't understand why my use of the plagiarism machine is okay." – some crybaby, probably.

The thing is, I do think the anti-AI folks are too hard-lined. They have good reason to be, but we really should, as a society, do better to distinguish between a grifter selling bullshit and a human trying to utilize AI in an interesting way.

Though, I'll throw in a big fuck you to anyone promoting AI as a replacement for humans instead of an augment that can help improve human productivity.

AI isn't as good as pro-AI people think, it's not as bad as anti-AI people think.

So, of course when a beloved website in the tech community (and several podcasts associated with editors on that website) veer sharply into "All AI all the time," there is going to be a negative reaction. Combine that with AI fatigue and the death of SEO and yeah, any publication is going to feel the pain.

However, leaving X and then running back to it because that's where the AI sympathizers are doesn't seem to be a smart move. You're not fixing your declining metrics, you're replacing a smaller more niche audience with morals and standards with unwashed masses that don't give a shit about you that also happens to be littered with right-wing extremists and Nazis.

I wish them the best of luck, truly. Independent media is struggling as it is and I hope they learn whatever lessons they need to learn from this move.

The worst way to handle this

I've barely listened to the recent podcasts because I skip entire chapters about how Claude helped automate some email thing. I'm sure someone finds that interesting, but to me, it all sounds like reading radio instructions to get a better understanding of how the volume knob works.

Shortcuts and automations are great! Chaining events together, and even using AI like Siri AI to make that better is interesting. You can talk about these things in interesting ways. But, somehow, they got drug into this continuous circle around the toilet bowl where they're exclaiming how amazing some new oddly named AI version is, token use, and do so breathlessly about achieving something that's so mundane I don't understand the hype.

Podcast or article, they never take the time to discuss the moral or environmental aspects of their AI use either. It's just all about spending hundreds of dollars on tokens to improve their productivity by 8% using an autocomplete engine dressed as a magic genie. It's all curtains and illusions – when you peek back there it's just technology, not some god creature.

Pair that sharp turn into worshiping this thing that should have been a footnote in a release cycle in 2022 that instead became a psychosis-inducing phenomenon with declining reader numbers, and you get the end result. Desperation.

They've run back to X, the Nazi bar, because some folks there are loud and proud about AI and spend an awful lot of time talking about how small, weak, and closed off Bluesky is.

To add insult to injury, they did so without alerting their contributors. I wish those departing the best and hope they can continue to create amazing things away from the Nazi bar.

Of course, the letter that followed attempting to explain things was hollow. Perhaps it was written with AI, I have no way of knowing at this point. But it attempted to justify why they went back against everything they stood for. That big show about creating Mastodon servers and seeking to be a part of the open web was all for the clout.

Clout chasers will always chase the potential clout, even if it's in a Nazi bar.

Don't argue with people on the internet

What's worse is this has created a bit of a schism in the tech community that I love. People I respect coming out to insult and defend one decision, while others publicly state how disappointed they are. It's a sad state of things.

I even saw some people break out into fights, some defending X and trying to distance it from Elon Musk, etc. Just a whole lot of cope.

I write all this to say I'm disappointed and hope we as a tech community can do better. If we can support each other and lift each other up, that's what matters.

Even if only 40 or 50 people get this newsletter in their email inbox, I'm happy to do so, because I did it without needing to rely on fascists.

I'll end with this thought:

If your ideas and opinions can only find support in a community filled with Nazis, its your ideas that are the problem, not the community that rejected you. Instead of pivoting to chase Nazi clout, perhaps pivot to topics that will make readers more interested in what you have to offer.

If going all in on AI means needing X to keep your business afloat, then maybe going all in on AI is the problem. But somehow, they didn't think of that.

Some stories

Hey, this section used to be at the top! Yes, that's right, I've moved it to the bottom. I figured the newsletter should start with, you know, the newsletter.

A Maine coon mix cat on a hutch with its paw stretched out — Marble doesn't have any use for AI or Nazis