sailor_sega_saturn

The Generative AI hype at my job has reached a fever pitch in recent months and this is as good a place to rant about it as any.

Practically every conversation and project is about AI in some way. AI "tools" are being pushed relentlessly. Some of my coworkers are terrified of AI taking their jobs (despite the fact that the code writing tooling is annoying at best). Generative AI is integrated with everything it can be integrated with, and then some. One person I talked to admitted to using a chatbot to write performance reviews for their peers. Almost everyone at my job who I'm not close friends with is approximately 300% more annoying to talk to than a year ago.

Normally if there's some new industry direction we're chasing people are almost bored about it. Like "oh dang I guess we have to mobile better". Or "oh gee isn't implementing cloud stuff fun whoop-dee-doo". But with AI it's more like everyone is freaking out. I think techies are susceptible to this somehow -- like despite not really working that way at all it feels close to sci-fi AI. So a certain class of nerd can trick themselves into thinking the statistically likely text generator is actually thinking. This can't last forever. People will burn themselves out eventually. But I have no idea when things will change.

Basically I should have gone into an industry with more arts majors and less CS majors sigh.

Zuck, who definitely knows how human friendships work, thinks AI can be your friend: https://bsky.app/profile/drewharwell.com/post/3lo4foide3s2g (someone probably already posted this interview here before but I wasn't paying attention so if so here it is again)

In completely unrelated news: dealing with voices in your head can be hard, but with AI you can deal with voices outside of your head too! https://www.rollingstone.com/culture/culture-features/ai-spiritual-delusions-destroying-human-relationships-1235330175/

(No judgement. Having had a mental breakdown a long long time ago, I can't imagine what it would have been like to also have had access to a sycophantic chat-bot at the same time.)

Indeed.

First, Chrome won the browser war fair and square by building a better surfboard for the internet. This wasn't some opportune acquisition. This was the result of grand investments, great technical prowess, and markets doing what they're supposed to do: rewarding the best.

Lots of credit given to 👼🎺 Free Market Capitalism 👼🎺, zero credit given to open web standards, open source contributions, or the fact that the codebase has a lineage going back to 1997 KDE code.

This is really setting students up for failure isn't it?

CS 100 and 200 level problems are well represented online enough that "vibe coding" will just write out the entire solution. But then the student may rely on it and either burn out in the harder courses or accidentally turn in something that oops was copy pasted directly from a previous year student off of github.

I agree. I spent more time than I'd like to admit trying to understand Yudkowsky's posts about newcomb boxes back in the day so my two cents:

The digital clones bit also means it's not an argument based on altruism, but one based on fear. After all if a future evil AI uses sci-fi powers to run the universe backwards to the point where I'm writing this comment and copy pastes me into a bazillion torture dimensions then, subjectively, it's like I roll a dice and:

1. live a long and happy life with probability very close to zero (yay I am the original)
2. Instantly get teleported to the torture planet with probability very close to one (oh no I got copy pasted)

Like a twisted version of the Sleeping Beauty Problem.

Edit: despite submitting the comment I was not teleported to the torture dimension. Updating my priors.

Along the same lines of LLMs ruining language stuff: I just learned the acronym MTPE (Machine Translation Post Edit) in the context of excuses to pay translators less and thanks I hate it.

Can't avoid slop reading translated books, can't learn the source language without dodging slop in learning tools left and right. It's the microplastics of the internet age.

Anyway my duolingo account is no more, I have better resources for learning German anyway.

Days since last "novel" prompt injection attack that I first saw on social media months and months ago: zero

Against my better judgement I typed steve.ai into my browser and yep. It's an AI product.

frodo.ai on the other hand is currently domain parked. It could be yours for the low low price of $43,911

https://www.latimes.com/california/story/2025-04-23/state-bar-of-california-used-ai-for-exam-questions

When measured for reliability, the State Bar told The Times, the combined scored multiple-choice questions from all sources — including AI — performed “above the psychometric target of 0.80.”

"I dunno why you guys are complaining, we measured our exam to be 80% accurate!"

I think this article fits here as how this fear mongering reaches government: https://futurism.com/google-ceo-congress-electricity-ai-superintelligence

Eric Schmidt tells the US government that unfathomable amounts of power need to be spun up ASAP because if the US doesn't make robot god first then:

If China comes to superintelligence first, it changes the dynamic of power globally, in ways that we have no way of understanding or predicting

Dedicating the US's entire economic output to scaling corporate LLMs and filling the air with smog would at least have somewhat predictable consequences I suppose.