Basically, the actual problem is with the capitalist system of relations and how automation is inevitably applied by the capitalists to harm workers.

Same, AI haters really don’t realize just how far this tech has come in just the past year. I’ve had to work on frontend Js projects at work, and I’ve been lucky enough to avoid Js for most of my career. I have lots of experience programming, and I know how to structure applications, but I’m not familiar with Js stack, libraries, and syntax quirks. LLMs help me paper over all that and use it like any other language I’m already well versed in. Without LLMs, I would’ve had to spent literally months ramping up on Js ecosystem to do the work I’m doing now.

There are countless constructive and legitimate uses for AI/LLMs. This just another form of automation, and it works well for many tasks already. Other people in the thread have already given a ton of examples, so I don’t really have to reiterate them here. Meanwhile, practically all the criticisms of this tech actually boil down to problems with how it’s applied under capitalism.

O7

Oh you can scan it, but as I recall they then force you to put your phone number in to finish the process.

That’s what google’s been using to lockout non official Android forks like GrapheneOS. You can click on the eye icon at the bottom to get the regular captcha though… for now.

Yeah, as I recall it was just codex repackaged, but there is also the aspect of the tooling itself that the model interacts with which they massively fumbled. People will still use codex on its own, but nobody wants to ever use copilot.

a very different culture here from reddit, and didn’t know that either about Yusupov

yeah can chat there

Not putting it into atmosphere is precisely how capture at the source works. Carbon is the input for the industrial process here. Every bit that isn’t captured is wasted.

Oh that’s the magic of tools like opencode, you run it in a folder and it acts as a harness for the model where it can interact with the filesystem. You could do the same with an IDE as well, making your own agentic harness is actually pretty straight forward. So you could make a plugin that talks to, say, ollama https://ampcode.com/notes/how-to-build-an-agent

I find these kinds of projects are neat, but if I’m being honest, I tend to just keep plain markdown files as well. The only thing I find that’s missing with that is searchability. Once you get enough files, it can get unwieldy. Although, I’ve been playing around with just using a local model lately as the interface. You can throw opencode at a folder with the files, and even a small model can find stuff fairly competently there.

kinda of hilarious that MS has a stake in both of the worst commercial LLM offerings

Same, I’m basically tone deaf and just assumed that Mandarin wouldn’t be accessible to me. But then I finally decided to give it a go, and turned out to not be much of an issue at all. I also find that it’s easier to remember the tones in a context of a sentence. It’s a lot like when you put an accent on different words when you speak English, so you can just memorize the cadence of the sentence, and you’ll start learning the tones implicitly.

That’s the approach I took as well. For the first year, I just stuck with pinyin, and then once I got comfortable enough conversationally, I started making some effort to learn the characters. I find drawing them out really helps you memorize them so you can recognize them later, so even if you’re going to write using pinyin, practising writing is still useful. And it does get easier, because there is a fixed set of symbols that all the characters are composed of. So, once you learn the first batch, it only gets easier from there. But yeah, putting that off is the correct call.

there is that too, but greed and corruptions are big factors to be sure

I suspect they know it doesn’t work, but people line their pockets and then sell it to the gullible public as doing something about the climate crisis.

There have been greenwash projects as well. The numbers are hilarious, like they remove 50 tons a year. 🤣

https://www.deepskyclimate.com/blog/canadas-deep-sky-to-pilot-dac-unit-from-dutch-startup-carbyon

I’m learning Mandarin right now as well, and in some ways it’s much easier than other languages I’ve learned. Grammar is pretty straightforward, there’s not conjugation or tenses to worry about either. If you learn a word, you just have to learn it once, there aren’t any variations. The whole thing with tones is largely overstated I find. Even if you get the tones wrong, people will understand you from the context. Where it gets trickier is with writing because the character based system is genuinely more difficult to learn than an alphabet. The characters are basically words written in two dimensions instead of one. Most are composed of subcharacters of which there is a common set of. But the upside here is that once you learn them, reading is a lot faster because each one is basically like an icon. So, you can scan through text a lot easier than with words all written out left to right.

There are two types of carbon capture. One is at the source, as is happening here, which is actually effective. The other is trying to scrub carbon from the atmosphere where the math does not work. If carbon capture starts producing useful outputs then there is a direct economic incentive to start doing it at the source. Incidentally, there are other similar processes that are also very promising https://www.foodtimes.eu/food-system/co2-upcycling-feed-protein-china/