That sort of implies they have been dethroned in some way, but the opposite seems to be true. They are eating OpenAI’s lunch. Claude is widely considered the best development AI, and they just passed OpenAI in business customers.

That said, they do post the most ridiculous blogs at times. Actually, I have yet to see a blog post by them that wasn’t complete bullshit, but that’s probably just selection bias. I assume they have some employees who aren’t complete morons.

I used Brave for a bit after it came out. All the built-in crypto bullshit wasn’t anything I ever touched. I can’t remember if I quit some time before Chrome announced the manifest-v3, or if that made me swear off all Chromium browsers for good.

Anyway, also apparently the creators are huge assholes, so there is that, but as I never had any intent to support them financially I just thought I’d use a privacy-focused browser. Meh. Firefox has all the plugins a person could ever want.

I mean, use whatever browser you like. Lemmy is cool and basically right about a lot of things, but fuck the zeitgeist and make up your own mind about shit. I went a different way, but forge your own path.

I, too, installed the speed drain.

Novel.ai comes to mind. It’s made for writing. It won’t write for you, but it’s better in some ways than the more general purpose AIs you named. I mean, it still needs editing and human input.

The interface is made for that, too. You can just edit the story like you can’t in typical chatbot interfaces. Pacing and omniscience are still a problem, but it does follow specific instructions better. You have lorebooks for specific details. And a very workable amount of tokens

Having stared at my wife’s vag as she gave birth, you don’t want to see that. It’s amazing watching your child be born, but there is… it’s not a vagina as you know it. If I wasn’t able to compartmentalize and I saw that image in my head every time I had sex with my wife, I’d probably never have sex again.

If someone offered to show me a video of my birth… I’d probably pass. The creature emerging from that portal to hell has no relation to what any is us became. Also, “I landed on a little poo,” isn’t something any of us wants to know.

Somewhere out there are fetishists claiming the exception to the rule but… I’d think twice.

All I can attest to is in Virginia, yes, in Michigan, no.

I’m Spartacus.

… Except I’m not on X or Reddit.

I have a feeling there’s a fine line between not hard enough and everyone spending $2k per month on tokens.

Those violent bastards. Someone oughta kill them…

I run this setup with 36GB (32+4). Local LLMs can be really effective BUT you are constrained by context size in a way you aren’t on cloud services.

Cline supports running a local model through lmstudio but my experience feeding it any significant tasks is it just can’t handle reading and holding the contexts to build components for enterprise scale applications.

I use Claude to write a lot of utility one-off scripts. With a maximum window of 1M tokens I can hit 30+% context just writing Python scripts. API contracts, development standards, existing reusable modules, and sometimes reading the code/documentation of the services I’m going to be calling.

My MacBook can’t handle 300k token contexts. 30k seems doable. I should see how it handles my utility script folder…

Anyway that’s still no Claude but if you need a cheaper model and you can afford for developers to spend time on it before ultimately deciding they need to spend for Claude or Codex or Gemini, then rubbing a local model on a beefy MacBook is 100% an option.

Stepping up from there to building a locally hosted LLM is probably the worst of all worlds. It will be a beefy CapEx, prone to saturation by all the users, and you will most likely still have to punt the hardest jobs to cloud AI. It can certainly be done and done well, but the best example I know runs on $250-500k worth of hardware (to service a pretty big number of users to be fair).

Everything is Turing complete these days. Someone out there is devising a test to prove Legos are Turing complete.

I suggested I’d probably be interested, and I am. I have to explore my thoughts on profit motive and owner class vs. working class. We don’t see entirely eye to eye, but there are a lot of layers to unpack and I’ve been at this reply off and on for some time now. I drafted a much more detailed response, but it’s undergone several full-draft rewrites, and I think I need to just do my thinking offline and this thread will be long dead before I come back around to it.

Anyway, I do appreciate your response.

I said vaguely. I’m not a communist or Marxist and recognize the limits of my understanding.

The difference between owning the means of production and sharing ownership with investors feels meaningful but not diametrically opposed. Without the investors, the workers would STILL have to weigh their ownership stake against working conditions and determine what is in their best interest.

I agree that bonuses being outside the workers control makes them not great overall, however in this case the bonus isn’t cash, but a stake in the company which again ties the payment to future performance. Not in a way the workers can directly control, but there is always going to be friction between what workers deserve to be paid for their work and what customers are willing to pay for the product. Ordinarily that friction serves to make investors fabulously rich and the workers largely get exploited.

Anyway, I said vaguely and I stand by it. If you want to go in depth on your views of capitalism and Marxism, I promise to read and likely be fascinated. But I think you read that with a lot more intent than I originally meant to impart. I probably should’ve just left that bit out, knowing Lemmy users.

I wouldn’t have expected it to. There are both cynical and earnest interpretations of that. Cynically, they are avoiding pissing off AIPAC. Earnestly, the American electorate as a whole isn’t moved by the plight of Palestinians.

I’ll leave everyone to their own choice of interpretations, but you don’t have to morally agree with the latter to feel it is true. Both could be true.

Here and I was reading it as “They are getting $400k, per Chip, employee of Samsung.”

Seems like a good thing. Aligning workers and shareholders seems reasonable. And it even seems vaguely lefty, with the workers having a bigger stake in the means of production and all. Valuation might go down, and if there is a global recession most of our 401ks will go down, too. I don’t see any disproportionate downside.

You either meet people where they are, you find a strategy to move the people, or you lose. None of that is on the people. Your blame changes nothing. It’s wishful thinking — you go into the election with the voters you have, not the voters you wish you had.

The voters are basically the same every election. Yeah they evolve over time but not quickly. You can’t blame populace for being exactly who they are.

They weren’t enthusiastic about Kamala, and if you ask me it has everything to do with the economy and approximately fuck all to do with anything else. But you can’t fault the people any more than you can fault the tide when there is a flood.

They are going to enshittify so hard trying to get their money back… it’s too bad, too. I use ChatGPT a fair bit for various things. But eras end and this could herald the end of it being useful to me. At some point I was always just going to set up my own local AI on a server here at home. Hopefully hardware prices can recover a bit before I feel compelled to move on from ChatGPT.

Hey well at least typos serve as a marker that everything is human written. You know, for all the comments about how AI is fancy autocorrect, why the hell is autocorrect still such fucking garbage??? Just do the things with the damn thing already!