Even for people who generally like the function of AI (which seem to be fairly rare here) the absolutely obscene climate impact and implications for peopes jobs and livelihoods, privacy breaches, and general internet enshittification is surely reason enough to be against it.
The jobs thing i don’t understand, its the distribution of productivity gains that’s the issue, why we keep voting for the same politicians ensuring it goes to the wealthy is the real mystery.
Oh, I absolutely agree. But currently, the people in charge of making those decisions have demonstrated moral bankruptcy and will absolutely ensure the productivity gains funnel to the top. Until that changes, AI impact on jobs will likely be devastating.
And I’m all for changing it. It’s just going to be a long and/or violent process.
Productivity gains are not across the board, and is a subject of scrutiny and debate.
But what AI really has done is basically redistributed American wealth to a smaller group of people, and therefore a smaller pool for the US politicians to focus on satisfying. If there is an AI bubble pop, what market watchers suspect is there’s actually no other American sector to mitigate what is otherwise a recession.
It has its uses but it feels like more of a 10-20% productivity boost when used effectively, not the 500%, “lets have openclaw replace my whole company!” kind of BS being pushed by AI companies.
If it is a productivity boost for you, it is at the cost of someone else who will have to proofread and test everything you do. LLMs (and genAI) are useless.
It’s no more work than proofreading any other code I write. Sounds like someone just slopped out code with an LLM and didn’t do the due diligence of checking it themselves. Using an LLM doesn’t mean no work. I think that’s when people get in trouble.
That I why I like small, specialized, locally hosted AI. Runs acceptably fast and quite on my gaming PC, it’s private, and I can give it knowledge is small doses in specific topics and projects.
Which model do you use and what are your specs? I ran a couple using an RTX5060 with 16gb and it’s too slow to be usable for larger models while the smaller ones are mostly useless.
I also have a 5060 (ti) with 16GB of RAM. I tend to use GPT-OSS:20B or Qwen3:14B with a context of ~30k. I have custom system prompt for my style of reponse I like on open web ui. That takes up about 14GB of my 16GB VRAM
But yeah it is slower and not as “smart” as the cloud based models, but I think the inconvenience of the speed and having to fact check/test code is worth the privacy and environmental trade offs
Ive had good success on similar hardware (5070 + more ram) with GLM-4.7-Flash, using llama.cpp’s --cpu-moe flag - I can get up to 150k context with it at 20ish tok/sec. I’ve found it to be a lot better for agentic use than GPT-OSS as well, it seems to do a much more in depth reasoning effort, so while it spends more tokens it seems worth it for the end result.
Even for people who generally like the function of AI (which seem to be fairly rare here) the absolutely obscene climate impact and implications for peopes jobs and livelihoods, privacy breaches, and general internet enshittification is surely reason enough to be against it.
The jobs thing i don’t understand, its the distribution of productivity gains that’s the issue, why we keep voting for the same politicians ensuring it goes to the wealthy is the real mystery.
Oh, I absolutely agree. But currently, the people in charge of making those decisions have demonstrated moral bankruptcy and will absolutely ensure the productivity gains funnel to the top. Until that changes, AI impact on jobs will likely be devastating.
And I’m all for changing it. It’s just going to be a long and/or violent process.
Productivity gains are not across the board, and is a subject of scrutiny and debate.
But what AI really has done is basically redistributed American wealth to a smaller group of people, and therefore a smaller pool for the US politicians to focus on satisfying. If there is an AI bubble pop, what market watchers suspect is there’s actually no other American sector to mitigate what is otherwise a recession.
It has its uses but it feels like more of a 10-20% productivity boost when used effectively, not the 500%, “lets have openclaw replace my whole company!” kind of BS being pushed by AI companies.
If it is a productivity boost for you, it is at the cost of someone else who will have to proofread and test everything you do. LLMs (and genAI) are useless.
It’s no more work than proofreading any other code I write. Sounds like someone just slopped out code with an LLM and didn’t do the due diligence of checking it themselves. Using an LLM doesn’t mean no work. I think that’s when people get in trouble.
That I why I like small, specialized, locally hosted AI. Runs acceptably fast and quite on my gaming PC, it’s private, and I can give it knowledge is small doses in specific topics and projects.
Which model do you use and what are your specs? I ran a couple using an RTX5060 with 16gb and it’s too slow to be usable for larger models while the smaller ones are mostly useless.
I also have a 5060 (ti) with 16GB of RAM. I tend to use GPT-OSS:20B or Qwen3:14B with a context of ~30k. I have custom system prompt for my style of reponse I like on open web ui. That takes up about 14GB of my 16GB VRAM
But yeah it is slower and not as “smart” as the cloud based models, but I think the inconvenience of the speed and having to fact check/test code is worth the privacy and environmental trade offs
Ive had good success on similar hardware (5070 + more ram) with GLM-4.7-Flash, using llama.cpp’s
--cpu-moeflag - I can get up to 150k context with it at 20ish tok/sec. I’ve found it to be a lot better for agentic use than GPT-OSS as well, it seems to do a much more in depth reasoning effort, so while it spends more tokens it seems worth it for the end result.