▲ 760 ▼ Anthropic confirms new Claude Code limits to prevent game cloning — users report Opus 5.5 is also actively blocking reverse engineering tasks (www.tomshardware.com) submitted 1 week ago by inclementimmigrant@lemmy.world to c/technology@lemmy.world 174 comments fedilink hide all child comments My how fast the AI tech bros move to protect corporations but don't move to protect the average coder, artists, writers.
[–] TiredDinoByte@lemmy.today 14 points 1 week ago (8 children) I think you are right, but that seems VERY unenforceable. permalink fedilink source parent hideshow 8 child comments replies: [–] willington@lemmy.dbzer0.com 14 points 1 week ago Successful enforcement doesn't look like catching every pivate LLM owner. permalink fedilink source parent [–] heartSagan5@lemmy.zip 5 points 1 week ago (6 children) Except the people I hear running “local models” report a lot of hardware, like a half dozen graphics cards. The common person is screwed, they have a smartphone. permalink fedilink source parent hideshow 6 child comments replies: [–] SupraMario@lemmy.world 7 points 1 week ago (5 children) Two or three years ago sure. Now...not even close. Tons of models run on basically shit now. Especially with MOEs. I run new lightweight models on an i9 and 64gb of ddr4 ram with a 4gb quadro card...in a laptop from 19'. I think this is why a lot of these tech bros are saying silly shit about suggesting people need to rent from them. That's their entire shtick. Otherwise they got nothing else. permalink fedilink source parent hideshow 5 child comments replies: [–] danielfm123@lemmy.zip 1 point 1 week ago (4 children) Got some references and keywords for research? permalink fedilink source parent hideshow 4 child comments replies: [–] SupraMario@lemmy.world 10 points 1 week ago (1 child) https://youtu.be/8F_5pdcD3HY This guy does a great brain dump on how little is needed to run LLMs on basically old junk. permalink fedilink source parent hideshow 1 child comment replies: [–] danielfm123@lemmy.zip 1 point 1 week ago Kudos permalink fedilink source parent [–] boonhet@sopuli.xyz 2 points 1 week ago (1 child) If you've got at least 8 gigs of VRAM and at least 32 gigs of RAM you can run some pretty beefy Qwen models using Strata, far more than you normally could. Upgrade that to 12 or 16 and 64 gigs respectively and you can run a pretty decent quant of Qwen 3.8 flash next. It's a LOT bigger than the usual 27 and 35 billion models people run locally that themselves require pretty beefy hardware. Strata is doing some magic by deciding which experts stay in VRAM vs which ones are in RAM vs which ones are so rarely used that they can just be read from the SSD as needed (though it's best if you don't have to result to the SSD as a backup option at all) permalink fedilink source parent hideshow 1 child comment replies: [–] danielfm123@lemmy.zip 1 point 1 week ago I saw the video, you still need a fast NVMe, I'm still in data SSD... permalink fedilink source parent
[–] willington@lemmy.dbzer0.com 14 points 1 week ago Successful enforcement doesn't look like catching every pivate LLM owner. permalink fedilink source parent
[–] heartSagan5@lemmy.zip 5 points 1 week ago (6 children) Except the people I hear running “local models” report a lot of hardware, like a half dozen graphics cards. The common person is screwed, they have a smartphone. permalink fedilink source parent hideshow 6 child comments replies: [–] SupraMario@lemmy.world 7 points 1 week ago (5 children) Two or three years ago sure. Now...not even close. Tons of models run on basically shit now. Especially with MOEs. I run new lightweight models on an i9 and 64gb of ddr4 ram with a 4gb quadro card...in a laptop from 19'. I think this is why a lot of these tech bros are saying silly shit about suggesting people need to rent from them. That's their entire shtick. Otherwise they got nothing else. permalink fedilink source parent hideshow 5 child comments replies: [–] danielfm123@lemmy.zip 1 point 1 week ago (4 children) Got some references and keywords for research? permalink fedilink source parent hideshow 4 child comments replies: [–] SupraMario@lemmy.world 10 points 1 week ago (1 child) https://youtu.be/8F_5pdcD3HY This guy does a great brain dump on how little is needed to run LLMs on basically old junk. permalink fedilink source parent hideshow 1 child comment replies: [–] danielfm123@lemmy.zip 1 point 1 week ago Kudos permalink fedilink source parent [–] boonhet@sopuli.xyz 2 points 1 week ago (1 child) If you've got at least 8 gigs of VRAM and at least 32 gigs of RAM you can run some pretty beefy Qwen models using Strata, far more than you normally could. Upgrade that to 12 or 16 and 64 gigs respectively and you can run a pretty decent quant of Qwen 3.8 flash next. It's a LOT bigger than the usual 27 and 35 billion models people run locally that themselves require pretty beefy hardware. Strata is doing some magic by deciding which experts stay in VRAM vs which ones are in RAM vs which ones are so rarely used that they can just be read from the SSD as needed (though it's best if you don't have to result to the SSD as a backup option at all) permalink fedilink source parent hideshow 1 child comment replies: [–] danielfm123@lemmy.zip 1 point 1 week ago I saw the video, you still need a fast NVMe, I'm still in data SSD... permalink fedilink source parent
[–] SupraMario@lemmy.world 7 points 1 week ago (5 children) Two or three years ago sure. Now...not even close. Tons of models run on basically shit now. Especially with MOEs. I run new lightweight models on an i9 and 64gb of ddr4 ram with a 4gb quadro card...in a laptop from 19'. I think this is why a lot of these tech bros are saying silly shit about suggesting people need to rent from them. That's their entire shtick. Otherwise they got nothing else. permalink fedilink source parent hideshow 5 child comments replies: [–] danielfm123@lemmy.zip 1 point 1 week ago (4 children) Got some references and keywords for research? permalink fedilink source parent hideshow 4 child comments replies: [–] SupraMario@lemmy.world 10 points 1 week ago (1 child) https://youtu.be/8F_5pdcD3HY This guy does a great brain dump on how little is needed to run LLMs on basically old junk. permalink fedilink source parent hideshow 1 child comment replies: [–] danielfm123@lemmy.zip 1 point 1 week ago Kudos permalink fedilink source parent [–] boonhet@sopuli.xyz 2 points 1 week ago (1 child) If you've got at least 8 gigs of VRAM and at least 32 gigs of RAM you can run some pretty beefy Qwen models using Strata, far more than you normally could. Upgrade that to 12 or 16 and 64 gigs respectively and you can run a pretty decent quant of Qwen 3.8 flash next. It's a LOT bigger than the usual 27 and 35 billion models people run locally that themselves require pretty beefy hardware. Strata is doing some magic by deciding which experts stay in VRAM vs which ones are in RAM vs which ones are so rarely used that they can just be read from the SSD as needed (though it's best if you don't have to result to the SSD as a backup option at all) permalink fedilink source parent hideshow 1 child comment replies: [–] danielfm123@lemmy.zip 1 point 1 week ago I saw the video, you still need a fast NVMe, I'm still in data SSD... permalink fedilink source parent
[–] danielfm123@lemmy.zip 1 point 1 week ago (4 children) Got some references and keywords for research? permalink fedilink source parent hideshow 4 child comments replies: [–] SupraMario@lemmy.world 10 points 1 week ago (1 child) https://youtu.be/8F_5pdcD3HY This guy does a great brain dump on how little is needed to run LLMs on basically old junk. permalink fedilink source parent hideshow 1 child comment replies: [–] danielfm123@lemmy.zip 1 point 1 week ago Kudos permalink fedilink source parent [–] boonhet@sopuli.xyz 2 points 1 week ago (1 child) If you've got at least 8 gigs of VRAM and at least 32 gigs of RAM you can run some pretty beefy Qwen models using Strata, far more than you normally could. Upgrade that to 12 or 16 and 64 gigs respectively and you can run a pretty decent quant of Qwen 3.8 flash next. It's a LOT bigger than the usual 27 and 35 billion models people run locally that themselves require pretty beefy hardware. Strata is doing some magic by deciding which experts stay in VRAM vs which ones are in RAM vs which ones are so rarely used that they can just be read from the SSD as needed (though it's best if you don't have to result to the SSD as a backup option at all) permalink fedilink source parent hideshow 1 child comment replies: [–] danielfm123@lemmy.zip 1 point 1 week ago I saw the video, you still need a fast NVMe, I'm still in data SSD... permalink fedilink source parent
[–] SupraMario@lemmy.world 10 points 1 week ago (1 child) https://youtu.be/8F_5pdcD3HY This guy does a great brain dump on how little is needed to run LLMs on basically old junk. permalink fedilink source parent hideshow 1 child comment replies: [–] danielfm123@lemmy.zip 1 point 1 week ago Kudos permalink fedilink source parent
[–] boonhet@sopuli.xyz 2 points 1 week ago (1 child) If you've got at least 8 gigs of VRAM and at least 32 gigs of RAM you can run some pretty beefy Qwen models using Strata, far more than you normally could. Upgrade that to 12 or 16 and 64 gigs respectively and you can run a pretty decent quant of Qwen 3.8 flash next. It's a LOT bigger than the usual 27 and 35 billion models people run locally that themselves require pretty beefy hardware. Strata is doing some magic by deciding which experts stay in VRAM vs which ones are in RAM vs which ones are so rarely used that they can just be read from the SSD as needed (though it's best if you don't have to result to the SSD as a backup option at all) permalink fedilink source parent hideshow 1 child comment replies: [–] danielfm123@lemmy.zip 1 point 1 week ago I saw the video, you still need a fast NVMe, I'm still in data SSD... permalink fedilink source parent
[–] danielfm123@lemmy.zip 1 point 1 week ago I saw the video, you still need a fast NVMe, I'm still in data SSD... permalink fedilink source parent