2026-06-08 15:40 UTC
Replies (3)
-
@TechTangents@dialup.space 2026-06-08 17:26
@FreakyFwoof@universeodon.com I draw the line at companies violating copyright on a massive scale and it being hand waved away. So if you have a small scale model made with cleared data, sure. We've had OCR since the 80s, google had real time video text translation over a decade ago. We don't need LLMs or transformer models for accessibility. We need purpose built products that actually make it viable.
-
@ClickyMcTicker@hachyderm.io 2026-06-08 17:34
@FreakyFwoof@universeodon.com @TechTangents@dialup.space AI is also useful for IT. By using it to watch for patterns that can allow you to mitigate issues before they cascade. I had an issue with storage latency causing what appeared as a memory leak on my SMB container, and I was struggling to solve it until I fed the logs live to a locally run AI. Once the next crash happened, I had a timeline of relevant needles plucked from the haystack. That said, I’m still predominantly opposed to generative AI. Most uses for it just result in noise that spams our senses. Plus they’ve been trained on uncompensated human labor, and are being used to degrade the living conditions of real humans. I’m not opposed to using technology to replace human labor, when done in a society that provides for those humans. Capitalism is not the society that provides for those displaced humans, nor is art a labor to be automated.
-
@Qybat@batchats.net 2026-06-08 20:51
@FreakyFwoof@universeodon.com @TechTangents@dialup.space I would draw a line as soon as any generative model's output is presented as human in origin.