apejcic
Re:DeepSeek: DeepSeek v4.1 Flash 1017
on Saturday October 17, 2026 @08:13AM (#50130421)
Its a great update, I ran a benchmark comparing verbosity and intelligence compared to old weights.
- Reasoning effort = 1 is now actually usable and produces 2.5x less tokens than old weights, without intelligence lost. - Overall 20% less tokens at each reasoning effort. - Same intelligence and same tool-calling pattern.
More details in blog post https://sference.com/blog/deepseek-v41f-1017-benchmark
Re:I tricked Claude into leaking your deepest, darkest secrets
on Wednesday July 15, 2026 @06:14PM (#48927801)
True
Re:I tricked Claude into leaking your deepest, darkest secrets
on Wednesday July 15, 2026 @10:57AM (#48921833)
You don't have to use Claude Code. Use PI (fully local harness) with ZDR open weight model provider (eg like sference.com).
I'm getting negative reaction for providing an obvious solution. Anthropic stores all your prompts as well and uses it for training.
Re:I tricked Claude into leaking your deepest, darkest secrets
on Wednesday July 15, 2026 @03:21AM (#48917366)
Use GLM-5.2 on ZDR inference provider like sference.com