Since I can’t post in AI Channel, I try here. Since everybody embraces Deepseek and it is also hosted on major platforms. Is anybody aware of the risks the local usage of deepseek poses ? Are there any security concerns and if so which ones ? Or can we trust that there is no backdoor or model inherent threat, e.g hidden backdoors, weight manipulation, unauthorised code execution or supply chain attacks. How are these and potentially other risks mitigated ? Do security checks on hugging face catch that? Or since it has a new novel lower level optimisations, these can’t be checked. Any help would be appreciated. Thx 🙏