Behind the scenes hackers and coders have been raising the alarm about the
privacy risks (reality, not a risk) of the use of commercial tools to access LLM'z, pushing themselves hard to provide advice and code to companies - at the cost of their sleep for the vast majority, under the pressure of management, in order for these companies to enter the market of "AI" to make profits at the cost of our security.
It makes me sad that today the majority of publication about #deepseek, (which is released in Opensource) availability turns around threats on privacy while our infosec community's work has been ignored for now more than two years. The topic of privacy will circulate on social networks to be ignored in a couple of weeks (I am optimistic), pushed out by another topic about LLM, eluding again the fact that any company can run its own models locally on dedicated servers.
Bests,
#opensource
#malinformation