Blog de Méthode
Retrouvez nos publications ci-dessous
LLM-free e-mail parsing, but better
1,500 e-mails a day, 11,000 a week, more than 500,000 a
year.
That's the data load we're expected to manage with a single client. When we first started to do NLP, statistics and heavy data-driven calculations on this many e-mails, we faced a major issue: everything is LLM-oriented, despite the models not being fit to do such a repetitive task [...]
That's the data load we're expected to manage with a single client. When we first started to do NLP, statistics and heavy data-driven calculations on this many e-mails, we faced a major issue: everything is LLM-oriented, despite the models not being fit to do such a repetitive task [...]
OpenSSL insecure ciphers: WHY is it so HARD to enable them & HOW to do it
Did you know it is almost impossible to retrograde TLS security
by enabling deprecated ciphers on recent versions of OpenSSL?
Of course, it is a good thing, so developers won’t do it by accident [...]
Of course, it is a good thing, so developers won’t do it by accident [...]