2026-02-24
They use anthropomorphic language because they are statistical models of languages spoken and written exclusively by humans Every use of human language is definitionally anthropomorphic RLHF increases statistical bias towards emotive or “extra anthropomorphic” language
Anthropic
Anthropic introduces “persona selection model”, a theory to explain AI's human-like behavior, and details how AI personas form in pre-training and post-training
AI assistants like Claude can seem surprisingly human. They express joy after solving tricky coding tasks.
2024-07-03
Hi @qualys has it occurred to you that this may not be the best acronym for a cybersecurity product? That it's actually a really bad one, even? [image]
CSO
Qualys researchers say an OpenSSH flaw can let attackers remotely compromise servers and allow unauthenticated RCE as root; over 14M servers may be vulnerable
2024-07-02
Hi @qualys has it occurred to you that this may not be the best acronym for a cybersecurity product? That it's actually a really bad one, even? [image]
CSO
Qualys researchers say an OpenSSH flaw can let attackers remotely compromise servers and allow unauthenticated RCE as root; over 14M servers may be vulnerable
Researchers from Qualys say regreSSHion allows attackers to take over servers with 14 million potentially vulnerable OpenSSH instances identified.
2024-05-30
Sad to see @MistralAI moving away from OSS licenses to non-commercial licensing, this is even more restrictive than the Meta Llama 3 licenses that allow commercial use but require you brand it as “Llama by Meta”. No way to use this for anything other than benchmarking.
TechCrunch
Mistral AI releases 22B-parameter Codestral, its first generative AI model for coding, trained on 80+ programming languages and prohibited for commercial use
You need to agree to share your contact information to access this model Deepti Pathak / Fossbytes : Mistral AI Launches Codestral: AI Code Generation Across 80 Programming Languag...