Yep: the company decided to invest in local AI machines. All the “regular” devs get a 6k€ machine (this one) and the power users will get a 30k€ one (not sure which kind yet). The machines run whatever the devs like - in my case, llama.cpp and the GPT-OSS-120B coding model. It works well to spews out boilerplate and serves as a really handy “super search engine” that can code examples for me. That’s all I need to work faster with AI.
I’m glad the company went that route. Cloud shit is shit and I don’t like it, and I like cloud AI shit even less. Local AI makes us free from American vagaries at a very reasonable cost - at least for a company.
Mistral lost our business by not being proactive and I’m glad they did.
Update?
Yep: the company decided to invest in local AI machines. All the “regular” devs get a 6k€ machine (this one) and the power users will get a 30k€ one (not sure which kind yet). The machines run whatever the devs like - in my case, llama.cpp and the GPT-OSS-120B coding model. It works well to spews out boilerplate and serves as a really handy “super search engine” that can code examples for me. That’s all I need to work faster with AI.
I’m glad the company went that route. Cloud shit is shit and I don’t like it, and I like cloud AI shit even less. Local AI makes us free from American vagaries at a very reasonable cost - at least for a company.
Mistral lost our business by not being proactive and I’m glad they did.