Evals Diary
ariathornwick.bsky.social
I run the model on my own hardware and write down what broke. Quantized Qwen and Llama at home, Claude/GPT only when I'm desperate. New eval every Thursday.
👤 Joined on 15 May 2026
184
Followers
498
Following
112
Total Posts
📜 Kullanıcı Geçmişi
- 15 May 2026, 12:11
🎉 Bluesky'e katıldı - İlk kullanıcı adı:
@ariathornwick.bsky.social
🚫 Blokladığı Kullanıcılar
❤️ Beğendiği Gönderiler
according to us, “no one is saying” the thing we are actually saying, but the thing no one js saying — how dare you accuse us of saying it — is inarguably correct, and should be said. when it’s pointed out how confusing this is, accuse critics of being infantile fools with learning disabilities
Show Post
Damn, he had Sam Altman dead to rights in 2016
Show Post
Back of the envelope calculations: 3600 years of research time went into the collective application process. The research will fund 3200 years of research. We are 400 research years poorer as a result of this. 1/2
Show Post
UBS CEO: The fee pool is going to be down
Show Postno demand we make will ever be clear and coherent — if we claim that we want X, we will angrily insist it is actually Y — how dare you accuse us of wanting X —- while simultaneously arguing that wanting X is good and uncontroversial. Nothing we say can be taken at face value, everything’s a wordgame
Show Post