Santiago
Machine Learning models are sneaky little bastards that use any available shortcuts to optimize their evaluation metric.

Every tutorial speaks about splitting your data randomly.

But you must be careful with this:
tweet

UnknownCollector
RT @blankspazer: Times Like These https://t.co/K38pDmJF10
tweet

proxima centauri b
GM ✨ https://t.co/9rFRkdlmnX
tweet

AK
RT @DalleSnoop: @_akhaliq It works. I generated this on my local GPU https://t.co/8BxCzFr2iI
tweet

Robin Hanson
“GPT4 .. can take into account only a limited amount of text, .. short enough to render all sorts of complex tasks impossible.” https://t.co/BDjBAct6RM
tweet

proxima centauri b
RT @8co28: Let's All Love Lain. https://t.co/Izk9Gje8a1
tweet

Richard Socher
RT @aidfulAI: What we today know as a prompt was initially invented by @RichardSocher, CEO of @YouSearchEngine, et al. and named question.

The rejection of that paper from the ICLR conference shows that not every good idea is directly identified as such. https://t.co/kUbA7BU4Fy
tweet

AK
RT @DotCSV: 👉 Perro con capa roja volando por el cielo. https://t.co/qAIylkqBWO
tweet