Post

By 2022 Google Translate had been using transformer models for several years. By 2014 Facebook was using deep convolutional neural nets (DeepFave) for face detection. Literally the things you think were “before LLMs” are either basically indistinguishable from LLMs or their direct predecessors.

This post is unavailable.

2/8/26, 4:07 pm7 reposts98 likes

aly

@aly.codes

2 Aug

Generalized computer vision is much newer and more generally useful than being able to pattern match facial measurements by dotting eyes and noses.

aly

@aly.codes

2 Aug

Seriously guys, Google’s Attention Is All You Need paper is almost 10 years old. The life you imagine from a decade ago is just a collection of extremely rudimentary systems of the same or similar form.

aly

@aly.codes

2 Aug

*DeepFace (if the folks at Apple could fix iOS 27 autocorrect before release, perhaps with Better Autocorrect, I would appreciate it)

affine

@refinement.systems

2 Aug

seq2seq RNNs were pretty impressive for the time too (many people remember the Tay drama, and it was likely one of these). I guess one could draw a line at "generating text message by message = virtuous ML" and "whole sessions = wretched GenAI" but that's just silly to me.