"We Have No Idea How Models will Behave in Production until Production": How Engineers Operationalize Machine Learning.

Systems can be dynamic. Machine learning models are inherently limited because they are trained on static datasets, so their “knowledge” is fixed. Therefore, developers need to combine models with other components, such as search and retrieval, to incorporate timely data. In addition, training lets a model “see” the whole training set, so more com

... See more

One interesting thing about LLMs is that they can actually recover (and without error loops). You can have a step that doesn't work right, and a later step can use its common-sense knowledge to ignore some of the missing results, conflicting information, etc. One of the problems with developing with LLMs is that the machine will often cover up bugs... See more

A rough analogy to the current LLM process is that making a new model is like baking a cake. You figure out your data and algorithms—like mixing the batter—and then you pretrain the model, that is, run it on a large number of computers for several months—like putting it in the oven—and then at the end you do some “post training”—like frosting and d... See more

One thing that is still confusing to me, is that we've been building products with machine learning pretty heavily for a decade now and somehow abandoned all that we have learned about the process now that we're building "AI".

The biggest thing any ML practitioner realizes when they step out of a research setting is that for most tasks accuracy has ... See more