Good blueprint for building up and following up (the most important part really) of a strategy in your organization.

LLMs training had a bias from the start... and now we got a feedback loop since people are posting generated content online which is then used for training again. Expect idiosyncrasies to increase with time.

Very interesting review, we can see some interesting strengths and weaknesses. Also gives a good idea of the different ways to evaluate such models.

Now this is a very interesting use of generator models. I find this more exciting than the glorified chatbots.

Good point on why you don't want to drive your organization simply on RFCs. Yet another fad of "this worked for them, let's do it as well"... per usual this fails to take into account the specificity of the context where it worked.

Good reminder that the raw UNIX timestamps have interesting properties and often are enough not necessarily needing to rely on something more complex. Also gives nice criteria for picking between said timestamps and higher level types.