Yeah I had a suspicion that the trend shown in the chart in that thread regarding the decline of DL job posts largely resembles the trend of total job posts.

I think the fact the original tweet was not normalized in this incredibly obvious way is at least one valid reason companies could use less deep learning folk

The original tweet is one of the authors of TensorFlow (specifically Keras, a large section of the API for v2), if it's any indication of the quality of the framework.

I wouldn't expect anything different from François..

That was my suspicion as well.

Btw. I don't like twitter's new feature that prevents everyone from responding to a tweet that was used by @fchollet. It no longer feels like twitter if you can't engage.

Once you reach 100k followers, you only need a 0.1% jerk rate, to always have a 100 people in your comment section that do nothing but troll, rile you up, or demand you defend your thoughts against their stupid uninformed disagreements. Chollet has 210k followers.

> demand you defend your thoughts against their stupid uninformed disagreements.

And I shall use this pulpit to demand, in a mixture of derision and righteous anger, that you defend your comme... ah never mind.

This may not be a new thought, but it's eloquently put. Thank you.

Disingenuous framing of data or a laughably fundamental misreading of it? This is akin to trying to gain insight from a bunch of data on a map that simply has a strong correlation with population density.

If only we could apply ML/AI to the data on ML/AI job postings.