The rise of Google, Facebook, Twitter, LinkedIn, and their ilk has driven much of the craziness. Semiconductor executives complain that the fact that these companies — and Apple — continue to staff up expanding semiconductor design teams is making it hard to retain good engineers.

Hey, if I was a smart, 20-something chiphead, I’d be shopping my resume to the hyperscalars, too. They have the youthful culture, moonshot projects, and career-making potential these days.

But this is just the stuff of a lively simmer. The red-hot wave of deep learning that’s sweeping through the industry has got the lid rattling off the pot.

Today, I’ll go to the second annual SysML event at Stanford. This group is moving so fast writing research papers on the new style of computing that it took a year to take a breath and post late last week its mission statement:

“We propose to foster a new systems machine learning [ML] research community at the intersection of the traditional systems and ML communities, focused on topics such as hardware systems for ML, software systems for ML, and ML optimized for metrics beyond predictive accuracy.”

The charter was signed by some 50 luminaries across more than a dozen top universities along with Amazon, Facebook, Google, IBM, Intel, Microsoft, Nvidia, and a small handful of newcomers I need to get to know.

What’s taking my breath away is that this is the third major new group spawned at least in part by deep learning in the last three weeks.

Last week, I attended the first workshop of the Open Domain Specific Accelerator group defining an open chiplet standard. The week before, I spent two days at the first meeting of TinyML, a group of top embedded systems and research veterans pushing on the nexus of AI and IoT.

Next week, I’m going to form a support group for tech journalists with aching fingers and worn-down keyboards. I’m going to call it PantingML.

Traffic around the Bay Area isn’t as bad as it was in 1999 at the height of the web craze, when optical networking startups were coming out of the woodwork, but it’s getting close. Cost and availability of housing is worse as the crane becomes a more pervasive, invasive species than crabgrass.

Clearly, there is a lot of important R&D to be done in deep learning. But just as surely, we are getting a bit ahead of ourselves.

Dozen of silicon startups have sprouted, many of them going after data center sockets that the hyperscalars are working on filling themselves. I get a couple of announcements a week about new AI software platforms claiming that every enterprise in the world is poised to embrace neural networks.

There’s no doubt that we are going to need more data scientists. But I think we could use a few specialists in common sense, too.

Beyond tech, economists have been saying for more than a year that we are poised for the next global downturn. They don’t know just what will trigger it or how hard it will hit, but it’s coming, they say.

A long hot summer is coming, too. I’m going to get a good air conditioner before the dog days arrive and they go out of stock.