I like Windows. But I like unified memory more.

It's super-trite, but true: technology can be used for good or bad.

I love the productivity gains and breadth of instructional knowledge AI has given me.

I hate that when I'm on Facebook I have to spend half my time blocking groups that generate AI summaries of classic TV shows and characters (that I'm otherwise a big fan of and follow).

...assuming the book can still be found in print form

CSS is not Turing complete

cf V'Ger

Of the three "godfathers of AI", Yann LeCun is the only one who is not concerned about the dangers of AI. https://www.bbc.com/news/technology-65886125

He disagrees with his fellow godfathers that AI is a risk to the human race.

By "prover", I meant that Agentic AI is not a single-shot execution engine, like LLMs of today or theorem provers of yore. By "planner" I meant externalizing the logic/reasoning. Perhaps I was aiming too much for alliteration.

Apple’s new paper on GSM-Symbolic shows that today’s best language models crumble when a gradeschool math word problem is re-phrased -- even if the logic is identical. It echoes 1969, when Minsky&Papert proved that a singlelayer perceptron could never learn XOR.

That blockade vanished in 1986 with backprop and nonlinear hidden layers. My bet: LLMs won’t need two decades to cross the reasoning gap. Why? Agents that call scratchpad Python or GraphRAG pipelines already externalize formal reasoning, turning the model into a planner rather than a prover.

Look, I’m thrilled Vox can read an SEER plot and notice that smoking, screening, and HPV vaccines matter (slow clap). But before we crown Big Tobacco lawsuits and Gardasil as the sole saviors of humankind, can we maybe glance at, oh, the last half-century of environmental regulation?

What about the asbestos bans that cratered mesothelioma in post’70s construction cohorts? 84% risk reduction -- ring a bell? What about Chile and Taiwan slashing arsenic in drinking water and watching bladder and lungcancer mortality do a Wile E.Coyote cliff plunge two decades later? Or the Mercury & AirToxics Standards that took nickel, chromium, and friends down by 80% -- something the EPA’s own Section 812 analysis credits with thousands of avoided cancer deaths?

But sure, let’s keep peddling the tidy narrative that medical tech alone bent the mortality curve. Those radon-mitigation building codes? Irrelevant. Beryllium and benzene exposure limits? Yawn. Apparently if the benefit isn’t measured in ninefigure pharma revenue or a primetime Super Bowl ad, it doesn’t make the Vox word count.

Pro-tip: pathology doesn’t care whether the carcinogen came from Marlboro Country or your municipal tap. Policy matters, and not just the ones that poll well on Twitter.

Way earlier than that.

https://en.wikipedia.org/wiki/List_of_predictions_for_autonomous_Tesla_vehicles_by_Elon_Musk

...so the opposite of Windows 8 Metro

Microsoft CEO Admits That AI Is Generating Basically No Value

I've never seen an O'Reilly, Manning, Packt, or Apress book in hardcover

Besides its use in software development (though perhaps AI will obsolete that), learning to think mathematically -- spatially, temporally, and logically -- makes one less susceptible to deception, rhetorical tricks, and subjectivity. It allows one to think objectively (phonics also helps in that regard).

Memorizing multiplication tables exercises language cognitive capability more than it exercises spatial capability, which is key to understanding pure mathematics. Montessori does especially well at connecting numbers both spatially and in a tactile manner. The much-maligned Common Core attempted to teach math how mathemeticians think -- problem was school teachers (or their parents) were not mathemeticians, propagating the cycle of math weakness. Another way to gain spatial reasoning capability is to explicitly train for it.