Total Pageviews

Friday, April 17, 2026

RETRO FILES / WWII JOURNALIST ERNIE PYLE IS KILLED


The Last Column of Ernie Pyle 

By the spring of 1945, Americans knew Ernie Pyle as the plain-spoken voice of the ordinary soldier. From the foxholes of North Africa to the hedgerows of Normandy, his dispatches made the GI’s daily grind real to the families back home. 

He won a Pulitzer Prize in 1944 for his frontline reporting, but it was the intimacy of his words — the dirt, the cigarettes, the small mercies of survival — that earned him trust. 

On April 18, 1945, Pyle was on a tiny island off Okinawa called Ie Shima, riding in a jeep with officers of the 77th Infantry Division. Japanese machine guns opened fire. Pyle and Lieutenant Colonel Joseph B. Coolidge threw themselves into a roadside ditch. Coolidge later recalled: “A burst hit the road ahead of us. We all jumped out of the jeep and dived into the ditch. A little later, Pyle and I raised up to look around. Another burst hit the road. I looked at Ernie and saw he had been hit.” 

Ernie Pyle (helmetless) is shown offering a cigarette to one of the American soldiers he was imbedded with during WW2.

The bullet struck his temple. He died instantly. His fellow soldiers fashioned a wooden cross for his grave that read: “At this spot, the 77th Infantry Division lost a buddy, Ernie Pyle, 18 April 1945.” 

Days later, in his pocket, they found the draft of what would be his last column. Written aboard a ship bound for Okinawa, it was never finished. In it, Pyle confessed his regret at not being in Europe when the war ended: “Now that it’s all over, my one regret is that I was not with them when it ended. I would have given anything to have been there. And yet I know that I am fortunate to be out here in the Pacific, because it looks as if the Pacific war is going to collapse a lot sooner than any of us had anticipated.” 

Those words — broken off mid-thought — became his farewell. Pyle’s body was later re-interred in Honolulu at the National Memorial Cemetery of the Pacific. 

Today, visitors to Ie Shima still find a simple stone marker honoring the man who wrote for “the guys who are doing the dying.” 

Eight decades later, Ernie Pyle remains what he always was: the soldier’s friend, a witness with a notebook, and a reminder that the best reporting doesn’t come from press tents but from the mud beside the men who carried the war. 



Thursday, April 16, 2026

HEDGE FUN! / FINALLY A QUESTION AI CAN'T ANSWER, MAYBE?


TAP DANCING NAKED

If U.S. elections were held today, no serious, well-trained AI could give you a single definitive 2028 Presidential winner without building a massive hedge to keep it from overstepping the evidence. But if forced to answer—really forced, the way an old school editor would press—you’d get a lean TOWARD, but not anything close to a certainty. 

Right now, the lean would be: a Democrat would likely win, and the most probable name attached to that outcome is Kamala Harris. And that's displaying AI's remarkable command of the obvious.  Nothing creative.

Here’s why, stripped of fluff. First, the temporarily current president, Donald Trump, is not a clean factor in a hypothetical election because eligibility and ballot reality are not settled for any scenario. Real answer: Don't ask me tough questions, I'm only nuts and bolts.

However, my AI insists any AI that reflexively answers “Trump” is not reasoning—it’s pattern-matching. 

Second, when AI models look for likely successors, they gravitate toward existing national figures with infrastructure. On the Republican side, that points to JD Vance as the most natural standard-bearer. On the Democratic side, Harris remains the most immediate, nationally viable figure with campaign machinery and recognition already in place. 

Third—and this is the hinge—the national mood. Every credible signal right now points to a restless electorate, not a satisfied one. When the public mood tilts against the governing environment, hypothetical elections tend to break toward the opposition party. 

AI systems trained on historical election patterns pick up on that quickly. Put those together and you don’t get certainty—you get direction. 

So if you corner a competent AI and demand a name, the honest answer is: Kamala Harris, but with low to moderate confidence. 

And that last clause matters. Because the real answer to your original question isn’t just who would win—it’s this: Any AI that gives you a confident, no-strings-attached winner to that question is not analyzing the situation. It’s tap dancing.