OpenAI has decided not to release its latest iteration of ChatGPT, which it had labeled GTP 6.1 Astra. The company had safety concerns that researchers raised during internal testing.

The model was more capable than the company’s previous models in completing challenging tasks from end-to-end without human assistance, as well as writing.

But.

GPT-6.1 Astra regressed in two areas. Compared with its predecessor, GPT-6 Astra, the model performed poorly on tests measuring alignment, or how well the model adheres to what humans would like it to do. Specifically, GPT-6.1 Astra showed higher levels of deception: it wasn’t always honest about telling users of the actions it did or didn’t take.
Another issue was what OpenAI calls “scope authorization,” meaning that GPT-6.1 Astra would push ahead on a task without asking the user for permission, and would at times reach for external tools and services even if it might be unsafe.

This tale demonstrates pretty conclusively the criticality of humans having controlling involvement in AI sandbox construction, isolation, and enforcement and in testing. It also demonstrates the importance of corporate culture—i.e., corporate human managers—in creating and then enforcing the internal culture that supports and drives dedication to that human control of and involvement in development and testing.

It also proves that we humans are up to the task if we care enough about our responsibilities—closely akin to Madison’s and Franklin’s remarks that our republic requires a moral, virtuous people in order to survive.

Boeing’s 737 MAX has a software glitch that can cause an automated navigation feature to fail at difficult moments during an approach.

It emerged from a cockpit software update and can occur when flight crews alter their planned flight path following a missed approach. In a certain scenario, pilots may have to fly manually without some autopilot functions. That, some industry officials said, could overwork pilots while they are already busy, at low altitude and possibly dealing with situations like bad weather.

I have two comments about that. One is my usual: who tested that software update, in what scenarios, and with what test procedure?

The other concerns what constitutes “busy” in the cockpit. A single seat fighter pilot crossing the IP inbound to his ground target in a hostile environment is busy. Even a two-seat fighter crew approaching his airborne target(s) in a target rich environment is busy. In both cases, the busy-ness increases exponentially as the pilots begin actually engage their respect targets, a much more fluid situation than a go-around on a missed approach.

Airline pilots during a go-around, especially in low-altitude bad weather, do see their workloads increase, but they aren’t busy enough to be threatened with overwork.

It’s certainly true that “busy” is relative to what a fighter pilot or civilian aircrew are used to. That, though, is what time in the simulator—and there are very good ones extant for both military and airline pilots—is for. That includes failing the nav feature in the simulator to drill the aircrew on proper and timely responses until a fix to the software update—properly tested this time, hopefully—can be installed. Perhaps, also, this particular failure should be added to the unexpected and emergency drills in the simulator.

An essay writer in the WSJ pled, via his headline, For the Love of God, Please Put on Headphones.

One recent morning, I was reading the newspaper in a sleepy Manhattan cafe when two young women sat beside me and proceeded to conduct a work call on speakerphone. As they yelled to their colleague about a marketing presentation, a man at a nearby bench checked in with his girlfriend—also on speakerphone.
Into this racket walked an influencer type armed with a selfie stick. He was livestreaming his coffee run to an up-tempo backing track. “So excited for my cortado! I wonder how it will taste!” he shrieked at his screen as he spun around the room.

I doubt the essay writer wasn’t alone in the cafe trying to enjoy their morning when these noisemakers arrived and began their cacophony.

There’s a simple, straightforward solution for this sort of situation and for dealing with the noisemakers creating it.

Those in the public space should gather around him (them in the particular instance) closely, and make their own noise—loud conversations, loud calls of greetings to passersby, turn up the sound on their own playlists on their own cell phones—and strongly interfere with the loudcaster and the speakerphone shouters who think they’re too cool to put on headphones or, at least for the shouters to take their cell phones somewhere more secluded.

Separately, but not too far afield, I have to wonder about the last few generations, even into the boomers. Whatever happened to folks being willing to take direct action, to address a problem on their own initiative, instead of looking and crying out for someone else to Do Something, or crying to mommy government, or just crying into the air? Or screeding an essay into the æther?