As all readers of this essay know, I am not in any way expert in machine learning (ML) and large language models (LLMs), so my descriptions and observations are, at best, lightweight cartoons of what is actually going on. Please keep this in mind as you read this. Some of you may remember Spock's death in Star Trek II (Wrath of Khan) and the brief scene where Spock mind-melds with Dr. McCoy: Spock says "remember" while depositing his katra in McCoy's brain in anticipation of self-sacrifice to save the starship Enterprise. As I read about yet another new breakthrough in artificial intelligence (AI) from Google Research, I thought of that scene. The new idea, christened "TITAN", is for a ML system to continue learning while in use after training.a

Authoring interactive stories where the player is afforded a wide range of social interactions results in a very large space of possible social and story situations. The amount of effort required to individually author for each of these circumstances can quickly become intractable. The social AI system Comme il Faut (CiF) aims to reduce the burden on the author by providing a playable model of social interaction where the author provides reusable and recombinable representations of social norms and social interactions. Motivated through examples from an in-development video game, Prom Week, this paper provides a detailed description of the structures with which CiF represents social knowledge and how this knowledge is employed to simulate social interactions between characters.

Eventually, vision scientists figured out what was happening. It was the mental calculations that brains make when we see. Some people unconsciously inferred that the dress was in direct light and mentally subtracted yellow from the image, so they saw blue and black stripes. Others saw it as being in shadow, where bluish light dominates. Their brains mentally subtracted blue from the image, and came up with a white and gold dress.

Hillsdale Middle School summer students lined up on opposite sides of the school cafeteria. They were getting into position to have a battle, one they had been preparing for through a summer of math, reading and problem solving activities. The battle would not be between the students but between robots, which the students programmed to turn and move in specific ways by writing code. The science, technology, engineering and mathematics teacher, Jenny Stump, counted down. While the actual battles -- there ended up being multiple rounds -- only lasted a few minutes each, weeks of preparation went into building the skills necessary to execute the activity.

Hillsdale Middle School summer students lined up on opposite sides of the school cafeteria. They were getting into position to have a battle, one they had been preparing for through a summer of math, reading and problem solving activities. The battle would not be between the students but between robots, which the students programmed to turn and move in specific ways by writing code. The science, technology, engineering and mathematics teacher, Jenny Stump, counted down. While the actual battles -- there ended up being multiple rounds -- only lasted a few minutes each, weeks of preparation went into building the skills necessary to execute the activity.

"The rule of thumb is that both images--the one from the satellite and the one from the autonomous vehicle--have to have identical content for current techniques to work. The differences that they can handle are about what can be accomplished with an Instagram filter that changes an image's hues," says Anthony Fragoso (MS '14, PhD '18), lecturer and staff scientist, and lead author of the Science Robotics paper. "In real systems, however, things change drastically based on season because the images no longer contain the same objects and cannot be directly compared." The process--developed by Chung and Fragoso in collaboration with graduate student Connor Lee (BS '17, MS '19) and undergraduate student Austin McCoy--uses what is known as "self-supervised learning." While most computer-vision strategies rely on human annotators who carefully curate large data sets to teach an algorithm how to recognize what it is seeing, this one instead lets the algorithm teach itself.

Imagine you and an intelligent robot are both before a judge who cannot see you. The judge will guess which of you is the human, and so will live, while the other will die. Both you and the robot want to live. The judge is fair and smart. The judge says: "You must each give me one word from an English dictionary. Based on this word, I will guess who is the human."

Imagine that you're living in some dystopian future, and you have been accused of being an advanced AI, which is outlawed in this society. The penalty is death, and in order to convince the judge who will decide your fate, you can utter just one word, any word you like from the dictionary, to prove that you're flesh and blood. What word do you choose? It sounds like the setup for a cheesy sci-fi short, but this is actually part of a curious paper from a pair of researchers at MIT on something they call the "Minimal Turing Test." Instead of a machine trying to convince someone they're human through conversation -- which was the premise of the original Turing Test, outlined by British scientist Alan Turing in his seminal 1950 paper "Computing Machinery and Intelligence" -- the Minimal Turing Test asks for just one word, either chosen completely freely or picked from a pair of words.

If you think your on-the-job training was tough, imagine what life is like for newbie surgeons. Under the supervision of a veteran doctor, known as an attending, trainees help operate on a real live human, who might have a spouse and kids--and, if something goes awry, a very angry lawyer. Now add to the mix the da Vinci robotic surgery system, which operators control from across the room, precisely guiding instruments from a specially-designed console. In traditional surgery, the resident gets hands-on action, holding back tissue, for instance. Robotic systems might have two control consoles, but attendings rarely grant residents simultaneous control.

The implication, of course, is that there is a potentially much better outcome for the hiring manager and employer in terms of candidate fit. While a human being would still meet and interview highly ranked, suggested candidates, it's clear that this kind of technology could make the recruiting process much more efficient and successful. In addition to solving a candidate volume problem, AI also has the ability to prevent an over-abundance of poorly matched job applicants in the first place. McCoy explained, for example, that some of Alexander Mann's clients employ chat bots to help prospective job seekers get the information they need or want to know that simply isn't available in a job description. McCoy told me that chatbots can be used to help job applicants understand whether they can keep their social media profile if they take a job or how flexible an employer is -- before they even apply for a role at a company.