I've posted some skepticism of the new AI models that are getting all the press--and all the money--in the past year. I said in "When AI Phones It In" that a lot of the fear of jobs being taken away is vaporous. And in "The AI Winter Shit-Winds Are Coming" I suggested we might be heading for an AI crash. Which would be unfortunate since we've been riding crashes in the markets already for a couple years. Last month, in "Smells a little bit like AI winter?" writer/scientist Gary Marcus asked if the simultaneous "implosion" of AI failures at Tesla, Google and Microsoft could lead to an AI Winter.

Microsoft's new A.I.-enhanced Bing search engine has gotten a lot of press over the past few weeks for being combative, rude, and just creepy. The chatbot told Kevin Roose of the New York Times that it loved him and that, although he's married, Kevin doesn't actually love his spouse. It told a Washington Post reporter that it can "feel or think things." On Twitter, a lot of the folks with early access to the chatbot (it's in a test stage and is not yet available to the public) posted screenshots of chilling conversations. Faced with the potential threat of runaway A.I., for years researchers have been working on safety guardrails, out of the public eye, hoping they could keep these systems functioning the way they're supposed to.

While many Americans were nursing hangovers on New Year's Day, 22-year-old Edward Tian was working feverishly on a new app to combat misuse of a powerful, new artificial intelligence tool called ChatGPT. Given the buzz it's created, there's a good chance you've heard about ChatGPT. It's an interactive chatbot powered by machine learning. The technology has basically devoured the entire Internet, reading the collective works of humanity and learning patterns in language that it can recreate. All you have to do is give it a prompt, and ChatGPT can do an endless array of things: write a story in a particular style, answer a question, explain a concept, compose an email -- write a college essay -- and it will spit out coherent, seemingly human-written text in seconds.

Over the last couple years, I've been spending time writing about creating ghosts -- perhaps an inevitability in the midst of a pandemic. While created by far-from-supernatural means, these are ghosts nonetheless; they are created from an essence of you -- from your voice, your data, your feelings, beliefs, habits, and history. Groups around the world are looking to take such information, this essence, and use it to create a digital version of you that may last once you are gone. Consider it a technological solution to the problem of death. Over the last couple years, I've been writing about creating ghosts -- perhaps an inevitability in the midst of a pandemic.

Therapy has an engagement problem. Despite the benefits of treatment plans and at-home exercises, people generally resist anything that feels like work, and this impedes the mental-health recovery process across the board. Clinicians have attempted to bridge this gap with various devices and reward systems, but still, it's often incredibly difficult to motivate patients to help themselves. Video games have the opposite problem. Players can spend hours immersed in a single digital experience, seated in one spot and lost in their own world, but they're often branded as "lazy" for this behavior.

In a mid-2020 issue of Supply & Demand Chain Executive, I had the pleasure of speaking with managing editor Brielle Jaekel on the "emerging technologies that claim to help companies in the supply chain." "In the near future, supply chain AI will begin to migrate to machine learning. Currently, supply chain AI consists of developers programming business rules, telling computers what to look for and what action to take when it encounters those situations, but as AI migrates to machine learning, it will begin to think for itself. As machine learning becomes more advanced, technologies will increasingly be able to make note of repetitive situations and past experiences to start learning and making recommendations on its own. Technology like this has already deployed on a wide scale in other industries, and it has the potential to rapidly automate and improve a wide range of supply chain processes."

You'd barely even know it, but CES just happened this week. In the last couple years the big technology trade show was very much a car show, where automakers and startups alike showed off the latest in hopeful self-driving systems always billed as just "a few years away." Barely a peep from anyone, and that may be because last year was the year everyone got real about autonomous cars. After a year that saw the first death of a human at the hands of a self-driving Uber prototype, and then an admission by Google's own self-driving technology chief that such cars won't ever be fully able to drive in all conditions, those in the automotive and mobility spaces seem to be taking a much more measured approach to things from here on out. Development of autonomous cars continues, to be sure, as does the advancement of semi-autonomous driving aids like those found on many modern cars. But we seem to be past the days where every car company, startup and government official swears we'll have a fleet of robo-taxis or delivery vehicles by 2020.

As we're nearing the end of 2017, we've come to the 5 year landmark of deep learning really starting to hit the mainstream. For me, I think of AlexNet and the 2012 Imagenet competition as the coming out party (although researchers have definitely been working in this field for quite a bit longer). It's been just 5 years and we've absolutely revolutionized the way we look at the capabilities of machines, the way we build software (Software 2.0), and the ways we think about creating products and companies (Just ask any VC or startup founder). Tasks that seemed impossible just a decade ago have become tractable, granted you have the appropriate labeled dataset and compute power of course. In this post, we'll overview the last couple years in deep learning, focusing on industry applications, and end with a discussion on what the future may hold.

As we're nearing the end of 2017 (and coincidentally the first day of NIPS 2017), we've come to the 5 year landmark of deep learning really starting to hit the mainstream. For me, I think of AlexNet and the 2012 Imagenet competition as the coming out party (although researchers have definitely been working in this field for quite a bit longer). It's been just 5 years and we've absolutely revolutionized the way we look at the capabilities of machines, the way we build software (Software 2.0), and the ways we think about creating products and companies (Just ask any VC or startup founder). Tasks that seemed impossible just a decade ago have become tractable, granted you have the appropriate labeled dataset and compute power of course. In this post, we'll overview the last couple years in deep learning, focusing on industry applications, and end with a discussion on what the future may hold.

Artificial intelligence (AI) has captured the imagination and attention of doctors over the past couple years as several companies and large research hospitals work to perfect these systems for clinical use. The first concrete examples of how AI (also called deep learning, machine learning or artificial neural networks) will help clinicians are now being commercialized. These systems may offer a paradigm shift in how clinicians work in an effort to significantly boost workflow efficiency, while at the same time improving care and patient throughput. Today, one of the biggest problems facing physicians and clinicians in general is the overload of too much patient information to sift through. This rapid accumulation of electronic data is thanks to the advent of electronic medical records (EMRs) and the capture of all sorts of data about a patient that was not previously recorded, or at least not easily data mined.