The biggest opportunity for the AI news room is to do news that is not being done by a human now because it is not profitable enough to pay the human. Using an AI powered newsroom to bring back local news is a great example, but there are others. You may still need a human in the loop but that human could cover a dozen towns instead of 1, taking it from a money loser to highly profitable.
Next-gen journalists need to follow the ‘fugu chef’ analogy: Just as fugu (pufferfish) chefs in Japan command top dollar because they’ve mastered a high-risk, high-reward skill, tomorrow’s most valuable roles will be those that require deep, hard-to-automate expertise. Commodity tasks will increasingly be handled by AI; what remains uniquely human (and high-value) are the “fugu” skills—those that combine domain knowledge, judgment and complex coordination. https://open.substack.com/pub/platforms/p/the-fugu-guide-to-jobs-in-a-world
Check out my substack site, saratogacivicpulse.substack.com, which extensively uses AI to help busy citizens stay civically engaged with their local government.
Ask ChatGPT to create a syllabus for your Yale seminar. Give it a one or two sentence summary of what you're trying to accomplish. Tell it how many weeks/sessions there will be. And see what it does.
I did that, once and rejected it, not for content (I forget what it told me to do, honestly), but because it felt like I'd be playing a cover song - someone/something/some conglomerate of predicted someone and something else's work.
Fair enough! Though you do play covers extraordinarily well and in a unique way... And maybe devising a curriculum isn't the value you bring to the class?
You mean this Substack? I'm doing that, and I ask my students to feel free to use AI on songwriting, PROVIDED they tell us all what they're doing so we can learn from it. Some of the musical uses are solutions to nuts and bolts problems that there used to be no way around. But the real differentiators in the class remain out of reach thus far.
Maybe the question is more personal: How to give up "control" of a process that feels "successful" only when it springs from someplace deeply subjective? I'm resistant, and don't want to be, but I have to be able to trust my tools, and there are many very good reasons not to trust this one. I doubt I need to outline them.
The biggest opportunity for the AI news room is to do news that is not being done by a human now because it is not profitable enough to pay the human. Using an AI powered newsroom to bring back local news is a great example, but there are others. You may still need a human in the loop but that human could cover a dozen towns instead of 1, taking it from a money loser to highly profitable.
Hard agree!
Just going to put this here: https://www.bostonglobe.com/2025/05/20/business/chicago-sun-times-philadelphia-inquirer-ai-summer-book-list/
Next-gen journalists need to follow the ‘fugu chef’ analogy: Just as fugu (pufferfish) chefs in Japan command top dollar because they’ve mastered a high-risk, high-reward skill, tomorrow’s most valuable roles will be those that require deep, hard-to-automate expertise. Commodity tasks will increasingly be handled by AI; what remains uniquely human (and high-value) are the “fugu” skills—those that combine domain knowledge, judgment and complex coordination. https://open.substack.com/pub/platforms/p/the-fugu-guide-to-jobs-in-a-world
Check out my substack site, saratogacivicpulse.substack.com, which extensively uses AI to help busy citizens stay civically engaged with their local government.
How does one learn how to use AI productively? Like where to go, what to read?
Ask ChatGPT to create a syllabus for your Yale seminar. Give it a one or two sentence summary of what you're trying to accomplish. Tell it how many weeks/sessions there will be. And see what it does.
I did that, once and rejected it, not for content (I forget what it told me to do, honestly), but because it felt like I'd be playing a cover song - someone/something/some conglomerate of predicted someone and something else's work.
Fair enough! Though you do play covers extraordinarily well and in a unique way... And maybe devising a curriculum isn't the value you bring to the class?
Hello, Mike! Start by reading this! ;-). And then experiment.
You mean this Substack? I'm doing that, and I ask my students to feel free to use AI on songwriting, PROVIDED they tell us all what they're doing so we can learn from it. Some of the musical uses are solutions to nuts and bolts problems that there used to be no way around. But the real differentiators in the class remain out of reach thus far.
Maybe the question is more personal: How to give up "control" of a process that feels "successful" only when it springs from someplace deeply subjective? I'm resistant, and don't want to be, but I have to be able to trust my tools, and there are many very good reasons not to trust this one. I doubt I need to outline them.