"Something Big Is Happening” racked up over 75 million views—but it’s a breathless argument built on shaky assumptions, and it shows how hard it’s becoming to separate signal from hype in AI discourse
From my perspective, what is strange and unfortunate about Shumer's article (apart from the fact it appears on Musk's platform) is that it focuses exclusively on outer changes to our societies, institutions and economies, and does not even mention the likely larger, deeper transformation we face--a transformation in basic human psychology, in our experience of being human, in what it feels like to be a "me."
Agreed--but so little of the discourse is discussing this "mind revolution," as I call it on my Substack, focused on what seems likely to become an inner transformation at least as significant at the outer one.
I view the changes and advances in AI to be beneficial. Yes - it will disrupt the job market. But I also view many AI tools as to opening more doors in the self-employed and/or content creator markets. The key is for people to take courses and educate themselves about the popular platforms instead of blindly using them.
Thank you. I agree. If it was framed as “I’m giving you one scenario which could be very disruptive” that would have been easier to accept especially if he’d been specific about the technical drivers instead of just fear mongering. There are still a lot of unknowns including cost of agentic workflows and whether hallucinations are tolerable.
Thank you, Sharon for keeping it real. I thought of Matt’s article this morning as I corrected three different Claude hallucinations in one conversation thread. LLMS, while more powerful with better prompt templates à la Cowork plugins, are still suspect and need to be managed closely.
The core issue with "the duty of those in the space" is that they actually possess a very limited conceptual understanding of something that's turning out to have massive contextual implications.
They experience conceptual success but think that equates to a contextual understanding of systems and the broader world.
I think humans are fundamentally incapable of fully internalizing the effects of exponential growth. And that is what the author expects will happen.
If we are at the precipice of an intelligence explosion, then we don't have decades to collect data and put out probability scenarios. The change will happen slowly (last 5 years), and then suddenly (this year, as the author believes).
What I loved about that article was how much he leaned into "this is bigger than any of us, and it WILL take our jobs" all the way into "so the best thing for you to do is to allow it to displace you by using it and training it". Spoken like a person who is making a lot of money off of trying to make people dependent on a tool that most of us don't actually want to use.
Sharon, you nailed it with “What’s been most frustrating to me is how difficult it’s become to separate real industry signals from overwrought hype.”
This has been/continues to be true across every major development in AI in the last two years (which is as long as I’ve been in the industry).
It underscores how complex this technology is and the very real gap between what it can do for the select few and its potential to transform business more broadly. IE: sure, Shumer can use AI to do most of his work really well but that doesn’t mean a F100 company can get agents off the ground and start canning the interns. Silicon Valley keeps skipping that part of the story- the part where AI can’t transform your contact center because you’ve got no way to make those legacy systems work with it.
Everyone keeps saying what industry leaders are saying won't happen keeps happening, and it keeps happening faster than they say it is. Some of the same people who say this also say we needed to change the world because of climate change, even though the impacts hadn't happened yet.
I think there’s an important distinction between being directionally right and being right about timing and scope. Many predictions do eventually come true, and I'm not saying Matt Shumer is fully wrong. But it often comes later, more unevenly, with lots more surprises along the way.
Climate change is a good example of this. Climate change warnings were grounded in decades of data and still framed in probabilities and scenarios, not *inevitability* on a near-term timeline.
My concern isn’t that AI won’t change work—it will!—but that treating uncertainty as destiny doesn't help people analyze things, it's just alarming and confusing, which doesn’t actually help people prepare.
From my perspective, what is strange and unfortunate about Shumer's article (apart from the fact it appears on Musk's platform) is that it focuses exclusively on outer changes to our societies, institutions and economies, and does not even mention the likely larger, deeper transformation we face--a transformation in basic human psychology, in our experience of being human, in what it feels like to be a "me."
also something we don't know how it will play out
Agreed--but so little of the discourse is discussing this "mind revolution," as I call it on my Substack, focused on what seems likely to become an inner transformation at least as significant at the outer one.
I think the discourse will increase, but that kind of evolution also might sneak up on us :)
I view the changes and advances in AI to be beneficial. Yes - it will disrupt the job market. But I also view many AI tools as to opening more doors in the self-employed and/or content creator markets. The key is for people to take courses and educate themselves about the popular platforms instead of blindly using them.
Thank you. I agree. If it was framed as “I’m giving you one scenario which could be very disruptive” that would have been easier to accept especially if he’d been specific about the technical drivers instead of just fear mongering. There are still a lot of unknowns including cost of agentic workflows and whether hallucinations are tolerable.
Thank you, Sharon for keeping it real. I thought of Matt’s article this morning as I corrected three different Claude hallucinations in one conversation thread. LLMS, while more powerful with better prompt templates à la Cowork plugins, are still suspect and need to be managed closely.
thanks Geoff!
The core issue with "the duty of those in the space" is that they actually possess a very limited conceptual understanding of something that's turning out to have massive contextual implications.
They experience conceptual success but think that equates to a contextual understanding of systems and the broader world.
I think humans are fundamentally incapable of fully internalizing the effects of exponential growth. And that is what the author expects will happen.
If we are at the precipice of an intelligence explosion, then we don't have decades to collect data and put out probability scenarios. The change will happen slowly (last 5 years), and then suddenly (this year, as the author believes).
What I loved about that article was how much he leaned into "this is bigger than any of us, and it WILL take our jobs" all the way into "so the best thing for you to do is to allow it to displace you by using it and training it". Spoken like a person who is making a lot of money off of trying to make people dependent on a tool that most of us don't actually want to use.
Ha ha that's a good point!
Really like where you went with this, and the comments
Thank you!
Sharon, you nailed it with “What’s been most frustrating to me is how difficult it’s become to separate real industry signals from overwrought hype.”
This has been/continues to be true across every major development in AI in the last two years (which is as long as I’ve been in the industry).
It underscores how complex this technology is and the very real gap between what it can do for the select few and its potential to transform business more broadly. IE: sure, Shumer can use AI to do most of his work really well but that doesn’t mean a F100 company can get agents off the ground and start canning the interns. Silicon Valley keeps skipping that part of the story- the part where AI can’t transform your contact center because you’ve got no way to make those legacy systems work with it.
100%! Thanks Molly - I do think SV keeps skipping that part of the story
Everyone keeps saying what industry leaders are saying won't happen keeps happening, and it keeps happening faster than they say it is. Some of the same people who say this also say we needed to change the world because of climate change, even though the impacts hadn't happened yet.
I think there’s an important distinction between being directionally right and being right about timing and scope. Many predictions do eventually come true, and I'm not saying Matt Shumer is fully wrong. But it often comes later, more unevenly, with lots more surprises along the way.
Climate change is a good example of this. Climate change warnings were grounded in decades of data and still framed in probabilities and scenarios, not *inevitability* on a near-term timeline.
My concern isn’t that AI won’t change work—it will!—but that treating uncertainty as destiny doesn't help people analyze things, it's just alarming and confusing, which doesn’t actually help people prepare.