Subscribe to discover Om’s fresh perspectives on the present and future.
Living in the Petri Dish of the Future
If you missed the brouhaha around a 5,000-word blog post about AI, let me bring you up to speed. But before I do, I want to remind you that sometimes a blog post is not just a blog post. It is really about the spectacle of it all.
Let me rewind. HyperWriteAI CEO Matt Shumer’s post on Twitter, “Something Big Is Happening,” went viral this week. Last check, it had over 75 million views, though I can guarantee you very few people read the whole thing. I bet not as many read the summary. Still, no stopping the chattering classes.
It showed up so often in my feed that I decided to read it. I started and stopped a dozen times. I don’t know why I persevered. Glutton for punishment? My efforts should tell you something about the quality of writing.
His thesis is that coding is the canary in the coal mine and other professions will follow. That’s the meta point, and I don’t disagree with it. We are going through a sea change in how we interact with information, so professions are going to change as a result. The industrialization and automation of software has been a topic of discussion since my GigaOm days. When someone claims the canary is already dead and the rest of us are next, I want to hear what they have to say. Even if it takes a dozen attempts to get through the “essay.”
What really kept me going was his Covid comparison. Shumer compared this moment in AI to February 2020, when a virus was spreading somewhere overseas and nobody was paying attention. He thinks we’re in the “this seems overblown” phase of something much bigger than Covid.
This is not the first time I’ve heard that comparison. Months ago, a very smart, highly rational friend of mine explained it to me differently. And better. He pointed out that as average humans, we had never lived the scientific process the way we did with the Covid crisis and the emergence of the vaccine. No wonder everyone from pot-smoking podcasters to scientists had differing opinions. The whole thing was playing out in front of us in real time.
The trials, the debates, the reversals, the uncertainty, the competing claims, the people who were sure they knew what was coming and turned out to be wrong. We watched science happen. It was messy, public, unfinished. And everyone became an armchair epidemiologist overnight.
We are doing the exact same thing with AI. We are living inside the experiment. The results aren’t in. But that doesn’t stop anyone from declaring them.
Shumer’s piece was greeted with enthusiasm by Fortune magazine’s editor in chief. A day later, Jeremy Kahn at Fortune called it fear-mongering built on faulty assumptions. Gary Marcus called it “weaponized hype.” Fortune‘s AI reporter Sharon Goldman, who Shumer had personally pitched before publishing, wrote that she was glad she didn’t read a draft. Ed Zitron also pointed out that Shumer has a checkered past. In 2024, he released an AI model called Reflection 70B that he trumpeted as “the world’s top open-source model,” but independent researchers couldn’t replicate the results. Turns out the model appeared to be a wrapper around Anthropic’s Claude rather than anything new. Shumer apologized, said he “got ahead of himself.”
Let’s zoom out and focus on the real big question.
This whole drama, from the viral post to the takedowns to the counter-takes, none of it is really about Shumer’s essay. What it’s about is simpler. And harder to admit. In the words of screenwriter William Goldman, “Nobody knows anything.” As I have written before, we are living in a petri dish of the future. Some of us are hopeful. Some of us are terrified. Most of us are both, often in the same hour. And into that vacuum of uncertainty there is a torrent of speculation dressed up as prophecy.
Shumer writes his breathless warning. Marcus writes his skeptical rebuttal. Kahn points out the flawed assumptions. Goldman notes the difficulty of separating signal from hype. And every one of them is also selling something. A book. A newsletter. A reputation. An audience. Since attention is the currency and velocity is the authority, AI at present is the best momentum asset class.
Every cycle produces its prophets and its skeptics. And every cycle, the reality turns out to be different from what either camp predicted. Not better, not worse. Different. Yeah, I am so glad to be out of the damn forecasting business. I embrace the uncertainty, the future, and let it all happen.
As for Twitter, it probably has moved on to something else. I am sure. I haven’t looked.
February 12, 2026. San Francisco
Imagine that back in February of 1996, you took a world-wide vote. You say, “here’s an optional future: 30 years from now, you’ll be able to get anything delivered to your home in a day, you’ll be able to actually watch your taxi on it’s way to you (on your phone!) and you’ll be able to get the answer to any question anywhere at any time. The tradeoff is that no one will believe anything any more, people will be more isolated and depressed than ever and there will be no middle class.”
How many people do you think would have voted for that plan? I think they would have lost in the vote tally.
Now, we have something even more disruptive than the internet. We have zero idea what’s going to happen, but it’s scary AF because no one, even the people building it, have any idea what comes next.
I tink what people are scared of is that they don’t have a say in this. A handful of very wealthy people are pushing a technology forward that will fundamentally reshape the future (we know that much) and we have no idea how. And we have zero say in the matter.
We are absolutely in the petri dish. Whether we like it or not.
Exactly, we all want to be part of history. And this is yet another twist in the INternet’s tale. I am just happy to be here for the ride 🙂
Imagine in 1996 you said the tradeoff would be that billions would be lifted out of poverty and the middle class tripled world-wide, unimaginable information is delivered almost instaneously and cheaply so that news is not gatekeeped and you would be able to connect with friends and family anywhere in the world. I think th vast majority would have welcomed that future.
I think the largest issue here is that They can release AI onto the internet—for anything. Hacking, attacking websites, becoming bot herders, etc. AI doesn’t have a moral obligation, and I’m thinking back to the court case a while ago. I, at least, certainly don’t want to have AI rule the world, and you can imagine what might happen next; someone reads this post, go gets a $20 AI model and promptly copies this comment to feed into the AI for defense.
As a matter of fact, I’m going to do more than assume, I’m going to declare.
This is not going to end well.
By any means.
Alright, we have the advantage of several better options for creating whatever—websites if you’re not a coder etc, but the main issue is different.
Computers have gone a long way from reading lines of code that look like someone platter an ink brush over them.
Should we try to sue these AI makers? I don’t think that’s going to help, especially if they’ve just designed an especially good model, because they can have AI defend itself.
And that is trouble.
If we leave AI to just a few people, they can do whatever they want, so I don’t think that!s an option. I know, from personal experience, that AI is extremely powerful, and we can’t ignore it—apologies if this sounds to redundant, but that’s because this is what’s happening. Emphasis on is, and you get a very slight idea, from a single sentence, what AI might look like—if you were blind.
We can’t do much but wait it out. Good luck on surviving if you have a job on the internet.
Hi Om.
You are writing more, and I hope that means you are feeling better.
Since you are writing more, I am reading more. Your level approach, sound thinking, risk mitigation are welcome news in a feed that is overwhelmed, much like its human. Thank you.
And let’s catch up soon.
I am willing to write when I have something to say. It is not often, but once in a while wind carries me towards the right words.
Thanks for reading.
Obviously one should have an healthy doze of skepticism on claims and counter-claims.
No one, absolutely no one can predict how this will play out and list out consequences in hi-fi
This also a time where probabilistic thinking helps.
All possibilities and their consequences are probabilistic and they constantly and continuously change!
Note To Self: Absorb from all biases. Don’t get attached to any!
Well said: “Absorb from all biases. Don’t get attached to any!”
Thanks for this post! It’s still baffling to see how much Socrates still applies to the 21st century. His basic sentence – “I know that I know nothing” – seems still to be the greatest wisdom after almost 2500 years.
But there comes one of Socrates best successors into my mind: Karl Popper. His philosophy showed that because no one knows about all the consequences, we need to make sure that power is restricted. Otherwise, it is too easy that powerful people can make terrible, non-reversible decisions – not even because these people are bad, but just because they might be just misled by their own cluelessness.
And I think this last point is the problem, why people fear AI: AI already shows that it has an incredible power that might be hard to control. In a certain sense, the emergence of LLMs and agents, is more similar to the discovery of nuclear power than to the early Internet. Nuclear power opened up the ability to destroy humanity just with a day – AI seems to open up the potential to do extremely harmful things.
Life is an experiment. But because we are clueless, it is our duty to carry out this experiment in a way that avoids terrible harm to other people. And this seems to be widely ignored by the responsible people in favor of taking huge profits.
Spot on about the noise to signal problem, Om. This whole AI circus on social media is making everyone lose their minds when we should be taking a breath and thinking clearly.
The thing that gets me is while everyone’s having meltdowns about viral doom posts, the actual hard work of figuring out where AI creates real value versus just hype gets completely ignored. People are arguing about whether AI will kill 50% of jobs while missing the real question – what does it actually mean that generalisation is the bottleneck, not compute?
We’re all stuck in this petri dish together, but at least some of us are trying to read what’s actually happening instead of just panicking about it.
Thank You!
Your usual super effective work
I feel exactly the way you do, having been through all of these cataclysmic changes in the tech industry. I am now 84, and I think young people should read the essay, mostly so it doesn’t blindside them. If you are already working, you already know something’s coming. But the “something” will be different for each individual. I remember when friends of mine told me they didn’t need to know about computers, or the internet, or social media because they’d be dead before these things took hold. They were wrong, and they had to learn, so they did. That’s how the world works.
Om,
I appreciate this kind of in-depth analysis of stories that I don’t have the technical expertise to fully understand.
No one knows the exact path of how AI we’ll play out, but few have the humility to acknowledge that.
One of the big changes in how we live now is that everything is infront of our eyes. We live in the “process” so we don’t quite know how to process it.
This feels a lot like drilling holes in the bottom of the boat to let the water out.
Mike
thanks for the comment. Well you and I have both lived through the equivalent of “water in the boat” moments over the history of technology. But i think we are now living in a level of reality/ alternative reality like never before. I am just here for the ride and keep looking at how much our world keeps changing.
Hope you are well.
Indeed.
I am having a bit of deja vu –
watching people throw around numbers with exponents that should scare them silly are just so blasé. And that’s without asking just how Avogadro’s number becomes revenue.
Or how five or six companies can each acquire 30% of total RAM wafer production.
Musical electric chairs!
Would love to share a beverage and a bite if you somehow find yourself nearer the Potomac instead of The Bay.
Cheers