I really liked this piece thanks. The point about tastes and how they change and are pretty hard to predict is really helpful. I tend to forget about that aspect when thinking about industry prospects. As a book lover I tend to think about prospects for books and bookshops. I see lots of bookstores closing and these days not much turnover at my local secondhand bookshop. But now I think of it, there’s lots of energy in other parts of my local book world. For instance, the thrift shops do a roaring trade and local secondhand book sales by charities are pretty huge community events so that’s something. So now I think who knows what the future will bring for book sales and who knows what new ways of producing, selling and reselling books will appear. It’s kind of exciting to think of as a book lover.
As a fellow book lover, both physical and digital, I love the democratization and spread of where to find books. The little free-standing neighborhood “library” is an example. Digital has helped self-publishers reach audiences they’d never have found. My favorite example was reading Hugh Howey’s Sand series for “free” on Kindle Prime (or whatever they call it now). From that to the streaming show on Apple and paying $15-20/copy now.
Bottom-line- more reading is a social benefit, at a lower price for a broader reach.
Books are doing quite well as an industry. Independent booksellers are seeing a resurgence. Audio books are huge now. Genres shift around as does distribution but on the whole book sales are up. Data says the global market is increasing at 8% YoY as a trend line.
This comment was originally posted on the a16z Substack:
As someone who’s building a GPU-powered Terminal App, the return to a CLI-first experience makes me giggle so hard. The times have changed but only in our memory perhaps! What is old is now new, etc. etc. etc.
I liked this quote but wanted to adjust it a bit:
> “New tools will be created with AI that do new things.”
to:
> “New tools will be created with AI that do old things in new ways.
The gutenberg press enabled writing and publishing (and copy-pasta-ing) at a greater scale. The typewriter did the same and then computer and then the internet and now AI.
Every major leap seems to follow the same pattern:
1. Reduce the cost of reproduction
2. Expand who can produce content
3. Trigger a crisis among existing gatekeepers (church, state, publishers, media companies)
4. Restructure power — those who control information lose monopoly
These are amazing times but our fundamental needs and even workflows really haven’t changed. We are still human after all these years.
As Ben Affleck said, the "death of ____" is mostly narratives that help the AI companies fundraise. Totally agree that there will be a ton of new software. But I do think there will be a new layer where companies are building light applications that fix certain things for them. Just like their do with Excel and other tools today. And with AI that becomes easier to do.
Everything looks like the end of history to some people. But it never is, and you'll be caught with your pants down if you act like it is.
Starting after the first world war (and really kicking in after the second), Europe bet that the Industrial Revolution was going to be so productive that they could skim socialism off the top and everyone could take it easy forever. The US did not make that bet, and got the driver's seat for the computing, Internet, and now AI revolutions.
Now there are people who should know better saying AI is going to be so productive we can skim UBI off the top and take it easy forever. Steven's got it right: don't believe it, and be ready for whatever is next.
AI won’t destroy software. I do think it will change who gets to collect ($$) and who gets left behind. It might easily shift the balance of power in society.
Over the last couple generations of financialization and the consumerization of computing, entrepreneurs and some established companies have created many amazing tools. There’s no doubt about that. But this period (for multiple reasons) has also seen a mass transfer of income from labor to capital, relatively stagnant middle class wages, and growing wealth inequality. The job creation from tech has disproportionately favored the economically and socially privileged.
We now have more access to news than ever but a handful of platforms are reaping the benefits while newsrooms are closing and journalists are losing their jobs. We can stream any song we want for free but artists earn pennies on the dollar while streaming platforms and record labors make bank.
We don’t know yet if AI will accelerate the concentration of wealth into fewer hands, but the land grab for compute among a handful of hyperscalers and model makers suggests it might. The types of high-paying jobs AI creates might ever relatively few, and the balance of those jobs might be for relatively low-paying service work. I’m still holding out hope for some amazing medical and materials sciences breakthroughs with AI, but those will take quite a while to develop.
I think you largely nailed it. However, there are a couple of nuances: some software is more vulnerable than others. For example, Adobe is easily replaced by an AI generative model that outputs better video and graphics. Application is based on search and manipulation of open source, text files such as legal financial 10k’s, etc. are also vulnerable.
On the other hand, enterprise systems are deeply embedded in the enterprise with highly customized schemas and while AI add new functionality to them , it will not replace them. More over, enterprise data is so incredibly messy and dirty and duplicated that it is not fit for purpose for many AI manipulation. What will change with enterprise software is that we see a movement away from seat pricing and towards volume pricing. This is also a great market for AI based data mastering tools to clean up enterprise data.
You do indeed present some valid points, but I disagree with what you say here: "I will never forget how bankers said they were going to be so much more efficient by having a few college kids use spreadsheets while they did all the hard work of deal-making. Little did they know that to even be a banker by 1995 you were going to be the person using the sheet to build a model. Repeat for consultants. Graphic artists. Writers. Lawyers. Doctors. And every single domain. This is going to happen all over again. Yes, some supporting jobs in domains went away, but they were not just replaced with even more skills, they were replaced with many more people. "
I hear this argument a lot but we have to compare apples to apples, which you are not.
For example, when spreadsheets came, indeed more people were needed, because someone had to operate the spreadsheet, since spreadsheets do not take actions on their own - ditto for illustrator for graphic designers, photoshop etc... Spreadsheets, illustrators, photoshops are tools, someone needs to operate these tools, spend a lot of time to get a meaningful outcome with a substantial learning curve; they do not have the level of intelligence to go beyond tools, to take action based on minimal instruction - which is what AI does.
Cluade, chatgpt and the like are not merely tools...
The invention of Cluade, chatgpt has never been done before, so no analogies exist from the past to compare and extrapolate.
Even for domain experts, how much time is needed before Claude catches up and these domain experts become obsolete?
“What person can imagine waiting seven days or an entire summer for the next episode of a show?”
IMO, this was largely a mistake driven by a need to profess differentiation by Netflix at the time. It was a flex, but it also cannibalized their audience’s potential for virally influencing each other in a more coordinated build-up of a show’s current hype. Now nobody can speak to each other without disclaimers when an entire season is dumped at once.
Notice how even Stranger Things Season 5 pulled back by staging it in three releases?
I really liked this piece thanks. The point about tastes and how they change and are pretty hard to predict is really helpful. I tend to forget about that aspect when thinking about industry prospects. As a book lover I tend to think about prospects for books and bookshops. I see lots of bookstores closing and these days not much turnover at my local secondhand bookshop. But now I think of it, there’s lots of energy in other parts of my local book world. For instance, the thrift shops do a roaring trade and local secondhand book sales by charities are pretty huge community events so that’s something. So now I think who knows what the future will bring for book sales and who knows what new ways of producing, selling and reselling books will appear. It’s kind of exciting to think of as a book lover.
As a fellow book lover, both physical and digital, I love the democratization and spread of where to find books. The little free-standing neighborhood “library” is an example. Digital has helped self-publishers reach audiences they’d never have found. My favorite example was reading Hugh Howey’s Sand series for “free” on Kindle Prime (or whatever they call it now). From that to the streaming show on Apple and paying $15-20/copy now.
Bottom-line- more reading is a social benefit, at a lower price for a broader reach.
I hadn’t thought of those examples they are excellent!
Books are doing quite well as an industry. Independent booksellers are seeing a resurgence. Audio books are huge now. Genres shift around as does distribution but on the whole book sales are up. Data says the global market is increasing at 8% YoY as a trend line.
I’m doing my bit for the increase in book sales!
This comment was originally posted on the a16z Substack:
As someone who’s building a GPU-powered Terminal App, the return to a CLI-first experience makes me giggle so hard. The times have changed but only in our memory perhaps! What is old is now new, etc. etc. etc.
I liked this quote but wanted to adjust it a bit:
> “New tools will be created with AI that do new things.”
to:
> “New tools will be created with AI that do old things in new ways.
The gutenberg press enabled writing and publishing (and copy-pasta-ing) at a greater scale. The typewriter did the same and then computer and then the internet and now AI.
Every major leap seems to follow the same pattern:
1. Reduce the cost of reproduction
2. Expand who can produce content
3. Trigger a crisis among existing gatekeepers (church, state, publishers, media companies)
4. Restructure power — those who control information lose monopoly
These are amazing times but our fundamental needs and even workflows really haven’t changed. We are still human after all these years.
—-
😬🙄🤣🥰
As Ben Affleck said, the "death of ____" is mostly narratives that help the AI companies fundraise. Totally agree that there will be a ton of new software. But I do think there will be a new layer where companies are building light applications that fix certain things for them. Just like their do with Excel and other tools today. And with AI that becomes easier to do.
Everything looks like the end of history to some people. But it never is, and you'll be caught with your pants down if you act like it is.
Starting after the first world war (and really kicking in after the second), Europe bet that the Industrial Revolution was going to be so productive that they could skim socialism off the top and everyone could take it easy forever. The US did not make that bet, and got the driver's seat for the computing, Internet, and now AI revolutions.
Now there are people who should know better saying AI is going to be so productive we can skim UBI off the top and take it easy forever. Steven's got it right: don't believe it, and be ready for whatever is next.
Once again, your essay illuminated the possibilities in a optimistic but balanced assessment. Thanks.
AI won’t destroy software. I do think it will change who gets to collect ($$) and who gets left behind. It might easily shift the balance of power in society.
Over the last couple generations of financialization and the consumerization of computing, entrepreneurs and some established companies have created many amazing tools. There’s no doubt about that. But this period (for multiple reasons) has also seen a mass transfer of income from labor to capital, relatively stagnant middle class wages, and growing wealth inequality. The job creation from tech has disproportionately favored the economically and socially privileged.
We now have more access to news than ever but a handful of platforms are reaping the benefits while newsrooms are closing and journalists are losing their jobs. We can stream any song we want for free but artists earn pennies on the dollar while streaming platforms and record labors make bank.
We don’t know yet if AI will accelerate the concentration of wealth into fewer hands, but the land grab for compute among a handful of hyperscalers and model makers suggests it might. The types of high-paying jobs AI creates might ever relatively few, and the balance of those jobs might be for relatively low-paying service work. I’m still holding out hope for some amazing medical and materials sciences breakthroughs with AI, but those will take quite a while to develop.
I think you largely nailed it. However, there are a couple of nuances: some software is more vulnerable than others. For example, Adobe is easily replaced by an AI generative model that outputs better video and graphics. Application is based on search and manipulation of open source, text files such as legal financial 10k’s, etc. are also vulnerable.
On the other hand, enterprise systems are deeply embedded in the enterprise with highly customized schemas and while AI add new functionality to them , it will not replace them. More over, enterprise data is so incredibly messy and dirty and duplicated that it is not fit for purpose for many AI manipulation. What will change with enterprise software is that we see a movement away from seat pricing and towards volume pricing. This is also a great market for AI based data mastering tools to clean up enterprise data.
Hi Steve, thanks for your article.
You do indeed present some valid points, but I disagree with what you say here: "I will never forget how bankers said they were going to be so much more efficient by having a few college kids use spreadsheets while they did all the hard work of deal-making. Little did they know that to even be a banker by 1995 you were going to be the person using the sheet to build a model. Repeat for consultants. Graphic artists. Writers. Lawyers. Doctors. And every single domain. This is going to happen all over again. Yes, some supporting jobs in domains went away, but they were not just replaced with even more skills, they were replaced with many more people. "
I hear this argument a lot but we have to compare apples to apples, which you are not.
For example, when spreadsheets came, indeed more people were needed, because someone had to operate the spreadsheet, since spreadsheets do not take actions on their own - ditto for illustrator for graphic designers, photoshop etc... Spreadsheets, illustrators, photoshops are tools, someone needs to operate these tools, spend a lot of time to get a meaningful outcome with a substantial learning curve; they do not have the level of intelligence to go beyond tools, to take action based on minimal instruction - which is what AI does.
Cluade, chatgpt and the like are not merely tools...
The invention of Cluade, chatgpt has never been done before, so no analogies exist from the past to compare and extrapolate.
Even for domain experts, how much time is needed before Claude catches up and these domain experts become obsolete?
“What person can imagine waiting seven days or an entire summer for the next episode of a show?”
IMO, this was largely a mistake driven by a need to profess differentiation by Netflix at the time. It was a flex, but it also cannibalized their audience’s potential for virally influencing each other in a more coordinated build-up of a show’s current hype. Now nobody can speak to each other without disclaimers when an entire season is dumped at once.
Notice how even Stranger Things Season 5 pulled back by staging it in three releases?