Friday Charts: Will natural intelligence keep us employed?
Bill Gates expects that within a decade, humans will no longer be needed “for most things”

thinkhubstudio/Shutterstock modified by Blockworks
This is a segment from the Blockworks Daily newsletter. To read full editions, subscribe.
“I do not fear computers. I fear the lack of them.”
— Isaac Asimov
Bill Gates expects that within a decade, humans will no longer be needed “for most things.”
We’re entering an era of “free intelligence,” he explains, in which expertise in all subjects will be commoditized — the best doctors and teachers, for example, will be AIs.
“It’s very profound and even a little bit scary,” Gates told Harvard Magazine, “because it’s happening very quickly, and there is no upper bound.”
Listening to ex-Google CEO Eric Schmidt tell it is even scarier: “Within the next one year, the vast majority of programmers will be replaced by AI programmers.”
Within five years, he adds, we’ll achieve artificial general intelligence, after which AIs will be as smart as the top expert in any field — and within six years we’ll achieve artificial super intelligence, where a single AI is smarter than the sum of all human intelligence.
“There’s no language for what happens with the arrival of this,” he warns.
But I wouldn’t worry too much just yet — these remain predictions about the future and the future, as Yogi Berra tells us, is famously hard to predict.
Also, evidence from the present suggests that humans won’t be so easily replaced.
For example, the founder of Cursor said this week that the AI code editor he developed is now writing one billion lines of code per day.
That’s a lot — almost certainly more than the total lines of code written by every software developer in the United States put together.
And yet, the Bureau of Labor Statistics estimates there are at least 1.6 million software developers still employed in the US.
That could be about to change, who knows?
But if AI is already writing a billion lines of code without any discernible effect on employment, I suspect that’s an indication that human programmers have many years of employment to look forward to still.
Another indicator: Subscriptions for Microsoft’s Copilot have reportedly been stagnant for a full year now — which suggests that Copilot really is an “AI companion” for human employees (as the company markets it) and not a replacement for them.
Similarly, OpenAI reports that ChatGPT has over 400 million active weekly users now — almost 5% of the world’s population!
But I still have to type this newsletter out myself.
LLMs don’t seem to be taking any other jobs, either: “We find muted effects of AI on employment due to offsetting effects,” one academic study reports. “Highly exposed occupations experience relatively lower demand compared to less exposed occupations, but the resulting increase in firm productivity increases overall employment across all occupations.”
Another study found similarly that despite widespread adoption of LLMs, “AI chatbots have had no significant impact on earnings or recorded hours in any occupation.”
So even newsletter writers appear to be safe for now.
And even if not, we’ll probably find something else to do.
Benedict Evans recently recounted the long history of new technologies leading to higher, not lower, employment, citing examples such as the invention of typewriters and adding machines: “What did that do to clerical employment? People hired far more clerks. Automation plus the Jevons paradox meant more jobs.”
Evans illustrates this point by noting that the vast floor of human calculators that Jack Lemmon worked on in The Apartment has since been replaced by a single spreadsheet.
But there are far more people employed using spreadsheets today than there ever were using adding machines.
In some places, the greater concern is that there are too many jobs and not enough people: Greece, for example, introduced a six-day work week last year “due to the twin perils of a shrinking population and shortage of skilled workers,” as The Guardian reported it.
Similarly, at least part of the reason for the persistent shortage of affordable homes in the US is that there aren’t enough humans available to build them.
This could change, of course.
ChatGPT may soon write this newsletter faster and better than I can — I’ve noticed, not without trepidation, that it’s slowly been getting better at proofreading individual sentences for me.
And even construction workers may someday fear for their jobs.
Analysts at Bank of America predict there will be three billion humanoid robots roaming the planet by the year 2060, doing everything from handling hazardous materials to caring for the elderly.
2060 is still a long way off, however, and chances are we (like Jack Lemmon) will have found better things to do by then.
In the meantime, though, let’s keep ourselves busy by checking the charts.
How it’s going so far:
A simple but amazing chart: Exactly one month after Liberation Day, Nasdaq is 3.2% higher. That is ….not what I expected. Also, Liberation Day was ONE month ago??? Feels like 10 years.
Not all is well in markets, however:
Gold is up, the dollar is down and oil is down a lot. This seems more consistent with yesterday’s news that tariffs are costing Apple $900 million a quarter and this morning’s news that Japan isn’t interested in the trade deal currently on offer.
Zooming out:
The US dollar tends to move in decade-long trends. With Japan now threatening to unload its massive stockpile of US Treasurys, it’s easy to see why the next trend might be down.
The economy’s firewall:
This morning’s report that the US added 177,000 jobs in April, which is encouraging, but of course backward looking, according to Mark Zandi: “The job market remains a firewall between continued growth and a downturn. But…the firewall feels fragile. Unless the trade war de-escalates in the next few weeks, the firewall will come down, and a recession will ensue.”
Looking forward:
Torston Slok warns that this morning’s employment data was collected the week after Liberation Day tariffs were announced and that the correlation to University of Michigan survey data suggest that employment is set to turn lower and “perhaps even negative.”
Don’t forget the little guys:
Apple can weather tariffs at even a cost of $1 billion a quarter, but smaller businesses dependent on imports from China probably cannot.
Congratulations to the class of 2025:
This chart from the Atlantic shows that contrary to the historical norm, recent college graduates now have a higher unemployment rate than the average American. (I’d probably still trade places with them though.)
It’s cheaper to stay in school, at least:
Bloomberg’s Jon Authers notes that Covid broke the multi-decade uptrend in the cost of college, which is now headed lower.
Recession odds:

The Polymarket odds have been ticking up, which seems weird given what stocks are doing. One analyst estimates that stock prices are now pricing in only a 34% chance of recession. Even more discordant is the estimate from Constance Hunter, chief economist at the economist intelligence unit, who puts the probability at 80%.
Not a prediction:
The latest data from Vizion’s Global Ocean Bookings Tracker shows exports from China to the US down 48.6% from a year before. “Maybe the children will have two dolls instead of thirty,” the president said in response this week. “Maybe the two dolls will cost a couple of bucks more.”
That sounds pretty recessionary to me — but maybe he’s right that the boats are filled with “stuff we don’t need.”
As long as we keep the jobs we need, I think everyone would be fine with that.
Let’s hope that’s what the markets are telling us.
Have a great weekend, naturally intelligent readers.
Get the news in your inbox. Explore Blockworks newsletters:
- Blockworks Daily: Unpacking crypto and the markets.
- Empire: Crypto news and analysis to start your day.
- Forward Guidance: The intersection of crypto, macro and policy.
- 0xResearch: Alpha directly in your inbox.
- Lightspeed: All things Solana.
- The Drop: Apps, games, memes and more.
- Supply Shock: Bitcoin, bitcoin, bitcoin.