This post was originally published on this site.
Last week, I shared how much closer we are getting to the singularity—when AI will be better than us at every human task, at which point nothing else is predictable.
Since, OpenAI has fully released its new model, o3, with the Deep Research mode. Today, we’re going to look at this release and how it will affect your jobs in the coming months and years.
An organization devised Humanity’s Last Exam two weeks ago:
The goal was to make the last exam we would ever need to gauge the intelligence of AIs: If they passed it, we could declare AIs are more intelligent than us.
Ten days later:
Let me graph this for you:
From Tyler Cowen:
“Adam Brown, who does physics at a world class level, put it well in his recent podcast with Dwarkesh. Adam said that if he had a question about something, the best answer he would get is from calling up one of a handful of world experts on the topic. The second best answer he would get is from asking the best AI models.”
So now only a handful of physicists are better than AIs in physics.
Arguably, Masayoshi Son, the CEO of SoftBank, one of the biggest investors in the world, had early access to o3. That’s why it decided to co-invest in the $500B AI investment announced last week. This is what he had to say:
Just a year ago, I thought AGI would arrive in 10 years. A few months after that, I said it would arrive in 2–3 years. But now, I want to correct it by saying it will arrive sooner than that.
That’s the context to understand this from Sam Altman:
In these three articles, I explained how we can expect AI to affect jobs:
-
If the demand for that job was not saturated, AI will improve that job’s productivity initially, creating more job opportunities and increasing the incomes of workers in that space.
-
As the supply of AI and automation saturates the demand for that job, AI and automation will start eliminating jobs in that space.
-
This process will be so fast that humans won’t have time to adapt into new industries: It will be faster to automate work with AIs than to train humans.
-
Today, the biggest limiting factor for this automation is the ability of humans to automate. It’s scarce. But automation capacity is getting better faster and faster. Soon, AIs will be more intelligent than humans. Once that happens, the speed of job elimination will accelerate. It won’t be limited anymore by the speed at which entrepreneurs can automate them (because going from idea to reality will be a matter of hours, not years and millions).
-
The real world is not safe, because robots are about to become good enough to replace humans in manual tasks that were safe from automation before.
-
The only jobs where humans will be safe long term are where humans appreciate having other humans interact with them, or where regulation is both strong and unavoidable.
-
AI is already taking over many tasks from different jobs, and reducing workload and jobs in specific ones.
I explored the industries of content, education, law, healthcare, logistics, childcare, construction, sales, customer service, illustration, modeling, and translation. It’s time to have another look to the progress I’ve noticed in some of these in the last few months:
Software developers are the Makers of God: They will be the ones to build superintelligence—and replace themselves in the process. So it’s important to track what’s going on in that field.
I already mentioned in the previous article that OpenAI’s o3 ranked as the #175 developer in the world.
The guy below is an expert realizing that o1 could write his PhD code (that took him a year) in 1 hour.
OMG it RAN!
OMG it ran.
That’s literally what my code does…
It ran…
Hahaha…
It did… my code…
Here’s an AI engineer at Meta, illustrating the idea that work becomes more valuable at the beginning of automation:
I shared this graph with you already:
Here’s a shocking anecdote to illustrate it: A few months ago, a Computer Science (CS) career fair was canceled in San Francisco because no companies wanted to hire anybody!
Programmers don’t ask each other questions anymore. AI answers most of them:
After spending millions to use Salesforce and Workday, Klarna stripped them out since it was able to rebuild their functionality internally with the help of AI (!!!).
We’re seeing across our whole business that things that previously took people a lot of time can be done much faster and much shorter with the help of ChatGPT, and we need fewer people to do the same thing.—Sebastian Siemiatkowski, founder and CEO of Klarna, suggesting that they will cut 1,800 jobs thanks to automation.
With the help of AI, the company is able to standardize and create a more lightweight tech stack to operate more effectively with higher quality.
And this was before the release of o1!
Now of course, we mentioned last time that AI companies are using AI to improve their AIs. Every week, this iteration loop will shrink until the speed of AI self-improvement makes most software developers redundant—and democratizes the job by making us all developers.
I have tried Replit a couple of times but couldn’t make it work with my software experience. I’ll keep trying tools until I find a combination that can get non-developers to develop apps.
If AI is automating the automators, how long do you think your job is going to last?
Next up: content, healthcare, driving, R&D and innovation, and takeaways. And later this week: How AIs are improving, and what’s left to AGI.