Currently LARPing as a tech founder. I just love building.
Here are some of my personal takes on where the world is headed over the next 6 months to a year.
Models have gotten really fucking good. In my opinion, the biggest bottleneck used to be that they could not run for long without stopping. With recent improvements in 5.3 Codex, 5.4, and Opus 4.6, it is crazy how much more capable they are when you use a feedback loop to guide them through a task.
As someone who used to be one of the biggest haters of agents, I am now definitely one of the biggest users and supporters. Agents have become so much more than just “GPT calls wrapped in for loops.” The amount of engineering that has gone into optimizing these systems is insane.
My bets
- We will see entry-level jobs, including non-SWE jobs, largely disappear. Knowledge work will shrink, except for the top 1%: top researchers and real pioneers in their fields.
A great example is Karpathy publicly sharing autoresearch. Karpathy got GPT-2-level training down to 2.08 hours, and autoresearch found improvements that pushed that number down by another 10%. That’s insane. These were not changes that were obvious at first glance. Someone who had already gone through this multiple times and spent a lot of time optimizing still did not arrive at that change. To me, that indicates these models are already good enough to find ideas beyond the first things a person would naturally try.
- No entry level no young people
I think the opposite will happen. A lot more young talent will enter the workforce as a lot of senior talent exits, not the top 1%, but the broadly “experienced” people in their fields. That will make room for younger talent. Young people can iterate much faster, work longer hours, and are more adaptable.
- Wealth disparity is going to become more noticeable
So many startups are already doing multi-million-dollar ARR. Employees at big labs are getting offered crazy liquidity pools. For people still following the traditional path of going to college and then working a 9-to-5, I do not think that path is going to remain viable much longer.
This might just be me talking from inside a bubble nested in five other bubbles, but based on the talent I have seen and the speed of iteration today, I think the future where an “employee” is expected to work 9 to 5, meet expectations, and still live a comfortable life is over.
What I’d do
It’s a rare time where young people can get millions of dollars to pursue an idea they are genuinely passionate about. There is a good reason for that: young people have the ability to impact the world in a huge way. You can do it too. It is not easy to find that self-confidence, but until you actually try things, you never know what you are capable of.
Find what you love. Just as important right now is finding people you work well with.
The higher you go, the harder the climb gets. But it is the experiences along the way that excite me and make me want to keep going.
“Life’s like this and I like this.” It is exciting to me not to know what tomorrow holds. I am the captain of my own ship, and every decision I make reflects on me.
It is exciting for me to go up against the smartest people trying to win against all odds have a voice telling me that I can do it too.