living in the future

Posted: March 12th, 2026 9:41 PM

  • "Oh, vibe coding is lame"
  • "Oh, AI is still not that good"
  • "Oh, being an aimaxxi is a waste of time"
  • "Oh, it's too expensive"
  • "But you aren't actually learning anything"
  • "You're offloading the cognitive work to the AI, that's the most important part"
  • "You're a glorified food taster and missing out on experiencing hard work"
  • "You must believe (insert random thing) since you think AI is so good"

this is cope.

I don't think that any of this is true. I don't think that it is a fair assessment or an intellectually honest perception of the current reality that we are living. This is coming from a place of p(DOOM). This is among the most ridiculous and close minded things that anyone could say in this very moment. Arguably the only worse thing one could say is that to use AI or to think that you could make something useful using only AI is ridiculous. These should be crimes at this point.

I realized how much progress AI had actually underwent when I started using Opus 4.6 via Claude Code. You can now do all of the things that were promised to us with GPT 4. Remember that time? Remember when GPT 4 was being marketed and talked about as AGI? It struggled to summarize a collection of articles into something unique and readable. Now, with GPT 5.4 you can not only get some sense of taste (through assigning a role or framework for the model to apply), but you can type a real thing into existence. That kind of progress is eerily similar to human progress

GPT 4 is like an early highschool student. They know some math. They know how to read (although they may not enjoy it). They know how to write (although you may get a wide range of quality). It's not bad, but objectively it is not the quality that an adult human knowledge worker would find compelling. GPT 5.3 is a college student approaching their peak. They are functional at a level that raises eyebrows when they talk about 'their major'. They can produce work that excites the working professional because in some ways, it's better than the deliverables they see at their day job. The ability for the college student to focus on a task for long hours (like reading an entire Russian book in a 7 hour stretch because the mid term is in a few days) is amazing. That's where we are now.

Sure, sometimes the model will struggle if it gets too deep; so do college students. Sure it is still rather impressionable. This is expected. AI and the eventual emergence of AGI is more like the development of a human like being than it is like the creation of a smart machine than we think. This realization has really humbled me. I have a deeper respect for my daughter who is learning to walk and taking her first steps. Her development from not being able to move on her own to crawling at rapid rates for long hours is amazing! And in less than a years time! At this rate she should be flying in approximately 29 months!

AI is at this inflection point currently. It is good. The barrier to entry is virtually null. There is a powerful free version of the models. But for only $100 a month you can get a college level student (even a professor level in other areas) that never needs sleep and has the only bottleneck of RAM usage and network speed. You can type almost anything into existence. You do not need to know years of jargon and nuance in order to build that thing you've always thought about. You simply need some source of income and focus (and a computer i guess). We are living in the future.

A good analogy that I enjoy using is the analogy of New York shortly after the first gasoline engine was made and the US saw the first versions of the automobile. Imagine with me:

We are living in the Eastern States. We hear that someone has made a machine that uses the wheels of a cart and some liquid to move. No legs required. No horses needed. We hear that they've made enough of them to race them the way we race horses. You work as a farmer outside of town and I work as a courier. You think about the time, effort, and money to put into maintaining horses. You think about the early morning feedings, the late night weening, the weekend birthings, the absurd amount of feed they go through in such a short amount of time. You think about all of the water you don't get to use for your crops, your food, or yourself, because the horses need it. You think about the vetinary costs and the depression that hits your daughters as they mourn the loss of a being that was a family member. You think about the smell. You think about your uncle who got kicked in the chest and died before he hit the ground because he was in the wrong standing spot. I think about the long hours, the chaffing, the rashes, the sweat, the heat, the winters, the back aches, the neck tightness, the throbbing migraines of hauling goods. I think about the small group of thugs that beat me into darkness and I had to see the doctor who told me, "You're lucky the good Lord didn't take you." I think about the number of pairs of shoes I've gone through. I think about the ripped pants. I think about how much time a pair of wheels would save me.

What neither of us can even dream of thinking about in this moment is that the other thousands of people in the area are thinking about this too. We can't even comprehend that people, investors, would give the inventors vast sums of money for this work. We can't comprehend seeing not horses, but only these machines in town. Even more, we can't imagine that our city will be unrecognizable in only 20 years.

We live in that moment where we have this realization and both of us are given a set of wheels and that engine. We may not be given all the parts and we sure don't understand how the thing works. We could use it to make a better version. We could use it to make a different kind of engine. We could use it to build the best kind of car. We could use it to compete in the open market. Or, as we both took time to imagine for ourselves, we could use it to improve our lives.

that is the opportunity

The labs have every reason to do the favorite activity of every SaaS business since 2019; price hikes. They should be charging more. Some of the labs need to be charging more. The opportunity cost of inaction has never been higher in our lifetimes than it is right now. We are not weighing the cost of inaction for not building the thing we want to. We are compounding that number by the potential increase for access.

I've built 3 projects recently using almost exclusively Claude Code. He (yes I want to anthropomorphize Claude) has done a great job at building out these projects. The loop I've followed is using Web or Desktop Claude to develop the loose plan and layout of the project and then to generate the relevant prompts that will be given to Claude Code. I specifically ask for .md output because I've noticed it does much better at structuring its output in that way. I also ensure that my prompt reads less like a stream of conciousness and more like a rough design doc. Using the prompts I'm given from Claude I then go to Claude Code and start pasting in the prompts. I ensure that actions are gated so I need to verify that whatever is being installed via curl is not malicious (I've had to check an install twice and neither were malicious). I then work through phases of the development and once I get to a point of testing I assume the personal role of a rigorous tester. I go in and try to break it and use it. I note everything I don't like and if there are simple things in functionality or UI that give an error output, I prompt Claude Code directly citing the errors or logs. If there is some different direction I want to take the development then I go back to Claude Desktop or Web and the cycle repeats.

So yes, I am assuming the personal role of a food critic. I tell the chef what to make and then I tell him how to make it better (in my perception of 'better').

The three projects I've made recently are:

We are living in the future. Take advantage of it now. Do something about it.

lesson

the world will never be the same. we are truly at a turning point in humanity. there are still people who have no real idea about what is happening. our jobs are probably safe for awhile because of all of the opportunity and work that AI will create. don't panic. and definitely don't doom.

you can both learn ai and learn how to use ai. these things are not mutually exclusive. i am going to learn mechanistic interpretability