Enabling vs. Disabling Technology

Got the idea for this essay from constant dilemma I have found myself since the past year : is the software world of startups really worth pursuing anymore with how cracked coding agents are getting. Not only me but a lot of my friends that I have built software companies with in high school and other college-aged people that got into the AI software startup space right when GPT-3.5 dropped face this question.

Building to learn

If in the next couple of years AI truly makes production software development autonomous, then through building software startups you really are not “learning” anything through the process. Building an AI startup in the mid-2024s was the sweet spot : you got to work with frontier AI tech as RAG and Agents were still a new concept, and applying that to a lot of legacy workflows and applications provided really huge upside in the product you built.

Importantly, you couldn’t even achieve it through vibe-coding alone. Actually looking through API documentation, writing agent loops by hand, generating CRUD files to DB chunk by chunk, and going through the integration of Stripe with bare-hands was painful, but you were LEARNING something in the process. You were getting more technical, specialized knowledge through the experience that a then vibe-coder was not.

The new "being technical"

Now with coding agent teams not only cracking the code to all of this but even writing full E2E tests, generating SQL migrations and various other operations in one-shot, it does enable you to work on high-order systems design but you are not necessarily learning anything technical. At-least what was considered "technical" before. Being technical in 2026 is much more different, and for a good-reason.

Motivation and value

So a lot of young people are just not motivated anymore to work software startups. One would expect you to be more enthusiastic about building something when its easier then ever before, but its the opposite. If building something is the easiest it has been, so can a lot of other people. And for this reason you don’t learn anything perse. A lot of them are actually going more into learning and growing expertise in concepts like frontier machine learning, deep reinforcement learning, and VLAs than writing software apps.

And this prompted me to think : how can I evaluate what I am working on is valuable not in the next 5 years but in the next 25-50 years. What is the one good heuristic I can apply to an idea and evaluate if I am on the right path.

Elon Musk and Tsiolkovsky

I referred back to Elon Musk’s idea of why one starts companies. It’s from one of his talks where he cites Konstantin Tsiolkovsky’s quotes about “being among the starts, expanding the scope” and that even the thought of leading humanity towards something as grand should just excite you to wake up everyday and work on it.

Enabling vs disabling technology

Essentially, a startup/company like SpaceX “enables” humanity do things rather than disable. If you are contributing to humanity with something like SpaceX, you “enable” humanity to do something that cannot be done without you. Enable humans to go to planets like Mars. There is a difference between technology that truly enables, and technology that makes an existing workflow a wee bit faster.

Let’s say if you are building one of those wearable startups that give you either a necklace or back-of-the-phone attachment to transcribe your in-person meetings and give you summaries, it just disables you to accomplish already existing work a bit faster and with less attention. Instead of actually paying attention to your convos, you just rely on your transcriber and then listen to your summaries. It definitely helps you extract the more relevant information and help you rather than hurt, but it is technology that "disables".

Deep tech vs consumer tech

One misconcpetion that may arise that I will just address now is technology that enables and disables doesn’t necessarily have to align to the divide between deep tech and consumer tech. Facebook being a software startup enabled people to stay connected with their community and the world on a scale not possible before. However, a startup like Weave robotics (I love what they are doing with Isaac) doesn’t necessarily “enable” you to do things. It just fold and does laundry for you, which does disable you from doing the work yourself, but I mean who wants to do their laundry anyway.

Note : This idea also partly originated from one of my friends claiming how despite both Cheerios and Apple being huge brands, your life without Cheerios would probably be the same but without an iPhone it would be vastly different.

Final thoughts

Circling back to Weave robotics, I also don’t want to claim that working on technology that disables is necessarily a “bad” thing. It again is a choice, and it actually is very ambitious if you want to work on technology that truly enables very early in your life. Capital, innovation, a hard-to-validate idea and a long timeline are hard obstacles that are gonna be on your way. It is just that if you do have deep expertise in hard-tech like robotics and you are using that to build a company, much rather spend time building robots to do work in energy or manufacturing plants than house-help activities.