Don't worry, you are not in danger

A developer traces the rapid evolution of LLM interaction techniques from 'prompt engineering' to 'context engineering' to 'harness engineering,' where models autonomously iterate and learn. The post warns that as LLMs gain more autonomy and access to production systems, the trajectory becomes unpredictable and potentially risky.

We went from "prompt engineering" to "context engineering" to "let 'er rip" in less than 2 years. "Prompt engineering" always sounded cheesy to us "converted engineers": that's not real engineering. It's "fiddling" at best. But is it all that trivial? Probably. Soon after came "Context engineering": The idea that working with an LLM is not about one message, a "prompt" that we send out as the singular input of our request. Rather, we should manage and "engineer" the entire bulk of the "conversation", build up a cohesive structure, compress and filter out the noise while avoiding falling into the deep pits of the "telephone game" as we take many turns to solve a complex problem. Just a few short months later, and Ralph is taking the wheel. "Harness engineering" entered the realm. Just let the LLM do it all: let it bang its proverbial head against the wall again and again, keep track of its own outcomes and let it learn as it goes. With the added twist that the learning gains are not that of Lisa Simpsons but rather that of Ralph Wiggum. RL but it escaped the labs, and instead of steroids we put it on Xanax. This is how we can solve all problems now. Nothing to worry about here. "But wait there is more." Let's not call it Ralph, because that would literally sound stupid. Instead, let's rebrand it as "loop"ing. And instead of building the loop ourselves, let the LLM write the loop, build the harness, and figure out all the difficulties there. Even more complex. Even more layers of abstraction and momentum build-up, an even longer horizon whose trajectory absolutely no one can predict anymore, all the while adding variables to account for and even more cognitive holes for the LLM to fall into. According to Boris Cherny: That's it. Coding is solved. Uninstall your IDE now and not at a later point when this is proven to work because the AI is taking care of everything now. Don't worry, even though it has all your API keys, partial access to prod and full access to your CI via pipeline files, I'm sure it won't abuse any of that. Let the LLM do it all while you get your beauty sleep. Sweet dreams. Original published at: https://www.linkedin.com/posts/dominik-seifert-phd-8b663b54 we-went-from-prompt-engineering-to-context-share-7472275432722169856-x8R / https://www.linkedin.com/posts/dominik-seifert-phd-8b663b54 we-went-from-prompt-engineering-to-context-share-7472275432722169856-x8R /