Making the easy thing the right thing
Listening to an interview with xAI software engineer Lauren Tan, she says:
"I think agents - from my experience, even the frontier ones - tend to take shortcuts. They tend to do the easy thing.
So a lot of the [agent skills] I've built have been around how do I make the easy thing the right thing."
The same in my experience, too.
And notice the balance between the old ways and the new here. How can you tell whether the coding agent is doing the 'right thing'? Writing code that is secure, maintainable, robust, and balances current needs with requirements you know are coming down the road?
That's where engineering foundations come in. The poor souls who lack those foundations look at the code the agent pops out, and have no clue whether it is good or bad. But when your foundations are solid, you can often tell right away.
Then comes the next part, where you must articulate clearly and thoroughly what you want the coding agent to do differently. So the same problem does not pop up again, next time.
That's where the 'agent skills' thing comes in, where you write out English words (or whatever language) that get injected into the coding agent's context. Making the easy thing the right thing.
Lots going on here, actually.
Because in order to pull this off, you need to know about:
- engineering foundations
- agentic coding
- content creation skills
- system design
(yes, the same writing skills that let you prompt a coding agent well, also let you create content that boosts your professional status)
If you are in Boosted Code, which of these four do you most need to improve in? That's what you want to focus on in the curriculum.
And if you want to become the kind of engineer who wields coding agents in this way, learn about Boosted Code here: