This is how you get to 10x with a LLM in my limited agentic experience for doing all the scratchwork of architecting a program before you generate it:
The user facing part of your program can be planned out using conceptual design https://essenceofsoftware.com/tutorials/ the author of that book teaches it in MITs old software studio course https://61040-fa25.github.io/schedule the point is to plan out modularity and the prof does enjoy using overly complex language to describe this method but once you read through the slides and tutorials you will understand why he describes it that way because he's trying to differentiate between features and concepts. For example HN has an upvoting concept who's purpose is to establish rank then a seperate concept karma which allows you to downvote but placing both functions inside the upvoting concept breaks modularity and conceptual design makes this obvious once you practice with it. Once everything is planned out this way then generating code is trivial again in my limited experience as I'm no expert on agentic coding but I've had success doing this.
All the code the user won't see can be modeled using one of the 'lightweight' formal methods out there like forge or alloy https://forge-fm.github.io/book/2026/ where a complex protocol you write or entire system can be tested first to find illicit states. Imagine you are designing some company app where there needs to be multiple logins of different security abilities this is how you would model your security plan first and make sure a combination of states doesn't produce an unexpected breach or that you missed a state. A custom network protocol that does kernel bypass is another example. The rules of a game you build is another you don't want the system to reach a state like winner unless they actually won. I now use Forge to plan css too because I don't want to show broken css states since I have limited design experience.
Now generate the whole system as modules and never look at the code. The same property tests I used for the Forge model I make into an Oracle and then blast the agent code with random inputs.
I built several gigantic prototypes this way mostly of papers I read in database designs and screwing around with graphical interfaces for them.
I still write anything critical by hand. For example some app that accepts uploads of PDFs to change format I'm going to escape and prep that data by hand so the vibe coded microservices only ever see cleaned/safe data.
If you can skip some or all electives and instead receive credits for joining research on campus as an undergraduate then it's worth it. There were a few TAs at my school who joined some professors project and only did that their last two years of undergrad where they had to produce a research paper for credit but they all ended up graduating with their name on multiple papers, going to invite only graduate summer schools, getting employment or grad school offers none of the rest of us were offered. Otherwise the cost seems too high now for a 4yr undergrad unless you scheme for a cheap bachelors (Europe? UoPeople?) and invest in a masters instead
AI copilots and prompts give me massive lines of imperative OCaml and the interface for that code always requires changing to properly describe the data it will receive when I can write it myself in a few minutes. I can however write a simulation of some hardware quickly with Java or C using claude code and then run my hand written programs in there for testing. An example is mimicking the runtime environment of some controller
Had to modify some program so just read the code and guessed mimicking what I read. It mostly just worked so continued doing that and reading enough docs I could write my own. Did that hacking for years until I finally read a real book on programming languages I think it was SICP but could have been something else before that like a SML compiler book
I assume this works where the ransomware authors, who likely are in some untouchable nation and the son of some major politician, provide a binary/kit with their own addresses to take the ransom then pay the person who planted it out minus their cut. Those wallets used for paying crime commissions are probably reused often or otherwise identified as they don't care if you get caught and you need to either sit on those coins for years until the limitations runs out or have enough knowledge to (correctly) wash them and anyone doing this is already making bad life decisions so likely greedy and cashed those in a traceable way like driving to work in his new Ferrari.
Sometimes except I learned the hard way that if you write everyday Python math code it's actually variable-time arithmetic and totally unsuitable for applied cryptography, oops
These sort of books are meant to be discursive in nature giving one an exposure to various facets in a domain. That is their "theme". For a layman/common reader this is actually quite welcome since it does not drown-him-in-depth/demotivate-him but gives an overview (possibly interconnecting) of the subjects involved.
For example the Stillwell book i mentioned above has chapters on Arithmetic, Computation, Algebra, Geometry, Calculus, Combinatorics, Probability and Logic thus making it an excellent book to peruse.
Math was always the stumbling block for me then I realized through work how easy it is with pure rational or reals except when you need to translate it to an algorithm now you are throwing out all the elegant linear algebra for a numerical representation and using obscure characteristic polynomial constructs just to run a program and now math sucks again I went back to the beginning.
A nice course for this is of course TAOCP volume 2 old testament or MITs math github https://github.com/mitmath/18335/tree/spring22 (change the yr to suit) like we can't even have nice things like gradient descent anymore because it zigzags and is too inefficient
The user facing part of your program can be planned out using conceptual design https://essenceofsoftware.com/tutorials/ the author of that book teaches it in MITs old software studio course https://61040-fa25.github.io/schedule the point is to plan out modularity and the prof does enjoy using overly complex language to describe this method but once you read through the slides and tutorials you will understand why he describes it that way because he's trying to differentiate between features and concepts. For example HN has an upvoting concept who's purpose is to establish rank then a seperate concept karma which allows you to downvote but placing both functions inside the upvoting concept breaks modularity and conceptual design makes this obvious once you practice with it. Once everything is planned out this way then generating code is trivial again in my limited experience as I'm no expert on agentic coding but I've had success doing this.
All the code the user won't see can be modeled using one of the 'lightweight' formal methods out there like forge or alloy https://forge-fm.github.io/book/2026/ where a complex protocol you write or entire system can be tested first to find illicit states. Imagine you are designing some company app where there needs to be multiple logins of different security abilities this is how you would model your security plan first and make sure a combination of states doesn't produce an unexpected breach or that you missed a state. A custom network protocol that does kernel bypass is another example. The rules of a game you build is another you don't want the system to reach a state like winner unless they actually won. I now use Forge to plan css too because I don't want to show broken css states since I have limited design experience.
Now generate the whole system as modules and never look at the code. The same property tests I used for the Forge model I make into an Oracle and then blast the agent code with random inputs.
I built several gigantic prototypes this way mostly of papers I read in database designs and screwing around with graphical interfaces for them.