Yeah - code rewrites are generally a thing to be avoided since the requirements of older production code is typically not fully documented, even if it originally had been. Production code accumulates bug fixes, performance tweaks, special cases, additional requirements, etc, with documentation and test suites not always being updated in sync.
Your rewritten codebase will presumably look much cleaner and more maintainable that what it's aiming to replace, but it's easy to miss the purpose/necessity of all of the old "cruft" which need not be just poor coding but rather later enhancements not anticipated by the original design whose implementation was therefore crufty.
Does anyone here work at Cisco, and have any insight as to why a rewrite is considered necessary and might be expected to succeed?
A rewrite is a slam dunk for any decent frontier AI. It has a penchant for maintaining API/etc contracts to a fault in the absence of instructions to the contrary. It can easily validate its work and loves building exhaustive test cases. You can tell it to get 100% code coverage if you want. It will laborously accurate document every accumulated edge case and help you find a better overall cohesive structure.
A rewrite has historically been a huge mistake in many situations, some of which I've personally had the agony of witnessing firsthand, powerless to stop it. But AI completely changes the calculus because it doesn't waste significant developer time or energy crawling back to the previous working state, it just does it while you go get coffee. And this I've found to be a huge benefit, that I can continue making my codebases nicer to work with and better documented/integrated with bugs stamped out for free. There are upsides to a rewrite, and they're now much more likely to be worth it as the downsides are automated away.
> It has a penchant for maintaining API/etc contracts to a fault in the absence of instructions to the contrary.
AI has no magic powers to figure out what the “API/etc contracts” are in a nasty old codebase. This problem is fundamentally hard and not always well defined.
AI is also (at least for now) not consistently good at writing tests that test the behavior that matters. Recent models seem to be getting slightly less bad, but again, this is a hard problem.
I’ve personally used AI to rewrite gnarly code where the goal is to have exactly identical results to the old code. GPT-5.6 Sol could do this correctly with some extra help guiding it to rest the right things, and I haven’t tried this particular task on a newer model.
But ancient code like in the article is probably an enormous buggy mess full of accidental behaviors and implementing protocols in ways that do not fully match the spec. Some, but not all, of the mismatches will be cases where the spec is wrong. Sure, an AI can probably generate a new implementation with a similar set of bugs, but is this what is actually wanted? Also, depending on the gnarly set of bugs, a clean implementation may be unable to reproduce the bugs, leading to a messy implementation instead.
If this was true then Python would already have been rewritten to be GIL-free, and the entire ecosystem of Python packages would have been rewritten to respect its API.
I'd even argue AI is better for maintaining/moderninzing existing code than writing it, just because there is so much less input from developer about architecture it requires, as it has it in the code itself.
But there is pretty big risk of false positives and negatives; some "bugs" are design decisions that were not documented well/at all, and inevitably some new ones will appear when it reasoned something wrong.
The entire point of a rewrite is to get rid of the behaviour you don't want, while you keep the behaviour you do want. Consider the pathological case: Rewriting your FORTRAN codebase in Rust would obviously be useless if you had to reimplement every bug in the rewrite.
An LLM doesn't have enough context to know what you so and don't want. If you give it call logs for every function that was called and how, maybe, but a seasoned engineer will have more context around what the company needs.
While I certainly don't work at Cisco, I've come to the conclusion that I need to re-examine my priors when it comes to code rewrites and general maintainability.
AI makes these projects significantly more feasible, at least in my domain. We have some critical-path software I wouldn't touch with such an approach, but the vast majority of our code is internal and/or supporting tooling.
For those cases, burning it down and rewriting via AI tooling has proven far more productive than anything we have attempted in the past. This is for old technical debt laden stuff everyone has been afraid to touch for well over a decade.
It's not perfect, but it moves the needle where before a lot of such projects were stuck on the backburner for years.
It certainly introduces new bugs and such, but at least this way forward momentum is happening vs. just being stuck in decision/analysis paralysis forever.
This also assumes some very senior SME folks are driving the process. I would not expect similar results with junior or even mid-level folks in these specific domains.
> While I certainly don't work at Cisco, I've come to the conclusion that I need to re-examine my priors when it comes to code rewrites and general maintainability.
One of the reasons to avoid rewriting has long been that the time spent rewriting code that works is time not spent writing code to meet new business needs. In other words, it's usually a huge opportunity cost, rarely balanced out by sufficient gains to be net positive.
I do think AI does challenge this prior, as the time needed to do a rewrite shrinks dramatically. It's not zero, of course, but it now becomes worth asking if it's worth doing, rather than defaulting to avoiding rewrites in nearly all cases.
It’s been a while, but configuring their hardware manually in 2016 was bug ridden ridiculousness. Awful UI/UX. Their own techs had specific field notebooks for our make and firmware version that we couldn’t access unless persuaded, detailing various land mines to avoid while working. Literal “don’t click these things in this order” kinds of notes.
A critical security firmware update once brought down redundant pair of firewalls we used because the bug didn’t manifest until a counter buffer overflowed months after application. Three hours into the outage the tech power cycles the thing and our services came back to life.
Cisco is only doing this because some other alternative is more expensive. They had more than a decade to get started.
We had to turn off a bunch of on-by-default features in the system I'm deploying now because IOS matches too aggressively and decides that arbitrary UDP datagrams are DHCP requests or PTP packets that need rewriting (and then dropping when they turn out to not be DHCP/PTP). I have to think that an AI rewrite with a today-good model would recognize the lax checks and do them better.
Yeah - code rewrites are generally a thing to be avoided since the requirements of older production code is typically not fully documented, even if it originally had been. Production code accumulates bug fixes, performance tweaks, special cases, additional requirements, etc, with documentation and test suites not always being updated in sync.
Your rewritten codebase will presumably look much cleaner and more maintainable that what it's aiming to replace, but it's easy to miss the purpose/necessity of all of the old "cruft" which need not be just poor coding but rather later enhancements not anticipated by the original design whose implementation was therefore crufty.
Does anyone here work at Cisco, and have any insight as to why a rewrite is considered necessary and might be expected to succeed?
A rewrite is a slam dunk for any decent frontier AI. It has a penchant for maintaining API/etc contracts to a fault in the absence of instructions to the contrary. It can easily validate its work and loves building exhaustive test cases. You can tell it to get 100% code coverage if you want. It will laborously accurate document every accumulated edge case and help you find a better overall cohesive structure.
A rewrite has historically been a huge mistake in many situations, some of which I've personally had the agony of witnessing firsthand, powerless to stop it. But AI completely changes the calculus because it doesn't waste significant developer time or energy crawling back to the previous working state, it just does it while you go get coffee. And this I've found to be a huge benefit, that I can continue making my codebases nicer to work with and better documented/integrated with bugs stamped out for free. There are upsides to a rewrite, and they're now much more likely to be worth it as the downsides are automated away.
> It has a penchant for maintaining API/etc contracts to a fault in the absence of instructions to the contrary.
AI has no magic powers to figure out what the “API/etc contracts” are in a nasty old codebase. This problem is fundamentally hard and not always well defined.
AI is also (at least for now) not consistently good at writing tests that test the behavior that matters. Recent models seem to be getting slightly less bad, but again, this is a hard problem.
I’ve personally used AI to rewrite gnarly code where the goal is to have exactly identical results to the old code. GPT-5.6 Sol could do this correctly with some extra help guiding it to rest the right things, and I haven’t tried this particular task on a newer model.
But ancient code like in the article is probably an enormous buggy mess full of accidental behaviors and implementing protocols in ways that do not fully match the spec. Some, but not all, of the mismatches will be cases where the spec is wrong. Sure, an AI can probably generate a new implementation with a similar set of bugs, but is this what is actually wanted? Also, depending on the gnarly set of bugs, a clean implementation may be unable to reproduce the bugs, leading to a messy implementation instead.
1 reply →
If this was true then Python would already have been rewritten to be GIL-free, and the entire ecosystem of Python packages would have been rewritten to respect its API.
11 replies →
I'd even argue AI is better for maintaining/moderninzing existing code than writing it, just because there is so much less input from developer about architecture it requires, as it has it in the code itself.
But there is pretty big risk of false positives and negatives; some "bugs" are design decisions that were not documented well/at all, and inevitably some new ones will appear when it reasoned something wrong.
The entire point of a rewrite is to get rid of the behaviour you don't want, while you keep the behaviour you do want. Consider the pathological case: Rewriting your FORTRAN codebase in Rust would obviously be useless if you had to reimplement every bug in the rewrite.
An LLM doesn't have enough context to know what you so and don't want. If you give it call logs for every function that was called and how, maybe, but a seasoned engineer will have more context around what the company needs.
5 replies →
While I certainly don't work at Cisco, I've come to the conclusion that I need to re-examine my priors when it comes to code rewrites and general maintainability.
AI makes these projects significantly more feasible, at least in my domain. We have some critical-path software I wouldn't touch with such an approach, but the vast majority of our code is internal and/or supporting tooling.
For those cases, burning it down and rewriting via AI tooling has proven far more productive than anything we have attempted in the past. This is for old technical debt laden stuff everyone has been afraid to touch for well over a decade.
It's not perfect, but it moves the needle where before a lot of such projects were stuck on the backburner for years.
It certainly introduces new bugs and such, but at least this way forward momentum is happening vs. just being stuck in decision/analysis paralysis forever.
This also assumes some very senior SME folks are driving the process. I would not expect similar results with junior or even mid-level folks in these specific domains.
> While I certainly don't work at Cisco, I've come to the conclusion that I need to re-examine my priors when it comes to code rewrites and general maintainability.
One of the reasons to avoid rewriting has long been that the time spent rewriting code that works is time not spent writing code to meet new business needs. In other words, it's usually a huge opportunity cost, rarely balanced out by sufficient gains to be net positive.
I do think AI does challenge this prior, as the time needed to do a rewrite shrinks dramatically. It's not zero, of course, but it now becomes worth asking if it's worth doing, rather than defaulting to avoiding rewrites in nearly all cases.
Does the coding assistant demonstrate any tricks to improve the likelihood of success for such rewrites?
1 reply →
> Your rewritten codebase will presumably look much cleaner and more maintainable that what it's aiming to replace
Until one day you leave and the next guy comes and says “what is all this shit”
It’s been a while, but configuring their hardware manually in 2016 was bug ridden ridiculousness. Awful UI/UX. Their own techs had specific field notebooks for our make and firmware version that we couldn’t access unless persuaded, detailing various land mines to avoid while working. Literal “don’t click these things in this order” kinds of notes.
A critical security firmware update once brought down redundant pair of firewalls we used because the bug didn’t manifest until a counter buffer overflowed months after application. Three hours into the outage the tech power cycles the thing and our services came back to life.
Cisco is only doing this because some other alternative is more expensive. They had more than a decade to get started.
Cisco iOS even on terminal servers in the early 2000s was...special.
It's because the old bits are starting to stink a bit from age. Get those stinky bits out of here, I say! Let the AI make some new bits!
Because it's Cisco?
We had to turn off a bunch of on-by-default features in the system I'm deploying now because IOS matches too aggressively and decides that arbitrary UDP datagrams are DHCP requests or PTP packets that need rewriting (and then dropping when they turn out to not be DHCP/PTP). I have to think that an AI rewrite with a today-good model would recognize the lax checks and do them better.
No goal, and no results, but a younger guy was enthused about AI!
Plus: publicity!
It is not in Rust.
All of the known bugs will be gone.
AI has electrolytes!
It’s what the shareholders crave!
They also seem to crave everything that has the word AI in it.
2 replies →
The hardware it runs on has capacitors as its components and some of them do in fact contain electrolytes.
It's what Cisco does, as they've done several times before.
It's got tabs instead of spaces. Yuck!
Somebody needs a promo packet. Duh
[dead]